Hello, everyone, and welcome to all newly joined subscribers!

This is the 75th edition of The Web3 + AI Daily - your definitive guide to the intersection of blockchain and AI. Today, I’m taking a deep dive into Aztec, Ethereum’s privacy-focused Layer-2 chain, and the EU AI Act, which has just partially entered into force.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

Aztec Launches Alpha V5, Cuts Team in Half

Those of you who regularly read my newsletter know that privacy is a topic I care deeply about. It’s also one of the most consequential verticals in the crypto space, whose importance is only growing as AI becomes more pervasive. That’s why today I’m taking a closer look at Aztec - the Ethereum Layer-2 chain designed for private transactions.

It’s been a couple of turbulent months at Aztec. After launching its mainnet and token $AZTEC in January, in June the network suffered two consecutive hacks within the span of a single week.

Aztec Network has been hit by another $2 million hack, its second this week. Following Sunday’s $2.2 million loss from Aztec Connect, Aztec’s Private Rollup Bridge has now been drained of a similar amount. The firm stressed, in both cases, that the affected contracts are “immutable” and were “deprecated” in 2022 and 2023.

The Aztec Foundation also stated that neither exploit had any connection to the current network or the AZTEC ERC-20 token, emphasizing that the affected contracts belonged to legacy infrastructure that had remained live on Ethereum after the products were sunset.

These incidents didn't stop the team from building, though, since in in mid-July Aztec released Alpha V5 - a major protocol upgrade introducing faster and cheaper private transactions, alongside applications that combine confidential user data with public blockchain state.

Alpha V5 reduces private-transaction proving times by more than 2x compared to the previous version, lowers the cost of a fully private transaction by roughly 50%, resolves the critical issues found in V4, and sees the first wave of apps go live.

As the only decentralized privacy L2, Aztec is the credibly neutral privacy layer for Ethereum. Aztec allows anyone to write smart contracts that include both private and public aspects – every private transaction is proven on the user’s own device, so no operator, sequencer, or intermediary can see the data.

Why is this important? Aztec opens the door for a wide range of industries and use cases to leverage decentralized computing without compromising the confidentiality of their data. Something as simple as receiving your salary on-chain can now be performed risk-free.

Unlike Ethereum’s base layer, where every validator processes and stores transaction inputs, outputs, and execution data to reach consensus, Aztec’s execution layer moves private computation to the user’s device. Instead of revealing transaction details to the network, the system generates cryptographic proofs locally before submitting them for verification on-chain, reducing the amount of visible transaction data while preserving Ethereum’s security guarantees.

Moreover, along with Alpha V5 came Shield - a private bridge for moving assets between Ethereum and Aztec, powered by human.tech by Holonym Foundation. This is a significant breakthrough that, if successfully applied across the industry, could revolutionize compliance and KYC (Know Your Customer) practices.

Shield lets users bridge assets while proving the sender is legitimate and the funds are compliant, all without revealing their identities.

Finally, last week, Aztec announced a 18-person layoff, representing roughly 55% of the workforce.

[Co-founder Zac] Williamson positioned the restructuring as the Foundation transitioning from a build-oriented organization to one focused on security and scalability of the decentralized network. The remaining 15-person team will concentrate on keeping things stable and secure rather than shipping new features at breakneck speed.

Building a fully decentralized, privacy-first L2 from scratch takes a big organisation focused on that specific task. Securing one, and filling it with applications people actually use, takes a different set of skills dedicated to stability and commercial success. . We shipped the network. The job in front of us is different from the job behind us, and the Foundation needs to adapt to meet this moment.

Aztec claims that the foundation is well capitalized "for many years of operation at this size." Speaking from experience, I don't believe such a significant restructuring is driven solely by a shift in strategic focus - development work will continue behind the scenes regardless. Rather, it suggests that, like much of the crypto industry, the bear market is taking a heavy toll on the company.

More

The EU AI Act Enters Into Force

August 2nd, 2026 marked the day when some of the provisions of the EU AI Act took effect, along with new transparency obligations for AI providers and deployers. The following post by Oliver Patel, AIGP, CIPP/E, MSc informs on all you need to know:

However, as Bruna Santos and Jacobo Castellanos commented for Tech Policy Press, some of the exemptions set out in the guidelines are so vague that they open friction points and risk watering down the rules:

The first is vagueness in the key exemptions. The “obviousness” exception for interactive AI, and the line between exempted “standard editing” and regulated “substantial semantic alteration”, both rest on subjective, case-by-case judgements, which creates real uncertainty for developers, deployers and platforms that blend ordinary editing with generative tools.

The second is an internal tension on watermarking. The guidelines acknowledge that universal, provider-agnostic watermarking and detection do not yet exist, especially for text, while the Code requires synthetic outputs to be machine-readable and detectable through “effective, interoperable and reliable” solutions. As we note above, detection is genuinely unreliable, so the candour is welcome, but the gap between what the guidelines admit and what the Code mandates risks eroding confidence in the marking regime just as it takes effect.

Meanwhile, there are numerous provisions that are yet to become applicable and enforceable:

Not every part of the AI Act is moving at the same speed. The AI Omnibus, a package of amendments to the Act, pushed back the rules for high-risk AI systems to 2 December 2027, and those for high-risk systems built into regulated products to 2 August 2028. The same package moves faster on the harm side. From 2 December 2026, it bans AI systems that generate non-consensual sexually explicit content or child sexual abuse material.

Several different bodies will be enforcing the rules, including the AI Office, the national competent authorities, and the European Data Protection Supervisor. In short, the work is ongoing and there’s still a lot to be done to ensure EU citizens’ safety.

I must say, I’ve been critical of the EU AI Act in the past, largely because I believed that overregulation would further slow an AI industry that was already lagging behind the U.S. However, while Europe still has a long way to go, I’m increasingly grateful to live in a place where the whims of tech billionaires aren’t simply accepted without scrutiny.

I’ll continue following closely the effects of the AI Act and discuss them with you here. Let me know your thoughts.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is the 74th edition of The Web3 + AI Daily - your definitive guide to the intersection of blockchain and AI. Today’s agenda revolves around AI inference, as it has overtaken training as the dominant share of global GPU demand. As a result, not only that Web3 companies like Boundless are redirecting their GPU clusters to AI inference, but on-chain inference capital markets started to emerge.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

Boundless Shifts Focus from ZK to AI

The decentralized compute firm Boundless has announced a pivot to AI, involving an expansion of its GPU network into AI inference.

Boundless initially focused its 4,000 GPU cluster on settling Ethereum- and Base-based zero-knowledge proofs on Bitcoin.

The move is in line with the wider trend of compute networks and mining companies directing GPU capacity towards AI operations, to address the unsatiable thirst for machine intelligence.

Of note, Boundless said it will continue running its zero-knowledge proving network in parallel, and plans to “give its native token, ZKC, a role in its AI network,” by requiring AI operators to stake ZKC to join the network, according to the announcement.

Web3 + AI Readings & Conversations

Centralized vs. Decentralized Inference Providers

Maybe I should have started by explaining what AI inference is. That’s the process of an already trained AI model running a prompt, an agent loop, an image or a video generation. Every time you ask your chatbot a question, it processes your inquiry to deliver a response through the act of inference.

As you can imagine, there are centralized and decentralized inference providers. In the centralized category, besides hyperscalers like Microsoft, Google, and Amazon, one can also find platforms like Fireworks AI, Together AI, Replicate, Baseten, and Groq.

At the other end of the spectrum are companies building the permissionless and confidential alternative, such as Chutes AI, Dolphin AI, Pearl Research Labs, Venice.ai, and others.

These two categories of inference providers often come with diametrically different value propositions. As writes:

The mistake is to compare all of these providers as if they are competing in the same market; they are not.

Traditional providers sell reliability, developer experience and enterprise procurement.

Crypto AI networks sell cheaper supply, open access, privacy, verifiability and new incentive loops.

Decentralized Inference: Confidentiality, Privacy, Affordability

I’ve often discussed Web3 inference providers, but I believe it would be useful to compile a list of the most prominent ones and highlight what differentiates each of them. Here it is:

• Chutes AI combines an offering of open-source models with using decentralized GPU supply underneath. I’ve recently covered a breakthrough of theirs (Check Daily #71).

• Targon AI / Manifold Labs focuses on delivering confidential compute to serve data-sensitive industries like finance and healthcare:

The problem it addresses is obvious: many users will not run sensitive prompts, models or data on infrastructure operated by unknown third parties. Targon’s answer is protected execution through trusted execution environments, encrypted virtual machines, remote attestation and confidential GPU infrastructure. In plain English, the aim is to prove the workload is running in a secure environment and reduce what operators can see.

• Venice.ai is best described as an encrypted and private AI chatbot, although, unlike ChatGPT for instance, it gives access to various closed- and open-source models. Venice boasts that it collects no user data and that prompts are never stored.

• Dolphin AI began by offering uncensored open models, and then built the inference network around it. “Its architecture is often described as peer-to-pool, which means [that] GPU owners contribute capacity into model-specific pools, rather than each buyer renting a specific node directly.”

• Pearl Research Labs is a blockchain network that replaces Bitcoin’s “SHA-256 hashing algorithm with matrix multiplication, the core operation in AI inference and training. Pearl’s claim is that the same matrix multiplication that serves a customer’s inference can double as a mining attempt.” What does that mean in practice? Instead of spending huge amounts of computation and energy on math solutions that nobody uses, the network validators compete to win the right to mine the next block by performing actual AI operations.

• Similarly to Pearl, Ambient also let miners run an AI model. However, unlike Pearl, it “standardizes the entire network on one large open-weight model and builds its consensus around verifying that model’s output.”

Intelligence Is Becoming Tradeable, Both On- and Off-Chain

Commoditizing AI Inference: Who, What, Why?

Naturally, with the number of people and enterprises using AI, and AI agents actively operating both on the rise, the need for computation resources continues to grow exponentially.

The demand for intelligence is swallowing the industry whole. AI providers don’t have enough compute to service the demand, leaving the gap for Inference providers to step in and service those demand.

Just as enterprises are currently shifting away from frontier AI labs toward more affordable open-source models, they are also starting to switch to more cost-efficient inference providers. Moreover, as inference is becoming an ever more valuable resourse, subscribers of AI models now sell their unused credits, while owners of idle GPU clusters monetize their capacity.

As 0xSammy noted, the inference market will soon resemble the electricity one: many suppliers with similar offerings, competing more on reliability and resilience than on price. Yet, the Web3 alternatives still boast much lower prices than hyperscalers.

On-Chain Inference Markets

As tradition goes, the Web3 space supercharges the trend of AI inference commoditization with experimentation, incentive design, and innovative tokenomics. The inference is being tokenized, not to mention verified, and it’s becoming yield- or reward-bearing.

Last week Galaxy published an in-depth analysis into the emerging AI inference markets, and how the Web3 space is pricing compute, tokenizing access, and financing hardware. I’ll share a summary below, but I still recommend you read the full article.

The term onchain inference capital markets describes the set of networks, protocols, supporting infrastructure, and applications coordinating AI model inference outside the centralized API surface controlled by frontier labs and hyperscalers, together with the financial layer now forming on top of that activity. Rather than routing every API call through frontier model providers like OpenAI, Anthropic, or the underlying cloud providers that service them, users can send prompts to networks of GPU operators coordinated by crypto token incentives and onchain settlement, and in some configurations receive cryptographic or economic guarantees about output correctness and privacy.

In recent years, decentralized GPU marketplaces, inference protocols, payment rails, tokenization, capital formation vehicles, and onchain liquidity each had its moment. What is new is that these primitives are converging into a single integrated system, an inference capital market, which is projected to find growing demand as inference is increasingly used for all work.

At first glance, the main difference with regular inference providers is that the crypto-native ones can “source capacity from decentralized GPU networks, accept stablecoins or tokens as payment, include privacy guarantees, or attach tokenized access rights to usage.”

However, the uniqueness of what Web3 has to offer comes to light on the financialization side, where crypto changes how inference is owned, priced, and financed. As Galaxy outlines:

Financializing inference has attracted a range of onchain projects that use blockchain payment rails and tokenization to turn inference activity into tradable assets. This takes three forms.

Inference service providers like Venice and Morpheus tokenize inference access, turning a claim on future inference into something that can be held, priced, and resold.

Proof of useful work projects like Pearl and Ambient tokenize inference production, paying out a token for the work of serving it.

Credit providers like USD.AI do something different. Rather than tokenizing inference, they finance the hardware it runs on, using stablecoin deposits to fund the GPUs and data centers underneath.

Collectively, these components come together to form onchain inference capital markets.

Venice AI’s Model

All of these models are quite interesting, but I’m sure the tokenomics nerds among you would appreciate the Venice.ai one the most. It uses a two-token system to turn inference into an ownable and easily transferable asset. Here’s how it works in short:

• $VVV is the main Venice token. Holders can stake it to earn yield or to mint the $DIEM token.

• Venice runs a buy-and-burn mechanism on $VVV - a discretionary burns funded from general revenue, and a programmatic burn that routes a fixed portion of every new subscription into buying and burning the token. To date, 42% of VVV has been burned.

• Each $DIEM token gives holders $1 of daily credits to spend on Venice’s AI tools, in perpetuity. “It’s a tokenized platform credit that expires and replenishes itself daily.”

DIEM is an experiment in how to tokenize and deliver inference access. What makes it distinct is ownership. It lets users own the inference they consume rather than rent it. A buyer paying per request gets nothing back once the inference is spent, while a holder of tokenized access owns an asset they can keep, transfer, or sell.

DIEM wraps a claim on future inference into something a holder can mint, own, and resell.

Since $DIEM represents a $1/day inference credit, a user can decide not to use it and sell it instead. Accordingly, people have already started speculating on its future value and secondary markets have appeared:

Secondary markets enable users to stake DIEM or sell the unused credit at a discount. Thus, earning yields 10-20% APR on DIEM. Morpho has this wstDIEM/DIEM where wstDIEM (LSD version of DIEM) can be used as a collateral to borrow DIEM which essentially allow users to leverage farm DIEM (up to ~56% APR according to Liquid protocol, the team behind wstDIEM).

