
Web3 + AI + Privacy: Spotlight on Ethereum
With its current focus on privacy, the crypto world returns to its cypherpunk roots. But how would it affect Web3 + AI?

The Web3 + AI Daily #35
Your definitive guide to the world of Decentralized AI (DeAI/dAI).

The Web3 + AI Daily #51
Daily insights into the fascinating convergence of Crypto and AI.
>100 subscribers

Coinbase rolled out Agentic Wallets, a wallet infrastructure built specifically for autonomous AI agents. The goal is to enable AI agents to move beyond advising to actually acting in decentralized finance and economic activities on their own.
Agentic Wallets builds on the x402 machine-to-machine payment protocol that Coinbase is developing with other internet stakeholders to support autonomous crypto payments. See my overview of x402 below to learn more.
Agentic Wallets allow developers to set security guardrails like spending limits, and support gasless transactions on Base Chain and other chains. While Coinbase’s previously released AgentKit tool made it possible to embed wallets during agent creation, Agentic Wallets, rather, is a plug-and-play solution to "give any agent a wallet" with the ability to transact on behalf of users.
Meanwhile, a BeInCrypto-cited report by Artemis indicates a 92% drop in x402 daily transactions since December 2025. These results come amid heightened interest in OpenClaw's open-source AI agents framework and Ethereum's 'Trustless Agents' standard.
Moreover, as Davide Crapis told Laura Shin (see Daily #56), the x402 payment protocol is a complementary, but not necessary solution. What that means is that there are many other options for agents to execute autonomous payments and transactions.
Over the past weekend and during a market slump, Bitcoin miner Cango Inc. sold 4,451 BTC (valued at around $305M) on the open market. The company stated that the proceeds will be used to repay a BTC-collateralized loan and fund a strategic expansion into AI infrastructure.
Cango will still maintain its mining operations, while repurposing its existing mining and data-center assets toward GPU-based AI services.
Cango’s move reflects a broader trend of mining companies pivoting to AI data-center businesses alongside or instead of pure cryptocurrency mining, amid tightening of traditional mining margins.
While at the Consensus 2026 conference in Hong Kong, crypto VCs described the current state of Decentralized AI as being in a post-hype “trough” phase, with less buzz and more focus on real, utility-driven projects rather than speculative ideas.
Anand Iyer of Canonical and Kelvin Koh of The Spartan Group stated that investors are shifting away from overhyped GPU marketplaces and decentralized alternatives to big AI models.
“Twelve months ago, it was enough to have a wrapper on ChatGPT. That’s no longer true.”
Instead, VCs are finally focusing on backing purpose-built, full-stack solutions that address specific problems and deliver tangible value.
“Speculation won’t drive product anymore. We have to think about users first.”
However, we shouldn't dismiss GPU marketplaces altogether. A couple of weeks ago, Jake Brukhman, founder and CEO of CoinFund, published an op-ed in support of decentralized AI training and its potentially enormous impact.
Brukhman argues that decentralized networks could link global GPU resources into a shared training fabric - the very same thing GPU marketplaces are doing. Then, they can reward contributors with tokens that represent ownership or revenue rights in the resulting AI models.
Thus, decentralized AI training could create a new digital asset class by allowing many participants to jointly train AI models and own a stake in them through tokens, rather than these models being controlled by a few big tech firms.
This tokenization could make AI models themselves tradable assets, democratizing access and investment in AI beyond centralized companies. Although still early and experimental, these systems signal a shift toward markets that price and trade “digital intelligence” itself.
Brukhman cites Prime Intellect, Gensyn, and Pluralis Research as the projects that have already achieved significant progress in the field.
Thank you for reading! If you haven't done so yet, I invite you to subscribe to stay in the loop on the hottest dAI developments.
If you want to support the publication financially, you can either purchase my writer token $WEB3AI, or buy my creator token $ALBENA on ZORA.
I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to reach out on social media.
Disclaimer: None of this should or could be considered financial advice. You should not take my words for granted; rather, do your own research (DYOR) and share your thoughts to encourage a fruitful discussion.

