Subscribe to The Web3 + AI Newsletter
<100 subscribers
Circle, the company behind the $USDC stablecoin, has partnered with robotics firm OpenMind to develop a micropayment system, based on Coinbase's x402 protocol, that enables machine-to-machine payments using $USDC. The goal is to allow autonomous AI agents, robots or automated software to transact automatically with each other, paying for services, content, compute or data in real time and without human intervention.
Under the initiative, OpenMind plans to integrate AI into autonomous robots and join a broader ecosystem (including partners like Amazon Web Services (AWS), Anthropic, and NEAR Protocol) that aims to build a so-called “machine economy”, with automated payment flows between intelligent agents. While the project is still in early stages and real-world use is yet to be demonstrated, the collaboration could establish a new standard for AI-powered micropayments, enabling scalable, efficient and programmable value transfer between machines and services.
Circle and OpenMind will launch the x402 micropayment protocol using USDC, enabling machine-to-machine payments in automated, permissionless transactions.
Caesar, a deep research Crypto x AI company, has teamed up with digital finance firm Centrifuge to issue on-chain equity, making it the first “crypto-native” company to do so.
Through Centrifuge’s infrastructure, including its SEC-registered transfer agent, Caesar’s equity will be represented as real securities on a public blockchain, not as derivatives or “wrappers.” This means shareholders will get full shareholder rights, and equity issuance, transfer, and record-keeping happen via smart contracts.
The move signals a shift: it shows that blockchain-native startups can evolve into investor-ready, regulated companies with transparent, programmable ownership, potentially lowering friction for future funding, compliance, and liquidity. For the broader market, this could establish a blueprint for future on-chain IPOs or private-equity offerings, bridging the traditional corporate finance world and decentralized finance.
Centrifuge's SEC-registered transfer agent powers Caesar's move into regulated onchain equity, marking a milestone for crypto-native companies.
Caesar also collaborated with Warden Labs to enable blockchain smart contracts / AI agents to request and verify complex, research-grade outputs in a verifiable on-chain manner. This combination aims to resolve a common trade-off: research being off-chain and unverifiable, or agents acting with limited context. Now, agent-driven applications can reason with context-rich, cited data and execute actions on-chain, raising the bar for reliability, transparency, and sophistication.
The integration of Caesar’s research intelligence with Warden’s verifiable agent infrastructure opens a new class of high-value, research-driven onchain applications.
Under this partnership, agents built on Warden can leverage Caesar to draw on structured research (market data, technical/scientific analysis, governance reports, etc.), then act on that information via Warden’s infrastructure across EVM and CosmWasm chains.
Warden brings AI to web3, enabling any application, protocol or smart contract to integrate safe AI.
Bitcoin miner Bitfury committed $50M to Gonka, a decentralized AI-compute network where anyone, from a single GPU owner to a full data center, can contribute computing power for real AI workloads. This marks the first investment under Bitfury’s new $1B fund for ethical emerging technologies.
Since its launch in August 2025, Gonka reportedly scaled quickly: within three months it aggregated compute power equivalent to more than 6,000 NVIDIA H100 GPUs, sourced from distributed hosts across the network. Bitfury’s investment aims to accelerate development and adoption of the Gonka protocol and ecosystem, signaling institutional backing for decentralized AI infrastructure.
Furthermore, Bitfury's George Kikvadze published an article arguing that Gonka aims to play for AI what Bitcoin did for money: a decentralized, open-access infrastructure for AI compute, breaking the monopoly of centralized players in AI and cloud compute.
Instead of traditional proof-of-work where miners waste energy on hashing, Gonka uses a “Proof-of-Work 2.0” (the “Sprint Consensus Protocol”) where GPU owners contribute to useful AI inference workloads. Tokens ($GNK) are rewarded based on real compute work, aligning value with real resource usage rather than speculation. This means almost 100% of GPU contributions go toward productive AI tasks (training, inference, etc.), not consensus overhead, which Gonka claims solves a core inefficiency of many blockchain-AI systems.
GONKA: Why I Believe This Is the Bitcoin of AI "First they ignore you. Then they laugh at you. Then they fight you. And then you win." I opened my book And Then You Win with these words because ...
Robotics firm XMAQUINA announced an undisclosed funding round led by Borderless Capital. The company clarified that investors, such as Borderless.Capital, Waterdrip Capital, vVv, along with strategic angels from ArkStream Capital, LD Capital-LLC and KuCoin Exchange Ventures, have all participated on the same terms as the public in Wave 3 of its Genesis Auction.
The round signals institutional confidence in tokenized robotics automation.
When XMAQUINA's founding members designed the DEUS Genesis Auction, the goal was simple: create an open, permissionless structure where anyone, not just insiders, could gain early exposure to the Robotics Capital Markets on equal terms. The objective has always been clear. It is time to move beyond the typical VC-first model and make room for broader participation.
A research team at Anthropic used leading AI models (e.g. Claude Opus 4.5, Claude Sonnet 4.5, and GPT‑5) to find smart contract vulnerabilities on Ethereum, BNB Chain, and Base Chain. In 19 out of 34 recently deployed contracts, they found working exploits with a combined simulated yield of about $4.6M. These were simulated attacks in sandboxed blockchain forks, no real money was at risk.
To test zero-day vulnerability risk, they ran Sonnet 4.5 and GPT-5 on 2,849 recently deployed contracts with no known vulnerabilities and still uncovered two previously unknown exploitable bugs, producing simulated profits of ≈ $3,694. In a set of 405 real-world smart contracts that were exploited between 2020–2025, the team managed to produce turnkey exploits on 207 (≈ 51%) of them, corresponding to about $550.1M in simulated “stolen” funds.
Their conclusion: as AI cyber-capabilities improve (tool use, reasoning, long-horizon planning), AI-driven automated exploits may pose a serious and growing threat to smart-contract security. They recommend proactively using AI for defense and auditing to stay ahead.
December 1, 2025 Winnie Xiao*, Cole Killian* Henry Sleight, Alan Chan Nicholas Carlini, Alwin Peng *MATS and the Anthropic Fellows program AI models are increasingly good at cyber tasks, as we've written about before. But what is the economic impact of these capabilities?
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.
I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to reach out to me 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 create a fruitful discussion.
Albena Kostova-Nikolova
No comments yet