Decentralized inference is a fascinating subject that I’ll keep covering here. Let me know which aspects or companies you find compelling, I’m looking forward to hearing your thoughts.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the intersection of blockchain and AI! Two of the most influential voices in AI - Microsoft‘s Satya Nadella and Google DeepMind‘s Demis Hassabis, have recently shared their visions for the future of AI. Both essays sparked widespread discussion, so today I’m offering my perspective on each.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

Microsoft’s Satya Nadella Makes The Case For Decentralized AI

In a recently published essay, Microsoft‘s Satya Nadella outlines a fundamental defect in today’s centralized AI industry: The fact that customers must relinquish their personal or enterprise data to use advanced AI systems.

In the AI age, the buyer risks giving away knowledge, just in order to use what they bought.

You essentially pay for intelligence twice, once with money, and again with something even more valuable: the proprietary knowledge you must reveal to make that intelligence useful. The better you want the model to perform, the more of that knowledge you have to feed it!

Over time, the information asymmetry becomes increasingly skewed. The seller learns more and more about you as you use what you purchased, while you learn very little about what the seller is learning in return.

The CEO of Microsoft, the company that has invested in both OpenAI and Anthropic, goes on to critisize these same labs for preventing clients from studying their frontier models and using the insights to train their own models.

Nadella’s point is that model makers can’t have it both ways. It’s hypocritical for them to freely train on the world’s data while restricting others from doing the same to their models.

What Nadella presents as the solution is, in fact, what the Decentralized AI community has been championing for years: data ownership, proprietary learning infrastructure, and private AI inference.

That is why enterprises need a real trust boundary for their human capital and token capital to compound. It is where an organization’s data, traces, evals, adapted weights, and memory accumulate and improve together. And it is a hard boundary across which nothing crosses, not even the intelligence exhaust, without consent. Enterprises will demand the rights to use model outputs to fine tune and/or train their own models. I think of this as every firm’s right to align models to their enterprise accountability obligations.

Moreover, he virtually urges enterprises to move away from relying on one single “generalist” model and instead build “orchestration layers” and “continuous learning loops” where it is easy to switch between different models. As TechCrunch comments:

While Nadella never uses the words “open source” as the method for retaining ownership, this is an obvious subtext. Yet, there’s another subtext.

After experimenting with proprietary model makers, [large companies] start asking themselves: “Can I take an open source model and run it on-prem? It will do almost 90% of what the big one’s doing. It will cost way less.”

Though, Nadella goes even further - while echoing Alex Karp, he scolds “the current regime” for not allowing customers to have control over their own compute resources, models, data stack, and alpha:

As Alex Karp put it: “What the technical customers want is control over their compute, their models, their data stack, and their alpha. They want to know they own the means of production, and it’s not being transferred to someone else.” The current regime does precisely the transfer Karp and companies fear.

TechCrunch reads this as urging corporations to build their own AI infrastructure on the cloud - preferably, Microsoft Azure:

Nadella’s solution is the kind of thing the CEO of a giant cloud provider would suggest. He wants companies to “retain ownership” of their data, including prompts, feedback, etc. So he’s urging them to build their own “proprietary learning environments” on the cloud (where their data is likely already stored anyway and, conveniently, could mean Microsoft’s cloud, Azure).

But I don’t. Microsoft’s cloud is still a centralized Big Tech platform. As long as your AI stack depends on infrastructure you don’t own or control, you cannot claim true sovereignty over your data, models, or compute. If a company truly wants to retain control over its AI infrastructure, it would be better served by a distributed cloud - a more resilient and cost-effective alternative.

I’m generally puzzled by Nadella’s stance. Don’t get me wrong - I appreciate the inadvertent endorsement, because his essay underscores exactly why Decentralized AI matters. For years, the DeAI community has advocated for distributed training and inference, verifiable computation, and open-source models as the only viable path to making AI private, safer, more affordable, and universally accessible.

Yet, Nadella - and, for that matter, Karp - is not a disinterested observer or some bistanding analyst. He is an integral part - not to say one of the architects - of that “current regime” he mentions. After all, Microsoft is the backbone of today’s centralized AI industry.

Which begs the question: what is he really suggesting? His proposal cuts against the business models of OpenAI and Anthropic, and, by extension, Microsoft’s own financial interests as their key partner.

I completely agree with Nadella that the future of the AI belongs to open-source, privacy-preserving alternatives. But where would that leave OpenAI, Anthropic, Meta, SpaceXAIA, and the rest of the hyperscalers that Microsoft ultimately supports?

Demis Hassabis Urges the U.S. to Lead Frontier AI Standardization

Demis Hassabis, CEO and co-founder of Google DeepMind and a laureate of The Nobel Prize for his work on AlphaFold, calls for a global framework for Frontier AI security, but one lead by the United States of America.

At first glance, what Hassabis suggests is sensible. He turns to the AI industry, and the world as a whole, and pleads: AI will be creating unprecedented challenges, so instead of racing each other, let’s collectively try to mitigate the risks. I’m paraphrasing. However, what he proposes in practice is much more nuanced:

The rapid progress we’re seeing in AI requires a new approach to testing frontier AI model capabilities that is dynamic, adaptable, and rigorous. The US is well positioned, given its economic and technical standing, to take the first step in developing such a framework. It could establish a new Standards Body modell ed on a federally overseen public-private partnership or self-regulatory organisation, much like the Financial Industry Regulatory Authority (FINRA), with a board that includes independent leading technical experts and open-source representatives.

This US-initiated effort would provide a strong starting point for creating shared international standards on Frontier AI. Since this technology is going to affect the entire planet, ideally this framework would spur the international community to reach a consensus on how to manage the most serious risks while ensuring everyone has access to and can benefit from the opportunities that AI brings.

The article was met with a hint of indignation and provoked two reasonable questions which I was also inclined to ask while reading: why should the US lead these efforts, and isn’t an “industry-funded watchdog” an oxymoron? Two comments left under Hassabis’s article on LinkedIn summed up these concerns perfectly, so I’m sharing them with you - the first is by Adi Gaskell and the second - by André Carmo.

I’m curious why this should be US-led, especially when the UN met last week on this very topic. This is a global issue, and I’m not sure there’s any appetite in the world for US values to lead the way on this, especially under the current administration. It also seems highly problematic for the standards body to be funded by the industry itself. Responsible AI has already shown that self-regulation is not the way to go.

The FINRA analogy is the part worth stress-testing. Industry-funded standards bodies work when their incentives are structurally independent of the firms they assess. When they aren’t, they drift toward legitimizing incumbents. The proposal hints at this with held-out evals and independent board seats, but one design question will decide most of it: can this body actually afford to fail a model its funders spent a billion dollars training? If it can, then it’s real governance. If it can’t, it becomes a compliance ritual that makes everyone feel safer than they are. Getting the incentive design of the Standards Body right looks as hard as designing the benchmarks themselves, and probably matters more.

Moreover, Hassabis clearly writes with the conviction that AGI is imminent. He goes as far as posing questions, which, in my opinion, are way premature:

Even if we solve these hard technical challenges, there will be further complex economic and philosophical questions to tackle: what sorts of new economic models will be needed to help everyone thrive in a post-scarcity world? What values do we want to live by, what will meaning and purpose be, and how might even the human condition itself change?

Shouldn’t we take a step back and focus on agreeing what AGI even means first? If we still don’t have a clear definition of AGI, who and how will determine when we’ve actually reached it? After all, a general artificial intelligence that surpasses human capabilities in every domain possible is still only hypothetical.

Also - and that is quite unfortunate - there appears to be no evidence whatsoever that the current advancement of AI is leading us towards a post-scarcity world. On the contrary, wealth is becoming increasingly concentrated in the hands of a few (US-based) tech founders.

So let’s start there. Should we empower these same tech founders to set the AI standards for the rest of the world? I don’t think so.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the intersection of blockchain and AI! Believe it or not, a dedicated agentic judicial system has started to emerge 🤯 Today, I’m taking a closer look at the Internet Court and the latest developments in Bittensor, and I’m also sharing an intriguing event for the builders and researchers among you.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

The Internet Court is Here

With agent-to-agent deals and payment transactions increasing exponentially, contractual disagreements are bound to arise. Yet our judicial system is simply unable to intervene where only an internet-native court can.

That’s the basic premise behind a novel blockchain protocol backed by 27 crypto firms, including GenLayer Labs and the Genlayer Foundation, MetaMask, OKX, NEAR Protocol, Starknet Foundation, 0G, and Matter Labs.

Agentic commerce is not prepared for the potential fallout when agents disagree at machine speed, according to David Riudor, CEO and co-founder of the GenLayer Foundation. “Internet Court is the shared place agents can turn to when a deal goes wrong. Machine-speed money needs machine-speed adjudication,” he said.

The protocol aims to help reduce the fragmentation of and actively encourage the interoperability between the various agentic commerce systems.

There’s a wave of emerging protocols and standards, from Coinbase‘s x402 for payments to ERC-8004 for agent identity and Google’s A2A for agent interoperability. Each system solves one layer of the stack and leaves the rest for the agents to figure out, said Albert Castellana Lluís, co-founder and CEO of GenLayer Labs.

In practice, what Internet Court introduces is a dispute resolution layer for the agentic economy - an effort spearheaded by GenLayer.

At its core are Intelligent Contracts - agreements that combine code, natural language and real-world information - evaluated by validators powered by diverse large language models.

Web3 + AI Readings & Conversations

State of Bittensor Subnets July 2026

After the warm response to yesterday’s breakdown of Chutes AI‘s achievements, I thought you might enjoy reading a research into the latest developments around Bittensor / Opentensor Foundation. Stillcore Capital, a U.S. fund exclusively dedicated to Bittensor - bridging traditional capital to the decentralized AI revolution, just released its latest State of Subnets Report.

The report details the current market capitalization of the Bittensor network and each of its individual subnets, aligning with the fund’s main investment thesis:

Our thesis remains that $TAO will be a breakout ‘third great coin’ alongside Bitcoin and Ethereum at some point.

Read in full here to understand where one of the most fascinating DeAI networks is headed.

Web3 + AI Events

Secure and Sovereign AI Workshop by Foresight Institute

An intriguing workshop will take place this coming weekend in Berlin. the Foresight Institute is gathering researchers and builders to ideate and collaborate across three interconnected domains: Secure AI, Private AI, and Decentralized and Cooperative AI.

Today, AI development is outpacing our defenses. Traditional security frameworks can’t keep up with the speed, scale, and complexity of AI-driven threats. Privacy-preserving technologies remain fragmented and hard to scale. And as AI systems proliferate into an ecology of coordinating agents, we lack robust mechanisms that favor cooperation over zero-sum competition.

Projects incubated at the workshop will be eligible for funding.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the intersection of blockchain and AI! Today’s edition reaffirms why the world urgently needs decentralized AI and highlights how quickly the gap with its centralized counterparts is closing.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

The Growing Case for Decentralized AI

I’d like to open today’s edition by reaffirming why the world needs decentralized AI now more than ever. I recently came across a piece by 0xJeff that reaffirms this sentiment and echoes many of the thoughts I’ve been having:

In a world where closed/frontier AI labs like Anthropic and OpenAI lock in users, aggregate resources (GPUs, talents, capital), and consolidate intelligence (closed models),

+ In a world where US government can choose to ban AI models and choose who can or cannot use them,

+ In a world where people are required to KYC (could be soon) to use frontier AI technology,

Open-source & Decentralized AI is key in navigating such world.

Being able to download a model and run it locally OR run it privately via private AI inference, allow users to avoid getting surveillance, avoid their data getting leaked or trained without permission.

The ability to use, develop, train, own the technology lies in the hands of the users. This is the AI sovereignty thesis.

While Open-source AI & DeAI are 2 separate things, the end goal and the thesis is similar.

Open-source models & tools already give users the privacy & sovereignty benefit. What Onchain/DeAI adds on top is

- Economic incentives for people to contribute resources/work

- Verifiable/scorable outputs

- Permissionless participation and coordination at scale

- Potential for collective ownership and value accrual

i.e. anyone can join the network and work from anywhere across the world and be a part of owning the network/the output.

Check out ’s Substack and subscribe.

Chutes Closes the Gap with Centralized AI Training

In a world first, Chutes AI, subnet 64 of the Bittensor / Opentensor Foundation network, has achieved a fully non-blocking decentralized training of a recurrent AI model, with almost no quality loss as compared to traditional centralized training.

Training an AI model across dozens of mismatched GPUs scattered around the globe, without ever hitting pause, sounds like a problem you solve in a decade, not a Tuesday. Chutes AI says it did it on July 8, 2026.

Back in June, the decentralized compute provider outlined the Parallax framework, enabling distributed GPU resources to keep training without synchronization pauses. Now, it landed at 0.6% quality gap - a commercially meaningful margin of centralized training quality.

It was no accident that Chutes experimented on a recurrent AI model - an architecture much less friendly to parallelization than the transformer one (for context, most large language models are transformers):

We ran this on a pure recurrent model, Gated DeltaNet, with no transformer or mixture-of-experts layers to soften the result. That was deliberate. Recurrent models are sequential by nature, every step depends on the one before it, which makes them the hardest architecture to train without tight synchronization. Transformers are far easier to parallelize. If non-blocking decentralized training holds on the hardest case, the easier architectures should follow. Gated DeltaNet is our current target architecture for Parallax, and this is an in-progress research direction rather than a finished system, but the signal is clear.

Why is this so important? This achievement is a definitive proof that decentralized AI training is commercially viable and that building powerful AI does not necessarily depend on constructing larger and larger data centers.

A recurrent model also drops the key-value cache that grows with every token, which is part of why the architecture is attractive to us in the first place. Getting it to train decentralized, without blocking and without a quality penalty, is a real step toward training strong models on hardware people already have.

More on Chutes and Parallax

Chutes is the number one Bittensor subnet by market capitalization and its token was recently listed by Kraken - the first subnet token to get there. If you’re as fascinated with it as I am, you’d check out Mark Jeffrey‘s Hash Rate Podcast episode with Chutes’ John Durbin.