Coinbase rolled out Agentic Wallets, a wallet infrastructure built specifically for autonomous AI agents. The goal is to enable AI agents to move beyond advising to actually acting in decentralized finance and economic activities on their own.
Agentic Wallets builds on the x402 machine-to-machine payment protocol that Coinbase is developing with other internet stakeholders to support autonomous crypto payments. See my overview of x402 below to learn more.
Agentic Wallets allow developers to set security guardrails like spending limits, and support gasless transactions on Base Chain and other chains. While Coinbase’s previously released AgentKit tool made it possible to embed wallets during agent creation, Agentic Wallets, rather, is a plug-and-play solution to "give any agent a wallet" with the ability to transact on behalf of users.
Meanwhile, a BeInCrypto-cited report by Artemis indicates a 92% drop in x402 daily transactions since December 2025. These results come amid heightened interest in OpenClaw's open-source AI agents framework and Ethereum's 'Trustless Agents' standard.
Moreover, as Davide Crapis told Laura Shin (see Daily #56), the x402 payment protocol is a complementary, but not necessary solution. What that means is that there are many other options for agents to execute autonomous payments and transactions.
Over the past weekend and during a market slump, Bitcoin miner Cango Inc. sold 4,451 BTC (valued at around $305M) on the open market. The company stated that the proceeds will be used to repay a BTC-collateralized loan and fund a strategic expansion into AI infrastructure.
Cango will still maintain its mining operations, while repurposing its existing mining and data-center assets toward GPU-based AI services.
Cango’s move reflects a broader trend of mining companies pivoting to AI data-center businesses alongside or instead of pure cryptocurrency mining, amid tightening of traditional mining margins.
While at the Consensus 2026 conference in Hong Kong, crypto VCs described the current state of Decentralized AI as being in a post-hype “trough” phase, with less buzz and more focus on real, utility-driven projects rather than speculative ideas.
Anand Iyer of Canonical and Kelvin Koh of The Spartan Group stated that investors are shifting away from overhyped GPU marketplaces and decentralized alternatives to big AI models.
“Twelve months ago, it was enough to have a wrapper on ChatGPT. That’s no longer true.”
Instead, VCs are finally focusing on backing purpose-built, full-stack solutions that address specific problems and deliver tangible value.
“Speculation won’t drive product anymore. We have to think about users first.”
However, we shouldn't dismiss GPU marketplaces altogether. A couple of weeks ago, Jake Brukhman, founder and CEO of CoinFund, published an op-ed in support of decentralized AI training and its potentially enormous impact.
Brukhman argues that decentralized networks could link global GPU resources into a shared training fabric - the very same thing GPU marketplaces are doing. Then, they can reward contributors with tokens that represent ownership or revenue rights in the resulting AI models.
Thus, decentralized AI training could create a new digital asset class by allowing many participants to jointly train AI models and own a stake in them through tokens, rather than these models being controlled by a few big tech firms.
This tokenization could make AI models themselves tradable assets, democratizing access and investment in AI beyond centralized companies. Although still early and experimental, these systems signal a shift toward markets that price and trade “digital intelligence” itself.
Brukhman cites Prime Intellect, Gensyn, and Pluralis Research as the projects that have already achieved significant progress in the field.
Thank you for reading! If you haven't done so yet, I invite you to subscribe to stay in the loop on the hottest dAI developments.
If you want to support the publication financially, you can either purchase my writer token $WEB3AI, or buy my creator token $ALBENA on ZORA.
I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to reach out on social media.
Disclaimer: None of this should or could be considered financial advice. You should not take my words for granted; rather, do your own research (DYOR) and share your thoughts to encourage a fruitful discussion.

Web3 + AI + Privacy: Spotlight on Ethereum
With its current focus on privacy, the crypto world returns to its cypherpunk roots. But how would it affect Web3 + AI?

The Web3 + AI Daily #35
Your definitive guide to the world of Decentralized AI (DeAI/dAI).

The Web3 + AI Daily #51
Daily insights into the fascinating convergence of Crypto and AI.
Share Dialog
Share Dialog
1 comment
@coinbase unveils Agentic Wallets amid 92% slump in the daily transactions using its x402 Agentic payment protocol. Why? Let's unpack all of this 👇 👉 Coinbase rolled out Agentic Wallets, a wallet framework built specifically for autonomous AI Agents. The solution lets agents manage funds, hold identity, and transact on-chain without human intervention. 👉 Agentic Wallets builds on Coinbase's x402 machine-to-machine payment protocol - the innovation that took the crypto space by storm in 2025. Yet, according to @artemis, the x402 daily transactions plummeted by 92% since December. 👉 Coinbase is building the agentic commerce infra, but it seems the demand isn't here yet. Or, the space, fascinated by OpenClaw and @ethereum's ERC-8004, is simply more interested in other payment solutions. After all, as @ethereumfndn's @davidecrapis said, x402 is a complementary, but not a necessary solution. Read more and subscribe 👇 https://web3plusai.xyz/daily_57