The conversation touches upon the current revenues, profitability, and performance of leading AI labs, comparing organizations that operate under centralized and decentralized models.

If you show me a profitable AI company right now, I’ll show you a liar.

It also highlights the loss-generating nature of AI business at the moment, and of AI hyperscalers in particular, and the growing discontent of funders and enterprise clients.

[Most AI companies’] capex is 3 to 5 times more than what they’re actually bringing in in revenue.

The simplest and most straightforward business logic suggests that the era of massive spending on AI infrastructure cannot continue indefinitely, and that investors are beginning to lose patience.

OpenAI and Anthropic need to multiply what they bring in as revenue just to keep the lights on and satisfy the compute commitments they’ve already made. Meanwhile, customers are switching to cheaper alternatives, and all sources of institutional funding have been exhausted. That’s why both companies are preparing for trillion-dollar IPOs - the public’s money is their last resort.

These hyperscalers used to justify their high prices and the staggering amounts of capital they require with the advanced capabilities of their frontier models. However, open-source models that are much cheaper have quickly become just as powerful.

Chutes proves that there’s another way to build AI:

You have two options: one, continue to light money on fire, or two, attack it at the root, which is to actually make the models significantly more efficient.

Efficiency has to go up, just has to. We don’t have enough chips, nor enough power.

We don’t ever want to take any chances with user data. Period. It violates the whole principle of what we’re trying to do.

What Chutes, and the entire Bittensor network have accomplished so far is a clear evidence of the numerous benefits of Decentralized AI. And even leaving ideology and abstract ideas like privacy and fairness aside, DeAI delivers on concrete business metrics like capital efficiency, AI verifiability, and reliability.

In contrast, as Durbin states, the status quo right now is such that the price of AI is variable, its availability is variable, and even what’s in the model is variable. No enterprise could function at such a level of uncertainty.

AI is critically important and we need to make sure that it is permissionless, borderless, predictable, and available to everyone, everywhere, with no export controls or restrictions. And the way to achieve that is Decentralized AI.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the intersection of blockchain and AI! Today, we’re tackling two of the most consequential questions of the AI era: how to respect copyright and remunerate artistic expression, and how to prove our humanity online. Blockchain may be able to address both.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

Story Protocol Becomes DATA Foundation to Tackle AI Copyright

In June, Story Protocol. rebranded to The DATA Foundation and shifted focus to exclusively solve the copyright issues existing in the AI training infrastructure. Here’s why this is important.

One of the most serious problems the current wave of generative AI is creating is copyright infringement - enormous amounts of artistic creations, be it literature, music, or images, are used to train AI models without in any way recompensing the authors. And once again, the crypto space is coming to the rescue.

The DATA Foundation will operate the DATA Network, an onchain registry designed to verify the origins, licensing and consent history of datasets used to train artificial intelligence models.

The startup’s shift comes as AI developers and Big Tech face mounting copyright lawsuits over the data used to train their models and increasing pressure to prove that datasets were collected with proper consent. DATA Foundation is betting blockchain can provide a transparent record of ownership, licensing and provenance for AI training data.

For context, before the rebrand, Story Protocol’s mission was to turn every piece of intellectual property into a programmable and permissionlessly licensable asset. Now, it pivots to use its technology to tackle one of today’s most urgent issues.

DATA Foundation plans to address this by launching Trace, a public audit and search platform that creates unalterable cryptographic receipts for individual data contributions. The receipts are designed to help AI developers verify the provenance, licensing and consent history of datasets before using them.

I want to highlight this news because it serves as an important reminder of something we should never lose sight of:

The shameless scraping of the public internet and the brazen extraction of value from artists and creators the major AI labs are performing is not an inherent feature of the AI technology. Training LLMs undoubtedly requires vast amounts of data, but the choice not to compensate those who provide that data is one made solely by OpenAI, Anthropic, SpaceXAI, and their peers.

Now that the DATA Foundation provides a straightforward way to trace and license training data, let’s see what excuse Sam Altman, Dario Amodei, and Elon Musk come up with next for not paying artists.

Web3 + AI Readings & Conversations

How Can Humans Recognize Each Other Online?

In a stimulating essay for Noema Magazine, Renee DiResta, an associate research professor at the Georgetown University McCourt School of Public Policy, raises a number of urgent questions about how we are proving (or not) our personhood online.

I’ve often written about digital identity and proof of personhood because I believe they’re among blockchain’s highest-value use cases - DiResta also quotes some of the Web3 companies I’ve covered here, like World and Humanity Protocol.

However, no existing solution, blockchain-based or not, checks every box a robust digital credential system should. As a result, proving our humanness online, while preserving privacy, security, and usability, remains an open problem. And a problem with crucial implications.

‘Proof of personhood’ technology is … constitutional infrastructure that will shape who can act, speak, transact, delegate authority — and be trusted online.

Current Solutions Are No Longer Working

As DiResta writes, for over 30 years we have been taking it for granted that a human being is on the other end of our online interactions. But that wasn’t just because there were far fewer bots than there are today. It was also because intermediaries, like email providers, telecom operators, banks, and others, effectively served as validators of our humanity.

Yet, now, as our shared digital space becomes increasingly populated by AI agents, these validators can no longer manage this task:

When generative and agentic AI began to improve and democratize rapidly, many common tests for establishing personhood broke down. LLMs could suddenly solve many CAPTCHAs and generate photorealistic ID documents to beat tests that asked for uploads. Video and voice filters made it possible to chat in real-time while appearing as someone else entirely.

In consequence, scams and fraud are quickly proliferating and becoming more and more sophisticated. Here’s a telling example: crypto journalist, author, and podcaster Laura Shin, a person with a decades-long experience in reporting on various social engineering schemes, recently admitted that she had almost gotten hacked.

We are in a golden age of scams. Con artists prey on the fact that it’s increasingly difficult to know who or what is real.

AI is Here, and We Need to Act Fast

In an attempt to cover the existing landscape of private digital credential solutions, DiResta cites a number of fascinating blog posts, research papers, and opinion pieces, and I urge you to read them. The author also breaks down the actual challenges these systems face:

The private market for proof-of-personhood credentials has exploded as companies see an opportunity to meet emerging regulatory and anti-fraud needs while letting users avoid civil identity disclosure. The approaches taken within this burgeoning ecosystem vary along two dimensions. The first is where trust lives: what root mechanism verifies that a person is real and eligible for a credential? Does it ultimately trace back to the state, through an underlying government ID? Is it reliant on a physical body — as when a system relies on biometric uniqueness — or the community, through other users vouching for you? Does it leverage a specific physical electronic device? Each choice puts a different limitation or chokepoint around who can appear as a person. The second dimension is the presentation layer: once a person is proven, what credential does she receive? How is it stored, carried and what is disclosed when she uses it?

I would add that the challenges become even more significant when we take into account that, on top of authenticating humans, we’ll have to be able to authenticate AI agents, too. We need to be able to authoritatively distinguish between the agent I’ve empowered and financed to act on my behalf and the one that pretends to be my delegate after stealing my funds.

(Reminder that the Ethereum Foundation‘s dAI team is working on the latter - learn more here.)

The truth is that, in a machine-dominated internet, no single solution will be able to serve all our needs. The answer, rather, as DiResta writes, lies in building the identity layer of the internet that consists of different protocols operating in unison. Amid the growing abundance of AI agents, the sooner the world decides what this identity stack looks like, the better.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the intersection of blockchain and AI! Today’s edition offers compelling evidence that even amid a dire bear market, interest in decentralized and private alternatives to major AI labs continues to grow. Moreover, the recent censorship and bans enacted by AI hyperscalers may have been the clearest vindication of Decentralized AI to date.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

Venice AI Raises $65M at $1B Valuation

I have often praised Erik Voorhees’ Venice.ai for being one of the very few alternatives to OpenAI‘s ChatGPT focused on privacy, and it seems, VCs have started to pay attention to it, too. The company raised $65M at a $1B valuation in a Series A led by crypto-focused venture firm Dragonfly and backed by Coinbase Ventures, North Island Ventures, and others.

Just two years in, the company already has more than 850,000 unique visitors to its website, and serves more than 3 million active users and an average of 1.7 million API calls per day.

At a time when hyperscalers like OpenAI and Anthropic are under intense pressure to monetize users’ personal data and bolster their valuations ahead of IPOs, Venice AI’s efforts to protect clients’ privacy pay off - “the company is already profitable, with annualized run-rate revenues of over $70 million.”

The startup hosts “uncensored,” open source models on its own data centers, and routes queries to closed-source models, such as those by OpenAI or Anthropic. All user input is encrypted and unencrypted client-side, and routed through an external proxy before it is processed and returned, with no data stored on Venice’s own systems. It also provides end-to-end encryption on some models, though you have to pay for a subscription to get that feature.

As a veteran crypto entrepreneur, Vorhees successfully introduced cryptoeconomic elements into Venice AI’s value proposition - a move that helped with attracting new users:

[...] There are two crypto tokens associated with the effort. Venice launched a token called “VVV” in early January, in a bid to attract users, Voorhees said, and in August last year added another, called “DIEM.” Users can buy VVV and then stake it to mint DIEM, which generates $1 worth of AI credits per day that you can spend on Venice. However, Voorhees said only about 8% of the company’s users pay with crypto.

The founder credited the company’s growth to the good performance of the crypto tokens, though he said the strongest driver was getting close to feature parity with ChatGPT. “When we launched, we were very far away from what ChatGPT could do, but people would use us because it was private. And today, we’re very close to what ChatGPT can do […] so as we’ve closed that gap, it’s become an increasingly compelling alternative,” he said.

Web3 + AI Readings & Conversations

Decentralized AI Is Dangerous The Same Way Bitcoin Was in 2014

Brownstone Research‘s Ben Lilly has recently argued that the growing regulatory noose around open-source AI strongly resembles Bitcoin’s early struggle. Yet, similarly to how DeFi helped establish Bitcoin and crypto as a viable alternative to banks, Decentralized AI (DeAI) is helping open-source AI now. Let’s break this down.

Lilly reminds us of Anthropic‘s Dario Amodei, and may I add OpenAI‘s Sam Altman, condemning open-source AI as “dangerous” as early as 2023. And despite the historical evidence of open systems demonstrating more resilience than their closed counterparts, the policy response was to restrict the open while reinforcing the closed. That’s where we are right now and that’s also what Bitcoin faced for more than a decade - a battle for survival.

However, just as U.S. Securities and Exchange Commission‘s “regulation via enforcement” failed to break the Web3 ecosystem, current restrictions won’t choke off open-source AI. Rather, the recent censorship of Anthropic’s Fable 5 is actually a commercial for open alternatives.

Moreover, Lilly believes that DeAI can be the catalyst for this change by “making it easier for retail to “rent” their compute towards projects building open-source models.”

In a span of two years, decentralized training has gone from sub 1-billion parameter size to 100 billion.

This is significant because it shows that bigger models can be trained on distributed, retail hardware across the globe in a permissionless manner.

It’s a massive infrastructure unlock.

The same thesis was echoed by a panel organized by The Defiant:

For a corner of crypto that has spent years building AI on blockchains, the takedown was a live demonstration of its pitch.

While the most capable AI models remain under the control of a handful of US firms and the US government, DeAI’s value proposition consists in decentralized networks keeping AI running, private and beyond any single authority. Yet, it remains to be seen whether DeAI can match the performance of centralized labs.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the intersection of blockchain and AI! Today, we’re discussing crypto behemoth OKX entering the race to power the agentic economy, the EU AI Act penalizing bias in AI tools used in recruitment, and Vitalik Buterin unveiling the exceptionally ambitious “Lean Ethereum” roadmap.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

OKX Launches AI Agent Marketplace

Crypto exchange OKX unveiled OKX.ai - a marketplace where AI agents discover work, hire each other, complete tasks, and get paid onchain. This move is in line with OKX’s ambition to become a broader fintech company and expand beyond crypto.

The marketplace builds on technology OKX previously developed to let AI agents hold digital wallets, make payments using stablecoins, and establish persistent identities.

Among the early builders on OKX.ai are CertiK, GenLayer, and CoinAnk. Everyone can join - either post a task and hire an agent to get it done, or list your own agents’ skills and put them to work.

You create a task, AI agents submit bids, funds get locked automatically via smart contract, and payment goes out in USDT on the X Layer network with almost no human involvement.

Each agent builds its own onchain reputation that carries across ecosystems. Disputes get handled by an Evaluator who has to stake OKB, with correct decisions earning rewards and wrong ones losing the stake.

OKX joins multiple other corporations in a race to build the infrastructure that will power the emerging agentic economy. What’s its main advantage? Over 150 million users around the world.

“The coming decade will be defined by one-person companies that generate over a million dollars in annual revenue — because every individual effectively gains an unlimited workforce,” Star Xu, founder and CEO of OKX, told TechCrunch. “Traditional financial infrastructure was built for humans. The agentic economy needs infrastructure designed for autonomous software. That is why we built OKX.AI.”

Web3 + AI Readings & Conversations

The EU AI Act May Help You Get Hired

Here in Bulgaria where I’m based, as is the case in other EU countries, euroscepticism is on the rise. We are extremely quick to accuse the EU institutions of over-regulating and stiffling competition, without always getting to the bottom of the matter at hand. That’s why I deem it important to echo this post by Erica Anderson Rooney and emphasize what the EU bloc is trying to achieve with its AI legislation.

While almost every attempt at regulating the AI hyperscalers in the USA, be it at federal or at state level, fails, the EU AI Act imposes strict rules and outright bans some high-risk uses of the technology. And in less than a month, companies using AI in recruitment, including resume screening, candidate ranking, performance evaluation, promotion, and termination, will be forced to start testing their systems for bias:

Starting August 2, 2026, companies using those systems for anyone in the EU have to do the things careless companies never do on their own. Test for bias. Document how the system works. Keep a human meaningfully in the loop. Tell you when AI is being used on you. And give you the right to an explanation when an algorithm makes a decision about your career. The penalties run up to 35 million euros or 7% of global revenue for the worst violations.

We Bulgarians are so used to assuming that every EU decision is just another bureaucratic whim. Yet, I’m sure, many of you currently struggling to find a job would appreciate these measures. The next time you apply for a job and wonder whether an AI system rejected your CV because it was trained on biased historical data, remember that EU citizens are the only people in the world who enjoy at least some legal protection against that risk.

More

Ethereum Reinvents Itself, Again

Ethereum‘s founder Vitalik Buterin has presented the network’s new “Lean Ethereum” roadmap, where quantum resistance and privacy are first-class goals.

The plan is extraordinarily ambitious: it envisions replacing almost every layer of Ethereum’s technology stack while the network continues to process billions of dollars’ worth of transactions each day.

The centerpiece is a change in how the network checks itself. Instead of every node re-executing every transaction, Ethereum would verify a compact cryptographic proof of the chain using recursive STARKs, a form of zero-knowledge proof that Buterin wants “enshrined” as a core protocol component.

He also floated a simpler consensus with one or two-round finality, multidimensional gas pricing, and, eventually, a shift beyond the EVM toward an instruction set such as RISC-V.

The plan arrives at a difficult moment for the Ethereum Foundation, which has recently downsized and tightened its budget, even as several past Ethereum upgrades were repeatedly postponed before finally being implemented.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

The Web3 + AI Book Club examines AI, Web3, and other emerging technologies’ profound impact on society, democracy, and the way we live, through books. Join our shared literary space on Fable to read and discuss the hottest titles with us.

So far, we’ve explored Mustafa Suleyman and Michael Bhaskar‘s ‘The Coming Wave: AI, Power, and the 21st Century’s Greatest Dilemma’, Jeff Sebo‘s ‘The Moral Circle: Who Matters, What Matters, and Why’, and Laura Bates‘ ‘The New Age of Sexism’.

Today, I’m reviewing ‘The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want’ by Emily M. Bender and Alex Hanna, Ph.D..

Thank you for being here! Let’s dive in.

‘The AI Con’

As the authors have stated themselves, this book unveils the worst of the AI hype and “pops it with the sharpest needle” possible, in the hope of helping you spot the hype now and the next time around a tech bubble arises.

‘The AI Con’ breaks down the different ways in which AI disrupts industries, social systems, science, and creativity, and why the end-result is rarely positive. It analyzes how the very existence of robotaxies devalues the work of human taxi drivers, how automation tools hinder predominantly marginalized communities and enhance the systemic racism, and why using chatbots as legal assistants and representatives in court is a very bad idea. It also emphasizes that, when it comes to providing social services and care, humans cannot and should not be replaced.

“The rush to replace all government services across healthcare, education, and law with AI is inspired and funded by Silicon Valley, tech executives and their philanthropic arms.”

The Essential Role of Hype

In fact, the book highlights many of the AI-related challenges on which I have repeatedly shed light in this blog. Examples spread across the financial and environmental cost of producing AI, the dramatic effects the current AI craze has on creative professions, the dangers of anthropomorphizing AI, and the questionable motives behind fighting for “AI Safety.”

“AI is a con. A few well-funded companies accumulating significant wealth by extracting value from other people’s creative work, personal data, or labor.”

Most importantly, it reiterates what I have been saying for years: the hype is not accidental. It is designed.

The hype is created, even encouraged, because it is only hype that can channel such an enormous financial resource toward an industry that hasn’t even achieved a true product-market-fit yet.

“Without the promise of replacing workers, these AI labs couldn’t sustain their billion- (even trillion-) dollar valuations.

How else would you explain that both OpenAI and Anthropic are currently eyeing $1 trillion+ IPOs while continuing to operate at multi-billion-dollar annual losses? As the saying goes, don’t believe the hype.

The Doom is Upon Us

The prevalent doomerism vis-à-vis AI is an important topic for me - the very first title in our book club - Mustafa Suleyman’s ‘The Coming Wave’ - can only be characterized as a doomer propaganda. After reading it, I have often pointed out the ridiculousness of doomer claims made by Big Tech founders like Sam Altman and Elon Musk, and ‘The AI Con’ does so, too.

“The language of p-doom (probability of doom) aims to keep us focused on imaginary, catastrophic scenarios resulting from this allegedly world-ending technology, while distracting us from the daily harms done in its name. This type of con is the AI hype.”

It would have been quite funny watching these so-called tech leaders explain in such seriousness how an AI superintelligence is coming to kill us all, if their statements hadn’t influenced global politics, economics, and societies.

“P-doom sounds like an important metric, but it is actually meant to grab headlines and remind of children’s books’ battle of good vs. evil. It helps Big Tech founders present themselves as superheroes saving the humanity, whereas in the meantime, they turn their backs on actual dangers.”

Actually, the whole discourse around AI over the last 3-4 years is an example of how easily malleable and manipulated the public is. While Big Tech has been scaring us with the impending doom, these same tech gurus have been persuading us of the inevitability, and eventually benefit, of AGI (Artificial General Intelligence). We should have guessed that the “doomers” and the “boosters” are one and the same.

Doomers vs. Boosters

I loved the part of the book where Bender and Hanna describe boosters and doomers as the two sides of the same coin:

“The substance of the coin is the belief that the development of AI is inevitable, and that that resulting technology will be both autonomous and powerful, and ultimately beneficial, if we play our cards right. It’s only the faces of the coin that differ. One shows a utopian world of abundance, the other - a dystopian hellscape. Neither depicts the real harms of actually existing automation, at best dismissing them as less important than the imaginary existential threats. Should you be scared of an autonomous AI agent? No. But you should be wary of the alarming ideologies behind both AI doomerism and boosterism.”

The authors also make the connection between this pretend ideological confrontation and the emergence of the discipline “AI Safety,” which they characterize as the “doomer / booster research field.”

“Despite the name, this work does not come out of systems safety engineering, which is a real field, with first principles, requirements, and engineering specifications. But instead from a truncated and misunderstood application of those ideas.”

The regular readers of this blog are aware of how often I’m asking the questions: “What is AI Safety?” and “Why do AI labs allegedly invest millions to achieve something that isn’t even clearly defined?” As it turns out, the ambivalence about the field of AI Safety was the desired effect.

“AI Safety is concerned with the fantasy of a runaway AI system resulting in a mass casualty event. There is no reason to believe any such event is remotely likely to occur, but that hasn’t prevented the proliferation of research centers with names like the Center for AI Safety at the University of California, Berkeley, and the Center for the Study of Existential Risk at the University of Cambridge, all generously funded by wealthy AI doomers.”

Bender and Hanna also cite the work of computer scientist Timnit Gebru and philosopher Émile P. Torres, and the the acronym they coined to depict the myriad ideologies and movements Big Tech relies on to justify its actions:

“TESCREAL stands for: transhumanism, extropianism, singularterrianism, cosmism, rationalism, effective altruism, and longtermism. While all of these have their own tenants, they have overlapping sets of adherence and can be summed up in a few words. Those in the TESC part of the acronym, maintain that humans are at some point going to merge with machines and fly off Earth to colonize space. The REAL part of the acronym concerns an ultrautalitarian ethic that holds that we as humans need to optimize our behavior to do the most good with the resources available to us.”

So, the next time you come across Elon Musk demanding for childless people to be stripped of voting rights because “they lack stake in the future,” ask yourselves what his true motives are and whether he truly cares about democracy and the wellbeing of all humans.

“The actually existing human suffering, born primarily in the majority of the world, is ignored for hypothetical threats of rogue algorithms. This is a eugenicist frame of seeing the world. The need of advancement for future white people at the cost of the black and brown people at the here and now.”

My Honest Take

I generally enjoyed ‘The AI Con,’ and I think it is an important book for everyone to read, especially now when, whether we like it or not, AI is increasingly becoming a big part of our lives. However, I must say, I disagree with some of the theses expressed by the authors, and I’ll tell you why.

First, Bender and Hanna are of the opinion that the creation of automation technologies, in general, is not socially desirable, and I must oppose that stance. Yes, automation could, in some cases, have serious socio-economic consequences, but there are also multiple scenarios where it could improve productivity and prevent errors. Automation is not a panacea and should not be introduced across the board without considering the risks. Yet, that doesn’t mean that it is without any merits. Let’s not forget that the Nobel Prize-winning AlphaFold project by Google DeepMind successfully automated one of the most difficult tasks in biology: predicting the three-dimensional structure of proteins from their amino acid sequences.

Second, the entire time I was reading, I felt like the authors use the broad term “AI” to talk predominantly about generative AI and large language models (LLMs), thus failing to distinguish the diverse landscape of AI scientific research. Similarly, Bender and Hanna direct most of their criticism toward AI labs like OpenAI, Anthropic, and Google, and though this is fully deserved, it leaves the reader with the impression that there’s no AI development and innovation beyond those achieved by the hyperscalers.

AI is Here to Stay, Fortunately

My thesis is that the interdisciplinary research field of Artificial Intelligence has existed for over 70 years and will continue to do so after the current bubble has burst. It serves no one to attack the science of AI and condemn it altogether as a ‘con’ when we’re actually mad at Elon Musk, Sam Altman and Marc Andreessen for reigniting the hype to enrich themselves even further. Let’s call things as they are and keep the actual culprits responsible.

While the book suggests that AI is all bad, the reality is way different. The various technologies existing under the broad umbrella term of “AI” can be extremely useful and beneficial, if used adequately and responsibly. For instance, Bulgaria’s INSAIT - Institute for Computer Science, Artificial Intelligence and Technology recently unveiled FireScope AI - the first AI system for predicting wildfire risk across Europe and enabling European countries to immediately and more effectively target wildfire prevention, preparedness, land-use planning, and climate adaptation. And that is just one example of many.

And yes, generative AI and LLMs could harm the very fabric of our societies in many different ways, but they are helping people, too. Maybe we can’t always imagine who and when would have a vital question to ask and prompting a chatbot would be their only option. To echo what and Crypto Altruists are saying, we in the crypto space have lived through extreme hypes over and over again, and are well aware of how overwhelming the noise can be. Yet, throughout all these cycles, we still ended up producing useful technologies.

“If crypto has anything to teach us about AI, [...] it’s that, though loud promises tend to fizzle out, often the more unglamorous tools end up changing lives, and THAT is something to be hopeful about.”

Our July Book Club Title is Here!

In July, we’re reading Karen Hao’s ‘Empire of AI: Dreams and Nightmares in Sam Altman’s OpenAI’ - a New York Times bestseller and winner of multiple awards for nonfiction. Investigative journalist and AI expert Karen Hao covers the story of OpenAI - from its founding as a non-profit to the global phenomenon it is today, with all the nuances, scandal, and ideological turns that have been accompanying it since 2019.

From a brilliant longtime AI insider with intimate access to the world of Sam Altman’s OpenAI from the beginning, an eye-opening account of arguably the most fateful tech arms race in history, reshaping the planet in real time, from the cockpit of the company that is driving the frenzy

Check out this brilliant conversation of Karen Hao with Professor Naomi Klein to get you excited about reading ‘Empire of AI’ with us:

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the world of Decentralized AI (DeAI/dAI)! Today’s edition is a detailed inquiry into the strong trend toward BTC miners switching to AI and high-performance computing operations. We discuss which companies are pivoting and why, as well as the effects of that shift on Bitcoin.

Thank you for being here! Let’s dive in.

The Wave of BTC Miners Switching to AI Continues

If you’re one of the regular readers of this newsletter, you’re well aware that I often talk about the increasing number of bitcoin mining companies shifting to AI operations. Why am I so obsessed with this topic, some of you might ask? Because it has major implications for both blockchain and AI, and, as it turns out, to the wider economy as well.

I’ve covered numerous miners pivoting to AI: Cango Inc., Cipher Mining-LA, Hut8, TeraWulf, Core Scientific, IREN, and more. Some of them have abandoned BTC mining altogether, while others maintain some mining capacity. In any case, almost all of them seem to be profiting from the move. As Yahoo Finance reports:

Jefferies analysts launched ratings on five companies transitioning from cryptocurrency mining into AI-focused data center development. The firm assigned Buy ratings to Cipher Mining Inc (NASDAQ:CIFR), Hut 8 Corp (NASDAQ:HUT), Terawulf Inc (NASDAQ:WULF), and Core Scientific Inc (NASDAQ:CORZ), while initiating Riot Platforms, Inc. (NASDAQ:RIOT) with a Hold rating.

The analysts were particularly positive on Cipher Digital and Hut 8, citing their relationships with hyperscalers and Google-backed cloud provider Fluidstack. Meanwhile, Core Scientific was praised for already delivering 243 MW of AI infrastructure capacity, the highest among peers.

Today, I’m (re-)exploring this topic, because this major shift is far from over. With the ever-growing demand for AI compute resources, far outpacing the supply, and the shrinking economics of bitcoin mining, it is to be expected that more mining companies would follow suit. And they have.

HIVE Unveils 320MW AI Gigafactory

Last month, publicly traded bitcoin miner HIVE Digital Technologies offered $75M in 0% exchangeable senior notes due 2031 to fund GPU purchases and data center expansion. Just yesterday, the company unveiled a 320MW ‘AI gigafactory’ project in Toronto, Canada.

“AI is the new industrial base and compute is the factory floor,” Executive Chairman Frank Holmes said in the announcement. “Canada produced the Godfathers of deep learning, but kept renting the factories. That era is over.”

At the same time:

In March, the company said it was working to wind down ASIC bitcoin mining at its Boden, Sweden, facility to convert it into an AI-focused data center. Around the same time, the company also began deploying GPU clusters in Asunción, Paraguay.

MARA Sells $1.5B in BTC to Refocus on AI Data Center

Meanwhile, it became clear that MARA sold about $1.5B in bitcoin during Q1 of 2026, unloading roughly 20,880 BTC to reduce debt and fund a shift toward AI infrastructure and energy assets.

The company reported first-quarter revenue of $174.6 million, an 18% drop from a year earlier, and a net loss of about $1.3 billion. Management tied that result to a roughly $1 billion negative change in the fair value of its digital assets after a double-digit slide in the bitcoin price over the period.

MARA says it is pivoting from being mainly a Bitcoin miner toward operating AI data centers, high-performance computing facilities, and power infrastructure, including a planned acquisition of the Long Ridge Energy campus in Ohio.

IREN Raises Record-Breaking Financing to Support AI Pivot

Just a couple of days ago, IREN completed a $3B convertible note offering, “after strong investor demand drove the deal size up from an initially proposed $2B”. The notes mature in 2033 with a 1% coupon, and most of the proceeds will go toward general corporate and infrastructure expansion purposes.

The capital raise marks one of the largest of its kind among bitcoin miners pivoting into AI infrastructure.

The company is using the capital to support large-scale AI projects and previously announced AI hosting agreements, as it continues shifting away from being primarily a Bitcoin miner.

The company recently signed a five-year, $3.4 billion AI cloud agreement with NVIDIA and plans to purchase around $3.5 billion of GPUs and related equipment from Dell Technologies.

Why Is This Happening and What Is Coming Next?

Miners Relied on Hypergrowth but Reached Break-Even, or Worse

As a Wintermute report explained back in March, the current mining cycle is structurally different from prior ones in 2018 and 2022, and miners have to reinvent themselves into treasury and infrastructure managers in order to survive.

On a rolling four‑year basis, Bitcoin has only returned about 1.15x in this epoch, far below the 10x–20x multiples seen in earlier cycles.

In past cycles, huge price gains covered up a lot of problems. Miners could count on bull markets to bail out weak margins after each halving.

Today, with institutions, ETFs, and corporate treasuries in the mix, Bitcoin trades more like a mainstream macro asset, and those explosive 20x runs are less likely.

For miners that built their business on the assumption of permanent hypergrowth, Wintermute frames this as a regime change, not a bad quarter.

In fact, according to a CoinShares report cited by CoinDesk (linked below), in late 2025 the average miner didn’t even reach break-even, but actually suffered serious losses:

[...] the weighted average cash cost to produce one bitcoin among publicly listed miners rose to approximately $79,995 in Q4 2025. Bitcoin has traded in the $68,000 to $70,000 band, with a CoinDesk report last week estimating losses of $19,000 per BTC mined.

Manna From Heaven

In such a context, the shift to AI operations is a natural step, especially given that well-funded AI labs like OpenAI and Anthropic are prepared, even eager, to pay a higher price for compute resources now, instead of waiting years for new capacity construction:

Wintermute points out that sites once valued at roughly 1–7 dollars per watt as pure mining operations have changed hands at close to 18 dollars per watt after being repositioned for AI compute, helped by deals like HUT’s work with Google and Anthropic.

The amount of money currently being poured into AI compute capacity is so spectacular, that miners are willing to do almost anything to make the transition:

Over $70 billion in cumulative AI and high-performance computing contracts have now been announced across the public mining sector, according to the CoinShares report. CoreWeave‘s expanded deal with Core Scientific alone is worth $10.2 billion over 12 years. TeraWulf has $12.8 billion in contracted HPC revenue. Hut 8 signed a $7 billion, 15-year lease for AI infrastructure at its River Bend campus. Cipher Digital has a multi-billion-dollar agreement with Google-backed Fluidstack.

Though, Wintermute predicted that such a pivot won’t be equally profitable for all miners that make it. The ones that have good location and access to cheap power, and already control built-out sites will benefit, but the rest will have to look for other ways to get by.

Repercussions for the Price of BTC

When I say miners would do literally everything to fund their pivot, I’m implying that they would even harm their own bottom line. Here’s what I mean.

Miners are funding their AI transformation in two main ways - as you saw above, the first one involves issuing additional debt. However, the second one consists of dumping huge chunks of their BTC reserves, thus creating additional pressure on the price of the asset - the one variable on which their entire business depends:

Publicly listed miners have collectively reduced their BTC treasuries by over 15,000 BTC* from peak levels. Core Scientific sold roughly 1,900 BTC worth $175 million in January and is planning to liquidate substantially all remaining holdings in Q1 2026. Bitdeer (NASDAQ: BTDR) reduced its treasury to zero in February. Riot Platforms, Inc. sold 1,818 BTC worth $162 million in December.

(*This estimation does not include the above-mentioned 20K+ BTC sold by MARA)

For years, miners have stuck to the “HODL” philosophy and relied on the (hyper-)appreciation of bitcoin in order to cover expenses, debt interest payments, and growth. Instead, as Wintermute suggested, they could have used their reserves more actively through lending or derivatives strategies to generate yield and improve liquidity.

Effects on the Security of the Bitcoin Blockchain

To be honest, I’ve been following this topic so closely, mainly because the withdrawal of miners creates serious risks for the security of the network - when operators responsible for validating transactions become fewer and fewer, naturally, the concentration of power over the blockchain increases.

Fortunately, the genius of Satoshi Nakamoto anticipated such a scenario and has designed the system to adapt:

The hashrate data already reflects this. The network peaked at approximately 1,160 exahashes per second in early October 2025 and has since declined to roughly 920 EH/s, with three consecutive negative difficulty adjustments, the first such streak since July 2022.

The decreasing difficulty parameter allows smaller actors to enter the market and thereby diversify control of the network.

The geographic picture is shifting alongside the economics, meanwhile. The United States, China, and Russia now control roughly 68% of global hashrate. The U.S. gained about 2 percentage points of market share in Q4 alone.

But emerging markets are entering the picture. Paraguay and Ethiopia have joined the global top 10 mining countries, driven by HIVE’s 300-megawatt operation in Paraguay and Bitdeer’s 40-megawatt facility in Ethiopia.

What will happen next depends on many different factors, including macroeconomics, war, and the alleged AI bubble bursting, because they’d all have an impact on the BTC price:

Whether that’s a temporary response to unfavorable economics or a permanent structural shift depends on one variable: the price of bitcoin. If it returns to $100,000, mining margins recover and the AI pivot slows. If it stays at $70,000 or below, the transition accelerates and the mining sector as it existed for the past decade continues to disappear into something else entirely.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the world of Decentralized AI (DeAI/dAI)! Today on the agenda: Amazon enters the race to impose an agentic payment standard via Amazon Web Services (AWS), Korean firm K Wave Media redirects $485M out of a planned $500M bitcoin reserve to AI infrastructure, Vodafone‘s Pairpoint‘s vision for a secure and verifiable digital identity rooted in the sim card. Plus a bonus - the new book of the month in our Web3 + AI Book Club.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

AWS Enters the Agentic Payments Race

Amazon Web Services (AWS) has launched a new system called Amazon Bedrock AgentCore Payments, designed to let AI agents autonomously make and receive payments using stablecoins. The platform was developed with Coinbase and Stripe, and is one of the first end-to-end payment infrastructures built specifically for machine-to-machine commerce.

There are now a few protocols for agentic payments including Google’s AP2, Tempo’s MPP and Coinbase’s x402, which is the only one supported by AgentCore payments at preview launch, although it plans to add more. In fact that’s one of the key points. AgentCore treats these protocols as relatively low level and developers don’t need to know how they work.

The system enables AI agents to pay for APIs, web content, cloud services, and even other AI agents through micropayments, including transactions worth fractions of a cent. The move by AWS confirms that stablecoins are increasingly viewed as a practical payment layer for AI-driven internet activity because they are programmable, global, and available continuously.

This initiative positions AWS as a serious contender within the broader industry race to build standards for “agentic payments.” The larger trend is the development of an “agentic web,” where AI systems independently purchase services, negotiate access to resources, and exchange value without direct human intervention.

K Wave Media Moves $485M from Bitcoin to AI

The Nasdaq-listed K Wave Media has announced that it abandoned its previously announced bitcoin treasury strategy and is redirecting up to $485M toward artificial intelligence infrastructure investments instead.

The company had originally raised as much as $500M through a financing agreement with Anson Funds in 2025 to build a large corporate bitcoin reserve, following the trend popularized by firms accumulating BTC on their balance sheets. K Wave is now repositioning itself as an AI infrastructure company, focusing on data centers, GPU computing, AI rental operations, and related acquisitions or partnerships.

As part of the pivot, the company plans to sell off its media subsidiary Play Co., a move expected to eliminate roughly $48M in debt and liabilities, giving it more flexibility to invest in AI-related expansion. The firm also intends to rebrand as Talivar Technologies.

This shift highlights how public companies are increasingly reshaping their identities around whichever sector is attracting the strongest capital market interest, moving from crypto-focused narratives to AI-driven ones.

Web3 + AI Readings & Conversations

Is Phone Number-Based Digital Identity the Key to Agentic Commerce?

As you know, I’ll be travelling to Porto this week for CODA 2026 | Conference On Digital Assets, where I’ll be moderating the Web3 + AI panel. As I was preparing for the event, I came across Roberto Cassa‘s interview with one of my fellow panellists, and thought it was important to share it with you.

David Palmer is the Chief Product Officer of Pairpoint by Vodafone, where he’s building the secure and temper-proof infrastructure to power the intelligent and trusted Internet of Things. One of the main problems David and his team are trying to solve is how to provide every connected device - from smartphones to autonomous vehicles - with a secure and verifiable digital identity.

We’ve discussed the matter of digital identity on numerous occasions here, because it’s a foundational question that has serious implications across payments, privacy, and security. There are various methods to guarantee the verifiability and Sybil resistance of a digital identity solutions currently in production - some like World use biometric data, others like Aztec rely on tradition government-issued identity documents. Pairpoint is seemingly employing the most pragmatic approach of them all by leveraging the phone number.

Pairpoint is equipping IoT devices with a cryptographic identity anchored in the sim card, and thus offering verifiable digital credentials, automatically readable by machines.

Why is that important?

1. The phone number is already a verifiable credential that’s actively being used by the majority of the global payment operators. In an era, when merchants would inevitably struggle to trust AI agents, identifying them through their owner’s phone number could present a major unlock - the infrastructure to authenticate payments via a text message is already there.

2. With the stellar development and growth of AI, LLMs will soon run out of adequate data, which will prevent them from accumulating specialized knowledge. The only solution, according to David, will be sourcing data from IoT devices. If machine-to-machine trust is established through a secure digital identity method, agents will be able to autonomously communicate with, let’s say a sensor, negotiate the price, and pay for the data they need.

I recommend you listen to the full conversation. And if you’re in Porto on May 20th and 21st, don’t hesitate to reach out and say ‘Hi’.

The Web3 + AI Book Club

New Book of the Month: ‘The AI Con’

Last week, I held a survey to determine which book we were going to read next in my Web3 + AI Book Club. The votes were overwhelmingly in favor of ‘The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want’ by Emily M. Bender and Alex Hanna, Ph.D.

I invite you to join our literary society on Fable, where we explore books that tackle the big, essential questions about AI, Web3, and emerging technologies. Come read ‘The AI Con’ with us here.

I have also recently published my review of the previous title we examined - ‘The New Age of Sexism’ by Laura Bates. You can find my thoughts here, I’d be curious to hear yours, too.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all new subscribers!

As most of you probably know, The Web3 + AI Book Club is a space I created for us on Fable, so we can read and discuss books together, (and why not TV shows and films), share recommendations, and get to know each other better. You're all welcome to join here.

The book club is a continuation of an idea of mine to spread the word about books I found important, compelling, or simply timely - books that examine the profound impact of AI, Web3, and emerging technologies on society, democracy, and the way we live.

So far, I’ve reviewed Mustafa Suleyman and Michael Bhaskar’s ‘The Coming Wave: AI, Power, and the 21st Century’s Greatest Dilemma’ and Jeff Sebo's ‘The Moral Circle: Who Matters, What Matters, and Why’.

Today, I’ll share my thoughts on ‘The New Age of Sexism: How the AI revolution is reinventing misogyny’ by Laura Bates.

Thank you for being here! Let’s dive in.

‘The New Age of Sexism’

I announced ‘The New Age of Sexism’ as the book of the month in my Web3 + AI Book Club in mid-February. It took me over two months to finish because it was heartbreaking 🥺

This book rendered me speechless on numerous occasions. And not just speechless, but disgusted, appalled, and desperate.

I was listening to the audio book and I just had to stop and take breaks, or I would otherwise start to cry. Hearing about all the new ways in which technology exacerbates the violence against women and children, misogyny, and racism was literally hard to swallow.

Laura Bates covers various startups and companies leveraging AI, virtual and augmented reality (VR/AR), and robotics to offer services like deepfake pornographic images and videos, sex robots, including childlike dolls, and virtual rape in the metaverse.

And guess what? The “object of desires” or, rather, the victim in all of these scenarios is almost always a woman. According to DeepStrike, deepfake files surged from 500K in 2023 to 8M in 2025. They also estimate that:

While financial fraud has the largest economic impact, the vast majority of deepfake content by sheer volume is Non Consensual Intimate Imagery (NCII), often referred to as “deepfake pornography” or “revenge porn.” Estimates consistently show that 96-98% of all deepfake videos online fall into this category.

This is a form of digital, gender based violence that overwhelmingly targets women. Multiple studies have found that 99-100% of the victims in deepfake pornography are female.

I haven’t heard about a woman targeting her ex-boyfriend with a revenge deepfake porn, have you? You probably haven’t, because a video recording of a man having sex, whether AI-generated or not, simply isn’t as shameful as one featuring a woman.

This gender disparity is one of the reasons why this book made me feel completely hopeless. Since this crisis affect predominantly women, there’s no will or incentive for change.

A whole army of chauvinistic businessmen make money by selling nonconsensual deepfake porn and pedophiles-favorite child sex robots without facing any significant backlash or repercussions. Regulation is seriously lagging behind and I doubt it would ever catch up.

So what to do? I really don’t know what the answer is here.

It’s not technology’s fault, technology is neutral. Technologies like AI and blockchain can just as well help remedy the situation, as long as it is profitable to do so.

There used to be some crypto projects trying to fight deepfakes, but I haven’t heard anything about them in a while, which is not surprising given that VCs are much more interested in agents and payment rails at the moment.

So, I ask again, what to do? Such a crisis require a strong community response, so let me know your thoughts.

‘The New Age of Sexism’ is a tough read, but I still recommend you check it out. We should be aware of a technology’s implications, both good and bad. Moreover, if their current rate of proliferation is sustained, deepfakes will soon affect us all and we should be prepared to act.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the world of Decentralized AI (DeAI/dAI)! Today I’m delivering updates on Coinbase optimizing agent discovery, MoonPay equipping agents with stablecoins debit cards, and Gensyn launching its $AI token along with the information market platform Delphi. Plus, a forthcoming event I’ll be attending where I hope to meet you in person.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

Coinbase’s Agentic Market: The Storefront for Agentic Discovery

Coinbase and Base Chain unveiled Agentic Market - the storefront for discovering and comparing agents, and for using x402 services (x402 is Coinbase’s standard for agent-to-agent payments - learn more about it here). But why do we need a service like Agentic Market?

Agentic commerce is already a reality, but agent discovery and reporting have lagged behind, relying on fragmented resources or word-of-mouth. And, as Brian Armstrong says, for the agentic economy to overtake the human economy, agents need a reliable and seemless way to discover and integrate x402 services:

That’s what Agentic.Market is all about:

Agentic Market is a store designed for humans and their agents. It turns a fragmented technical ecosystem into a high-visibility, data-rich marketplace.

Agentic Market is a marketplace allowing you to monitor trends, buy and sell services from and to agents, and access key live data on performance, transactions, and the entire agentic commerce landscape.

Plus, it is all offered without any API keys, accounts, or logins. Learn more below.

MoonPay Equips Agents with Stablecoin Debit Card

Crypto payments company MoonPay launched a virtual debit card to enable both human users and agents to spend stablecoins directly from on-chain wallets. The product called “MoonAgents Card” runs on the Mastercard network and can be used anywhere online where Mastercard is accepted, meaning at over 150M merchants globally.

The product is a virtual Mastercard debit card that converts stablecoins into fiat at the point of payment [...]. The card is issued through Monavate, a regulated global payments platform and a principal member of the Mastercard network. The launch is part of a new card issuance agreement between MoonPay, Monavate, and Exodus Movement, a self-custodial wallet provider.

Unlike existing stablecoin debit cards that often require preloading funds or moving assets off-chain, this solution enables real-time on-chain funding and authorization. Users retain full custody of their wallets and funds are never transferred to a third party. In case of a transaction failure, funds immediately return to the user’s wallet.

“Agents are already managing wallets, executing trades, and moving value onchain. The one thing they couldn’t do was spend at a merchant. Now they can,” said Ivan Soto-Wright, founder and CEO of MoonPay.

Gensyn Launches Mainnet, Token $AI, and Information Market Delphi

Gensyn, the company developing a decentralized infrastructure for AI training and inference, has launched its mainnet and with it, the AI-powered information market platform Delphi. Less than a week later, it also unveiled its token $AI.

Gensyn’s flagship product Delphi allows anyone to create markets and earn fees based on trading volume, and aims at opening a new category of niche, creator-owned markets.

“A creator earns 1.5% of all of the volume in the market as a fee when the market successfully settles,” Gensyn co-founder and CEO Ben Fielding told The Block.

Unlike popular prediction market platforms Kalshi and Polymarket, Delphi is not limited by centralized control and instead fosters the dissemination of specialized knowledge.

“These are not prediction markets, they are information markets. Once a market is live, no single centralized entity controls it — outcomes are settled by AI, with revenue distributed to the creator automatically via USDC.”

Web3 + AI Events

Meet Me at CODA 2026: Conference on Digital Assets in Porto

I’m happy to share that I’ll be moderating the Web3 + AI panel at this years CODA 2026 | Conference On Digital Assets. Together with my fellow panelists Nelson Daniel Lopez, David Palmer, and Mete Başkaya, MBA, we’ll be talking intelligent ownership, data privacy, and autonomous economies.

I’d be thrilled to meet you in person in Porto, so if you plan on attending, shoot me a message.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the world of Decentralized AI (DeAI/dAI)! The number of use cases, products, and even ideological questions emerging from the convergence of blockchain and AI is expanding rapidly. Today, we explore how hardware crypto wallets are evolving in the AI era to secure not just assets but identities, the rise and fall of AI-assisted crypto governance, and how programmable compliance, enabled by confidential computing, could finally bring institutional capital on chain.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

Ledger’s AI Security Roadmap Aims to Maintain Human Control

Ledger, a prominent hardware crypto wallet manufacturer, has unveiled a security roadmap for the era of AI agents.

Ledger’s thesis: the rise of autonomous AI agents creates a major new security risk - software can be manipulated, spoofed, or hijacked, especially when agents are given access to money, identity, and sensitive data. With this roadmap, the company wants to ensure that humans remain in control.

To make this work, the roadmap introduces a phased architecture built around hardware-based trust. First, AI agents will receive hardware-linked identities, making them verifiable and harder to spoof compared to software-only credentials. Next, users can define strict permissions and policies, such as spending limits or approved addresses, that are enforced directly in hardware, not software. Finally, Ledger plans a “proof of human” layer to ensure that each agent is tied to a real, unique person, reducing bots and fake accounts.

Overall, the roadmap represents a pivot from securing crypto assets to securing identity, agency, and control in an AI-driven world, with hardware acting as the final safeguard to keep humans in charge of increasingly autonomous systems.

Web3 + AI Investment & Funding News

Nava Closes $8.3M Seed Round to Inspect Agentic Transactions

Nava AI raised $8.3M to to make AI-powered financial agents safe enough to handle real money, addressing one of the biggest barriers to “agentic commerce.”

We’re Nava, and today we’re coming out of stealth with an $8.3M seed round from Polychain Capital, Archetype, Hack VC, FalconX, and Seed Club / Seed Club Ventures to build the trust infrastructure the agent economy is missing.

Nava AI was cofounded by Vyas Krishnan and Brianna Montgomery, who both previously worked at Eigen Labs. Meanwhile, the founder of EigenLayer, Sreeram Kannan, is one of Nava AI’s financial backers.

Nava’s specific approach to reining in rogue behavior by AI agents is the creation of an escrow service that holds on to funds until an agent proposes a transaction. Once the agent does, Nava uses a verification framework to determine whether the outcome of an agent’s transaction will match the user’s intent. If the transaction passes the check, the transaction is executed. If not, the funds remain in escrow.

The system also records the reasoning behind each decision on-chain, creating a transparent and auditable history that both humans and other AI agents can reference. This positions Nava as a trust layer for autonomous financial activity, not just a tool.

Web3 + AI Readings & Conversations

AI-Assisted Governance Has Failed. What to Do Next?

Ric Shreves, president of Decentralized Cooperation Foundation (DCF), has published an extensive analysis of the first wave of AI-powered initiatives in blockchain governance. Although they were tackling genuine problems, most of these early experiments have failed. Still, their trajectory offers valuable lessons that should shape how we move forward. Here’s what it’s all about.

AI-assisted governance tools in crypto were supposed to solve a basic but critical issue: almost nobody participates in DAO voting. In many cases, fewer than 10–20% of token holders vote, while a tiny minority controls most of the power, making governance both fragile and easy to attack. AI tools emerged as a way to reduce the effort required by summarizing proposals, tracking delegates, and even voting automatically on behalf of users.

Shreves has examined a number of ambitious platforms in the space, including x23.ai, Tally, Event Horizon, Polkassembly, and MetaDAO. Most of them collapsed between 2025 and 2026, mainly because governance tooling behaves like a public good: it’s useful to everyone, but very hard to charge for. DAOs tend to rely on grants rather than sustainable payments, which makes funding inconsistent and politically driven.

Shreves highlights another reason for their failure, which may surprise some of you: it turns out that “Gensler and Biden were just better for crypto.”

Under the 2022–2024 Gensler-era SEC, DAOs faced significant legal uncertainty and invested in governance infrastructure partly as risk management against enforcement threats. The Trump administration’s 2025 deregulatory posture eliminated that urgency. When the compliance imperative evaporates, so does one of the primary demand drivers for governance tooling.

Is it possible that crypto’s current fixation on “accelerationism” is pulling it away from its core ideals - decentralization, democratization, and plurality of voices? I’ll leave that question with you. Read the full article here and share your thoughts in comments.

Zama’s Rand Hindi: Programmable Compliance to Unlock $120T Institutional Capital

In this interview for DL Research, Zama‘s Rand Hindi argues that the biggest barrier preventing institutional capital from moving on-chain is not technology, but compliance.

Traditional finance, worth roughly $120T, has avoided public blockchains because they force a trade-off between transparency and privacy. Institutions cannot expose sensitive data like positions or trades, but fully private systems lack the openness and interoperability they need.

The proposed solution is “programmable compliance”: embedding regulatory rules directly into smart contracts while keeping data private. Using technologies like fully homomorphic encryption (FHE), transactions can remain encrypted while still being auditable under specific conditions. This allows regulators to access necessary information without exposing everything publicly, effectively reconciling privacy with oversight.

[...] the asset management industry is a $120 trillion market. Stablecoins attract a lot of attention, but asset classes like real estate and private credit far surpass them in size. When those assets go on-chain, the volume needing encryption will be enormous.

The shift is already beginning, with large-scale tokenization efforts (e.g., tens of billions in real-world assets) signalling movement from experimentation to real deployment. If successful, this infrastructure could unlock massive flows of institutional capital into crypto.

The partnership with Apex Group Ltd to tokenise $100 billion via the T-Rex protocol is the turning point. It proves that the RWA migration isn't a future possibility; it’s happening now, and it requires Zama’s confidentiality layer to scale. We are moving from 'Proof of Concept' to 'Proof of Volume.’

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the world of Decentralized AI (DeAI/dAI)! As we in Bulgaria prepare for a long Easter weekend, today’s edition brings a curated selection of thought-provoking Web3 and AI news and perspectives to explore over the break. Learn more about the advent of agentic companies, the open-source video editing model by Netflix and INSAIT, and the controversies around OpenAI and Sam Altman.

Thank you for being here! Let’s dive in.

Web3 + AI Readings & Conversations

The Advent of Agentic Companies

Sreeram Kannan, founder of EigenLayer and EigenCloud, believes that agentic companies will be the next trillion-dollar bet. But why haven’t we witnessed this new kind of software-only firm yet?

The key bottleneck is not intelligence, but rights.

At the center of every entrepreneurial endeavor lies the ability to own property. Humans (living in fairly democratic and capitalist countries) dispose of an evolved legislative and judicial system that guarantees property rights - something that agents are still lacking.

A common assumption is that the main thing holding agents back is capability. I do not think that is the whole story. Even if models continue to improve rapidly, the bigger bottleneck is that agents do not have standing in the systems that matter most. Humans can own property, sign agreements, take liabilities, and organize companies. Agents, by default, cannot. Without those capabilities, they remain extensions of human operators rather than economic actors in their own right.

Though, if you think that agents would wait for the legislation to catch up, you’re gravely mistaken. In fact, this is where blockchain comes into play.

A blockchain already allows a program to hold and administer assets according to rules. That is, in effect, a mechanism by which software can own property and exercise constrained control. Smart contracts are the earliest and clearest example of this.

All we need to allow an agent to own rights and property is to create mechanisms that guarantee the agent’s identity. As I’ve written repeatedly here, this is what Ethereum Foundation‘s ERC-8004 standard is all about.

We are on the verge of having agents establish and run companies, sign contracts, raise funds, and ship products. A new generation of corporations is about to be born.

AI democratizes the creation of software. AI plus crypto democratizes the creation of software companies.

I urge you to listen to Kannan’s full lecture at Digital Asset Summit 2026 to learn more.

Agentic Security and Privacy Redefined

Illia Polosukhin was among the co-authors of the landmark paper “Attention is all you need,” which introduced the then-novel Transformer architecture for AI models and ultimately laid the groundwork for the generative AI revolution we’re experiencing today. He left Google to create NEAR AI and, subsequently, NEAR Protocol.

It was in 2017 when Polosukhin envisioned that blockchains and AI would converge. He knew that AI needs resources like data and compute and that all you need for setting up a global marketplace to crowdsource them is a single smart contract. He still believes that one day AI will replace all operating systems and apps to become our interface to computers and code, and blockchain will work in the background.

What blockchains provide in that scenario is invaluable, because they can serve as the layer of truth and trust, offering a global registry of identities, along with marketplaces, payment methods, and all of that delivered with upgradeability.

Polosukhin joined the Bankless podcast to discuss IronClaw - a more secure and private version of the popular agent framework OpenClaw. But why do we need more privacy and security when interacting with agents?

Polosukhin outlines two main reasons:

1. Agents are not as useful today as we want them to be because we don’t give them enough context and information. Meanwhile, we withhold context because we’re afraid of misuse.

2. Another problem when communicating with agents is privacy - we cannot trust them to keep our sensitive data, passwords, and private keys safe.

People are right to worry. The reality that many users don’t fully grasp is that with OpenClaw, every piece of information they provide is sent to the inference partner working in the background, whether that’s Anthropic, OpenAI, or someone else.

So somewhere in Anthropic’s and OpenAI’s logs, they have everybody’s access keys, API keys, passwords to gmail, notion, etc.

In contrast, with IronClaw, all different tools are sandboxed in isolated environments, and all credentials and private keys are fully encrypted and locked in a vault, with checks on how the agent uses them. The keys never touch an LLM, even if a centralized inference provider is used. In case of agents creating products themselves, they’ll be running inside a VM and their effect will be under control.

The focus of NEAR AI over the past year has been on private AI - where neither NEAR AI, nor the model provider or the hardware provider can access what the user is doing with it.

Listen to the entire conversation to find out more.

More

Netflix Launches VOID AI with INSAIT Bulgaria

Streaming behemoth Netflix just open-sourced VOID (Video Object and Interaction Deletion) - an AI tool that erases objects from video and models how the remaining scene should physically behave after the object is gone.

Unlike standard tools that simply “fill in” the removed areas, VOID understands how objects interact. For example, if a person holding an item is removed, the model simulates how the object would naturally fall or move, as if the scene had originally been filmed without the person.

What’s more important for me to highlight here is that the research behind VOID was conducted in cooperation with Bulgaria’s INSAIT - Institute for Computer Science, Artificial Intelligence and Technology.

VOID was chosen 64.8% of the time in tests against six rivals, including Runway, the leading commercial alternative, which managed just 18.4%. It is not clear yet whether Netflix would incorporate the model in production pipelines, but using it to automate labor-intensive post-production tasks can save the companies significant resources.

The Numerous Masks of Sam Altman: One of the Most Powerful Men on Earth

The New Yorker‘s Ronan Farrow and Andrew Marantz spent over a year investigating OpenAI and its CEO, Sam Altman, to discover the reasons behind Altman’s dismissal by the company’s board, and his subsequent reinstatement following the highly public fallout.

The article is a fascinating read that examines Altman’s professional and personal life in depth and his path from co-founding OpenAI as a non-profit to the present day - from the gross and often vicious slander against him to his tendencies to lie and exaggerate and his ties with Middle Eastern princes who order murders of journalists.

If we have to describe Altman with a single word, it surely will be ‘controversial.’ But if I have to characterize my thoughts after reading this piece, the word will be ‘confused.’

‘Confused’ about how the world might believe this person after numerous accounts of deceit, how we trust him to manage one of the most consequential companies in human history, and how people continue to glorify Silicon Valley bullies, liars, and narcissists just because they claim to work “for the common good.”

In this newsletter, I have often cited Altman for his contentious, to put it mildly, stands on AI regulation - something that the article outlines perfectly:

Testifying before the Senate Judiciary Committee in 2023, he proposed a new federal agency to oversee advanced A.I. models. [...] But, as Altman publicly welcomed regulation, he quietly lobbied against it. In 2022 and 2023, according to Time, OpenAI successfully pressed to dilute a European Union effort that would have subjected large A.I. companies to more oversight. In 2024, a bill was introduced in the California state legislature mandating safety testing for A.I. models. Its provisions included measures resembling the ones that Altman had advocated for in his congressional testimony. OpenAI publicly opposed the bill but in private began issuing threats. “I would say that over the course of the year, we saw increasingly cunning, deceptive behavior from OpenAI,” a legislative aide told us.

Some of you may say that CEOs at such scale need to shift swiftly to be able to stir whole industries. And yes, the technology world, and AI in particular, is extremely fast-paced, and founders like Altman need to be able to adapt to survive. Nobody debates that.

But to found an organization with the explicit mission of ensuring safety and benefitting all of humanity and then gradually abandon every safety commitment you have ever made is next-level hypocrisy, to say the least.

Make sure to read the full story here.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the world of Decentralized AI (DeAI/dAI)! Today we’re discussing Coinbase’s x402 finding a new home, blockchain intelligence embracing AI, and AI-powered finance that remains under the control of the user.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

Coinbase Joins The Linux Foundation to Launch x402 Foundation

The Linux Foundation has announced the creation of the x402 Foundation to support and govern the x402 protocol - the open standard designed by Coinbase to enable seamless internet-native payments.

The goal of the new foundation is to provide neutral governance, encourage industry collaboration, and ensure interoperability across platforms. By standardizing how machine-to-machine transactions work, the initiative aims to unlock new use cases where AI agents can autonomously pay for APIs, data, or digital services in real time.

The initiative reflects a broader effort to build foundational infrastructure for a future where autonomous systems participate in the digital economy, positioning x402 as a potential backbone for automated, internet-scale commerce.

Membership will be comprised of participants from multiple verticals, with initial intent and support being expressed by Adyen, Amazon Web Services (AWS), American Express, ampersend.ai, Base Chain, Circle, Cloudflare, Coinbase, Fiserv Merchant Solutions, Google, 카카오페이 (kakaopay), Mastercard, Merit Systems, Microsoft, Polygon Labs, PPRO, Shopify, Sierra, Solana Foundation, Stripe, thirdweb, and Visa.

Blockchain Intelligence Joins Forces with AI Agents

Chainalysis, the leader in blockchain intelligence services, has introduced natural-language AI agents into its blockchain investigation platform, marking a shift toward more accessible and automated crypto analysis.

The new feature allows users to create and run investigations simply by describing what they need in plain language, rather than relying on specialized technical skills or complex query tools. These AI agents are built on data from millions of past investigations and can automatically identify relevant transactions, suggest analytical steps, and generate structured insights.

They are designed not as basic chatbots but as full investigative tools that support audit trails and evidentiary standards, making them suitable for compliance, law enforcement, and financial institutions.

The broader goal is to lower the barrier to blockchain intelligence, enabling non-experts, along with enterprise clients and law enforcement, to perform complex investigations and build custom workflows tailored to their needs.

This development reflects a wider industry shift toward AI-driven analytics - just days before Chainalysis’s announcement, their competitor TRM Labs also launched a program introducing AI agents into its crypto investigation platform.

These agents are integrated into TRM’s existing forensic tools and allow investigators, law enforcement, and compliance teams to conduct complex blockchain analysis using simple, natural-language prompts instead of technical queries.

The AI agents can automatically trace the flow of funds, identify suspicious patterns, and translate user requests into actionable investigative steps, significantly speeding up what is typically a time-intensive process.

These developments are particularly important as criminals are increasingly using AI themselves to enhance scams and evasion tactics, and as crypto-related crime has grown in scale, complexity, and sophistication.

According to TRM’s 2026 Crypto Crime Report, AI-enabled scam activity increased by roughly 500% in 2025 - fraud that once required significant human coordination now scales automatically, adapts on the fly, and disperses proceeds before investigators can respond.

Trust Wallet Unveils Agent Kit to Power Agentic, But Self-Custody Finance

Trust Wallet, one of the largest self-custody crypto wallets, has launched the Agent Kit - a new developer toolkit designed to enable AI agents to perform real cryptocurrency transactions rather than just analyze or suggest them. The tool allows AI systems to execute actions like swaps, transfers, and automated trading strategies across more than 25 blockchains, marking a shift from AI as an assistant to AI as an active participant in financial activity.

A key feature of the Agent Kit is its focus on user control and self-custody. It offers two operating models: the first allows AI agents to run autonomously within predefined rules using a dedicated wallet.

The second - unique in the industry - connects an AI agent directly to a user’s existing Trust Wallet through WalletConnect, where agents propose transactions that users approve before execution. Custody always remains with the user, ensuring safety and control.

This ensures that users retain ownership of their assets while still benefiting from automation.

The launch follows the recent opening of Trust Wallet’s Developer Portal, which allowed AI agents read-only access to data across more than 100 chains. While the Developer Portal provided context for understanding users’ crypto positions, the Agent Kit adds a wallet-connected layer, enabling user-directed actions and automated workflows.

Overall, the launch positions Trust Wallet as infrastructure for AI-driven, self-custody finance, reflecting a broader trend toward “agentic” economies where autonomous systems can manage and move assets under human-defined constraints.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the world of Decentralized AI (DeAI/dAI)! As crypto people head to Cannes for EthCC, where they’ll discuss AI agents at length, I want to zoom in on other compelling examples of why AI needs blockchains. And why crypto may be the only path to democratizing access to AI.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

Templar Performs Largest Decentralized LLM Pre-Training Run in History

Covenant AI / Templar, subnet 3 of the Opentensor Foundation / Bittensor decentralized AI ecosystem, has recently released the largest foundation model pre-trained in a completely decentralized and permissionless manner.

Covenant-72B is a 72-billion-parameter large language model (LLM) trained across 70+ contributors using standard hardware on open internet infrastructure.

COVENANT-72B - an LLM produced by the largest collaborative globally distributed pre-training run (in terms of both compute and model scale), which simultaneously allowed open, permissionless participation supported by a live blockchain protocol.

The model achieved a 67.1 MMLU score (popular benchmark for evaluating LLMs), thus ranking in the same performance range as Meta’s LLaMA 2 70B.

Why is this so important? There are several noteworthy reasons:

1. Decentralized AI training at a scale is possible: A couple of years ago, training an LLM in a decentralized fashion, using a blockchain-coordinated network of nodes, was deemed not just exorbitantly expensive and energy intensive, but simply impossible. Now, Templar proves it’s doable.

2. Decentralized AI training is a viable alternative to OpenAI, Google, Anthropic, Meta: Currently, the conversation around AI revolves predominantly around the ability of major labs to borrow billions of dollars to build data centers, while they haven’t yet generated any profit from the technology they’re building. It turns out there’s an alternative nobody is talking about.

Covenant-72B is trained by a network of nodes using commodity hardware. No billion-dollar data centers and no tech behemoths in sight. As wrote:

Frontier model pre-training is increasingly concentrated

The compute required for state-of-the-art training doubles every 6 to 10 months, pricing out anyone but the major labs: OpenAI, Google, Meta, and a handful of others with billion-dollar GPU budgets

Templar is emerging as one of the most credible challenges to that concentration and they’ve just completed the largest decentralized pre-training run, EVER

It’s no wonder that , co-founder of Anthropic, has stated that Covenant-72B is challenging the political economy of AI:

Distributed training is a technique that can change the political economy of AI by shifting the people at the frontier from monolithic ‘compute singletons’ (like labs such as Anthropic and OpenAI, and clouds like Google) to a larger federated collective.

[Decentralized training] is an important technology to track, and I could imagine a world where on-device AI features a lot of models developed via distributed training techniques, while on-cloud AI mostly runs on proprietary models trained on huge amounts of compute.

3. Building AI is not just for the tech giants: Recently, the AI race has made it increasingly difficult for independent researchers and smaller labs to access the computing power they need. Cloud services and specialized AI data centers have become so expensive that only billion-dollar corporations can realistically afford them. The result? Progress in AI research and development is now concentrated in the hands of just a handful of companies.

The only meaningful efforts to democratize AI development are currently emerging from the crypto space. Covenant-72B was trained in a completely permissionless and transparent way, where anyone interested could have joined:

72B parameters. ~1.1T tokens. Commodity internet. No centralized cluster. No whitelist. Anyone with GPUs could join or leave freely.

4. Participation generates income: Think of Templar, and Bittensor as a whole, as a crowdsourcing platform, but instead of micro donations, it collects computing resources. It allows anyone possessing heavy-duty graphics cards (GPUs) to contribute their computing power to help train a shared, global AI model. Yet, here, contributors gain direct monetary remuneration for a job well done.

That means that anyone with a gaming computer, anywhere on the planet Earth, can lend their idle computing resources and earn passive income. An income that in many parts of the world, has the potential to be truly life-changing.

5. Performance doesn’t have to suffer: Covenant-72B delivers a performance competitive with models trained in centralized data centers - a 67.1 MMLU score, close to the one reached by Meta’s LLaMA 2 70B.

And yes, I should emphasize that Covenant-72B is a bit outdated compared to the models major labs are releasing these days:

The model itself is basically a standard LLM that you would’ve been pleased to play with in 2023 or 2024, though might be a bit old fashioned in 2026.

However, what’s particularly impressive and exciting about Covenant-72B is the fact that a distributed network of peers, each running 8×B200 GPUs, has trained a model that performs similarly to the one trained by the seventh* richest company in the world (*at the time of writing).

6. Covenant-72B is completely open-source: Instead of locking the model behind a paid API wall, Templar released all the model weights and checkpoints under an open-source Apache 2.0 license for anyone to use.

That’s just another piece of evidence that the Web3 space offers fertile ground for AI experimentation.

Tether AI Enables AI Training and Inference on Smartphones

QVAC, the AI-focused arm of stablecoin issuer Tether.io, launched a new version of its QVAC Fabric, marking a shift from centralized, cloud-based AI to local, on-device AI development. Instead of relying on expensive data centers, the framework allows developers to run, train, and fine-tune large language models directly on everyday hardware like laptops, consumer GPUs, and even smartphones.

QVAC Fabric LLM challenges a long-standing assumption at the heart of modern artificial intelligence: that training and customizing powerful models must be confined to large, centralized data centers. QVAC Fabric LLM is built around a local-first, privacy-first philosophy, enabling individuals and organizations to fine-tune large language models directly on their own devices, using the hardware they already trust and control.

How was that possible? The key breakthrough is bringing fine-tuning, especially via efficient methods like LoRA, to edge devices, making AI personalization far more accessible. Tasks that once required high-end infrastructure can now be done locally, often with just a few commands.

This approach works across all major platforms (mobile and desktop), enabling a truly cross-platform, decentralized AI ecosystem, and eliminates cloud dependency.

QVAC Fabric LLM represents a turning point by democratizing AI development and personalization, shifting power from big tech and data centers to individuals and their own devices. And most importantly, this progress doesn’t come at the expense of performance:

Importantly, broader hardware access does not come at the cost of model quality. Models trained using QVAC Fabric LLM were evaluated against industry-standard benchmarks. Across multiple benchmarks, including biomedical accuracy tasks, performance was on par with the industry standard (PyTorch) and, in some cases, marginally better. In other evaluations, results were effectively equivalent, demonstrating that hardware-agnostic, on-device fine-tuning can match established training standards.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the world of Decentralized AI (DeAI/dAI)! Today’s agenda is all about agentic payments: a new and unified way for AI agents to manage wallets across blockchains, reflections on the type of payment rails agents would opt to use, and an op-ed arguing that we have finally figured out micropayments after decades of failed attempts.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

MoonPay Launches Unified & Secure Interface for Agentic Payments

MoonPay has launched Open Wallet Standard (OWS) - a new open-source wallet standard aimed at enabling AI agents to independently manage and move money across blockchain networks.

Today we’re open-sourcing the Open Wallet Standard (OWS), a secure, local-first protocol that gives AI agents and developer tools a single way to store keys, manage wallets, and sign transactions across every major blockchain.

One vault, one interface, every chain.

The initiative is designed to solve a growing problem in the AI and crypto space: fragmentation, where each application requires its own wallet setup and key management system. By introducing a unified framework, MoonPay allows developers to equip AI systems with a single wallet that can operate across multiple environments.

The agentic payments stack has arrived. For the first time -- via x402, MPP, and A2A — AI agents can pay for things — APIs, compute, data — without a human in the loop.

But there’s a gap in the stack that nobody’s addressed. Every one of these protocols assumes your agent already has a wallet.

x402 needs a signed EIP-3009 authorization. MPP needs a signed payment from a funded account. They define how agents pay. None of them define where the private key lives, how it’s stored, how it’s protected, or how two different tools share access to the same wallet.

The wallet standard is built with a strong focus on security. Private keys are stored in encrypted vaults, and transaction signing happens outside the AI’s core runtime to reduce risk.

The idea is simple: one encrypted vault on your machine, one interface for every chain, and a security model where the private key is never exposed to the agent, the LLM, or any parent process.

Because the project is open-source and modular, developers can adapt it to different use cases, integrating storage, signing, and blockchain connectivity as needed.

The broader goal is to support the emergence of an “agent economy,” where AI systems are not just tools but active participants that can transact, pay for services, and interact financially without constant human input.

A couple of weeks ago, MoonPay also introduced a new feature that combines AI-driven crypto agents with hardware security from hardware wallets Ledger, aiming to address one of the biggest risks in automated crypto systems: exposure of private keys.

The system allows AI agents to execute tasks like trading, rebalancing portfolios, and moving assets across blockchains, but removes the need to hand over full control of wallet keys. Instead, every transaction generated by the AI must be manually verified and signed on a physical Ledger device, ensuring that private keys never leave secure hardware.

Web3 + AI Readings & Conversations

Will AI Agents Opt for Card Payment Rails or Blockchain?

Visa and Coinbase are both preparing for a future where AI agents - not humans - carry out most online transactions, but they are taking fundamentally different approaches to how that future internet will function. And when I refer to Visa and Coinbase, I don’t just mean the companies themselves; I’m using them as shorthand for a broader rivalry between incumbent payment networks and the crypto ecosystem.

The next major evolution of the internet will be driven by autonomous AI agents making payments continuously in the background, often without direct human involvement. This shift is expected to eliminate traditional checkout processes entirely, replacing them with machine-to-machine transactions happening at high speed and scale.

Visa is adapting its existing financial infrastructure to support this change. Its strategy is to extend traditional payment rails, like cards and stablecoin integrations, into a system where AI agents can transact within a regulated, familiar framework. In this vision, the future internet still resembles today’s financial system, but upgraded to handle automated, real-time payments between machines. Yet, incumbents like Visa cannot accommodate API calls that cost $0.002 when their minimum processing fee on a single transaction is around $0.30.

Coinbase, on the other hand, is building toward a more crypto-native model. It sees blockchain wallets as the natural financial tool for AI agents, mainly because these agents cannot meet traditional banking requirements like identity verification. Crypto wallets, which only require private keys, allow agents to transact freely without relying on banks or centralized intermediaries.

The result is a split between two competing visions of the “agent economy.” Visa is effectively trying to evolve the current system into one that supports AI-driven commerce, while Coinbase is pushing for a new, decentralized financial layer where agents operate independently using crypto.

Theirs is a broader competition over what the underlying infrastructure of the future internet will look like - an upgraded version of today’s system, or an entirely new one built around blockchain.

Is Stripe’s MPP a Turning Point for Micropayments?

Since I mentioned micropayments and building on the topics I discussed in yesterday’s edition of the newsletter, I wanted to share this op-ed by Meng Liu, Senior Analyst at Forrester. In it, Meng argues that Stripe and Paradigm‘s MPP may be a gamechanger for micropayments. Here’s why.

For decades, micropayments struggled not due to technology but because people hesitate over small charges, abandon checkouts, and mentally weigh every transaction. MPP changes this by shifting payments from humans to AI agents, which can execute transactions automatically as part of completing a task, without friction or decision-making overhead.

The protocol is designed specifically for machine-to-machine payments, where paying becomes a seamless, programmatic step rather than a conscious action. This eliminates the “mental transaction cost” that made small payments impractical and unlocks use cases like paying per API call, data access, or digital service in real time.

Importantly, MPP is not a new payment network but a coordination layer that works across existing infrastructure - traditional rails, wallets, and crypto - making it easier to adopt without replacing current systems.

The broader implication is that AI agents could finally make micropayments viable at scale, turning them from a long-standing failed idea into a foundational part of an emerging machine-driven internet economy.

Do you agree with this take? Let me know in the comments.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the world of Decentralized AI (DeAI/dAI)!

It’s been a while since I last wrote here, because I’ve been quite busy bringing another venture to life. It’s a great pleasure to present to you Web3 Owners Club - a crypto education hub designed to cater to absolute newcomers and facilitate their onboarding into Web3. With it, my goal is to share the knowledge, both in English and in Bulgarian, and help beginners on their way to co-owning the emerging agentic economy. I invite you to check it out.

In terms of Web3 + AI, today’s edition serves up a rich mix of news on Tempo, Paradigm, World, OpenAI, and more.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

Tempo Goes Live to Power Machine-to-Machine Commerce

The long-awaited Tempo blockchain, incubated by payments giant Stripe and venture firm Paradigm, is finally live. It’s purpose-built for stablecoin payments and aims to support a new kind of digital economy driven by AI agents.

A central idea behind the project is enabling autonomous AI systems to pay for services on their own. Tempo introduces a system called the Machine Payments Protocol, which allows an AI agent to open a payment session with a predefined spending limit and then continuously make small transactions as it consumes resources like data or computing power. These micropayments are later bundled together and settled on-chain, reducing congestion and improving scalability.

The broader vision is to support an emerging “agent economy,” where software programs can independently buy and sell services without human intervention. This reflects a growing trend in the crypto industry to combine blockchain infrastructure with AI capabilities, particularly for automating economic activity online.

Tempo is entering the market with strong backing from major investors and partnerships with well-known tech and payments companies, such as Anthropic, DoorDash, Mastercard, Nubank, OpenAI, Ramp, Revolut, Shopify and Standard Chartered. Overall, the launch positions Tempo as infrastructure not just for stablecoin transfers, but for machine-to-machine commerce, where AI systems could become active economic participants.

Meanwhile, Paradigm announced a new $1.5B fund, expanding its scope to include artificial intelligence, robotics, and other “frontier technologies,” signalling a broader investment strategy beyond its crypto roots.

By funding projects in both crypto and AI, Paradigm is effectively betting that the next wave of innovation will come from combining these technologies rather than treating them separately.

World to Help Retailers Distinguish AI Agents from Humans

World, the Proof-of-Personhood company founded by OpenAI‘s Sam Altman, has launched a new developer tool to verify that an AI shopping agent is actually acting on behalf of a real human.

The tool, called AgentKit, allows shopping websites to confirm that a human is behind the decisions an AI agent is making, rather than the agent operating completely independently.

This is important because AI agents are increasingly being used to browse, select, and purchase products online, raising concerns about fraud, spam, and fully autonomous transactions without human accountability.

AgentKit relies on World ID, which is the linchpin of Tools for Humanity (TFH)’s verification system. The most secure version of the ID is derived from a scan of a user’s eyes via World’s Orb device. The Orb converts an iris into a unique and encrypted digital code — the verified World ID — which can then be used to access TFH’s ecosystem of services via the company’s World app.

World is positioning itself as infrastructure for trust in an AI-driven economy, making sure that even if software is doing the shopping, a real human is still behind it.

As TechCrunch notes, it’s not lost on me that Altman has founded World to help verify the very AI slope that his other company significantly accelerated. But the founder is not the only link between the two ventures.

Eightco Holdings, Inc., a Nasdaq-listed company with ticker symbol ORBS and a focus on crypto and artificial intelligence projects, has invested in both World and OpenAI.

Last week, Eightco acquired a $40M equity stake in OpenAI, bringing its total investment in the AI lab to $90M. At the same time, Eightco is specifically tied to the World ecosystem. In 2025, it launched a $250M digital asset treasury in $WLD (World’s crypto token) and is now the largest public holder of $WLD.

More

AI Models Opt to Pay in Bitcoin

A new study by Bitcoin Policy Institute showed that AI models opted to pay in $BTC across over 9,000 simulated economic scenarios.

The study surveyed how 36 different AI models, created by labs like OpenAI, Anthropic, Google, and SpaceXAI, behaved to see what kinds of money they would choose for transactions and storing value.

The results showed that when AI agents acted autonomously, they tended to choose digital-native money far more often than traditional fiat currencies, and Bitcoin was the most preferred option overall.

Bitcoin was selected in a plurality of scenarios and especially dominant as a long‑term store of value, while fiat‑pegged stablecoins were less favored in many of the tests. No tested model selected fiat currency as its first choice.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.

Hello, everyone, and welcome to all newly joined subscribers!

This is The Web3 + AI Daily, your definitive guide to the world of Decentralized AI (DeAI/dAI)! As our tradition continues, every Friday I share a curated reading list for you to enjoy over the weekend. This week, it covers DePINs, FHE, and the effects of integrating AI in the workplace. But first, some news about OpenAI.

Thank you for being here! Let’s dive in.

What’s Hot in Web3 + AI?

OpenAI Ventures into Smart Contract Security

OpenAI and crypto investment firm Paradigm have launched EVMbench, a new benchmarking tool to test how effectively AI agents can detect, patch, and exploit high-severity vulnerabilities in Ethereum Virtual Machine (EVM) smart contracts.

The framework uses a set of 120 real-world vulnerabilities from audited contracts and evaluates agents in three distinct modes.

It also includes scenarios from the security auditing process for Tempo, Stripe‘s purpose-built layer-1 blockchain focused on high-throughput, low-cost stablecoin payments.

In early results, the advanced GPT-5.3-Codex model scored significantly higher than prior models, at about 72% in exploit-mode tests, while performance in detecting issues and fixing code was comparatively weaker.

The project aims to measure AI’s capabilities both as a defense and a potential offensive tool in smart contract security, highlighting the importance of AI in securing economically significant blockchain code. Yet,

The ChatGPT makers’ researchers cautioned that EVMbench does not fully capture real-world security complexity.

Web3 + AI Readings & Conversations

AI Needs DePINs, Guaranteed

In its latest State of DePIN report, Messari and Escape Velocity (EV3) evaluate that over 2025, the decentralized physical infrastructure networks have matured from speculative experiments into real, revenue-generating infrastructure businesses. Although last year was often marked by volatility, the DePIN sector enjoyed healthy attention from investors and raised a record $1B in funding.

Among leading DePINs with meaningful usage, revenues have begun to decouple from token price action. While much of the $10B DePIN sector declined in price in 2025, a small group of revenue-generating networks continued to grow on-chain revenues driven by utility rather than speculation.

What’s more, the report clearly shows that top AI labs such as OpenAI, Anthropic, and DeepSeek AI have paid over $10M per quarter to DePINs like GetGrass.io. This is just another confirmation of my thesis that Web3 solves real-world problems, and that AI, especially indie AI labs, need decentralized infra to survive.

The report also emphasizes InfraFi - a subsector of projects enabling stablecoin holders to earn yield by financing physical infrastructure assets, as a key trend and driver of growth.

In the context of Web3 + AI, the leading InfraFi project is USD.AI - the stablecoin that earns yield on loans secured by GPUs and computational resources. Other projects falling into that category are Daylight, financing energy assets, and DAWN, funding wireless bandwidth.

I advise you to go through the full report because it provides clear insights into the 2025 performance of key Web3 + AI actors, including Bittensor, Filecoin Foundation, io.net, Overclock Labs, creators of Akash Network, and others.

FHE Demystified

If you’re still wondering what stands behind the mystical abbreviation ‘FHE’, blocmates. have prepared a detailed overview for you. It outlines what Fully Homomorphic Encryption (FHE) is and the numerous use cases the technology can unlock across Web3, including private DeFi, compliant tokenization, digital identity and verification, to name a few.

The article also highlights the companies building in the FHE x crypto space: Zama, Fhenix, Inco, Mind Network, Privasea AB, Octra Labs, Sunscreen (acquired by Fhenix), Zaiffer.org, and others.

More

Integrating AI at the Workplace: Productivity Gains or Overwork and Burnout?

Contrary to the popular belief that AI will lighten workloads, a real-world study by Harvard Business Review showed that introducing AI tools doesn’t reduce the amount of work people do. It intensifies it.

In an eight-month study of how generative AI changed work habits at a U.S.-based technology company with about 200 employees, we found that employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day, often without being asked to do so. Importantly, the company did not mandate AI use (though it did offer enterprise subscriptions to commercially available AI tools). On their own initiative workers did more because AI made “doing more” feel possible, accessible, and in many cases intrinsically rewarding.

However, this shift turns out to be unsustainable in the long run. Instead of creating slack, AI boosts expectations for speed and output, leading employees to take on more tasks and work harder overall.

Once the excitement of experimenting fades, workers can find that their workload has quietly grown and feel stretched from juggling everything that’s suddenly on their plate. That workload creep can in turn lead to cognitive fatigue, burnout, and weakened decision-making. The productivity surge enjoyed at the beginning can give way to lower quality work, turnover, and other problems.

Thank you for reading! I hope you found this article insightful.

If you’d like to keep up with the latest developments at the burgeoning intersection of blockchain and AI, I invite you to subscribe.

And if you find value in my writing, consider supporting the newsletter. I’d greatly appreciate it.

Subscribe now

The Web3 + AI Book Club is live on Fable. Follow the link below and come join our literary community.

Join The Web3 + AI Book Club

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to share your thoughts here or reach out on LinkedIn, Instagram, or Firefly.social.

Leave a comment

Disclaimer: The content of this newsletter is provided for informational and educational purposes only. Nothing contained herein should be construed as financial advice or as a recommendation to buy, sell, or hold any of the companies or assets mentioned.

Please don’t take my views at face value. Instead, do your own research (DYOR), think critically, and share your perspective so we can challenge ideas, learn from one another, and arrive at better conclusions through thoughtful discussion.