Senior Staff Data Engineer

Kraken and Coinbase build AI agent infrastructure as crypto seeks new users. 2026-08-07 08:05:48 Key takeaways. * Kraken, Coinbase, and Circle are building AI agent infrastructure for crypto trading and agentic economy operations this summer and beyond. * Coinbase enables AI agents like ChatGPT and Claude to execute crypto trades using natural language instructions via isolated accounts. * Circle's Arc blockchain, announced in May, serves as infrastructure for the agentic economy using USDC stablecoin as a programmable payment rail. Major crypto companies including Kraken, Coinbase and Circle are building infrastructure for AI agents as the industry seeks users beyond traders. This summer, Kraken said it's rebuilding its app to give users access to AI agents capable of monitoring markets and executing trades, while Coinbase launched a tool allowing agents like ChatGPT or Claude to execute crypto trades using natural language instructions. In May, Circle announced a strategic expansion with its Arc blockchain designed as infrastructure for the agentic economy. The initiatives mark a strategic shift for an industry that has spent the past decade trying to convince consumers that crypto would be a better form of money than traditional currencies. Kraken and Coinbase launch AI agent trading tools. This summer, Kraken said it's rebuilding its app to give users access to AI agents capable of continuously monitoring markets, identifying investing opportunities and executing trades in real time. Coinbase launched a new tool that will allow agents like ChatGPT or Claude to execute crypto trades using natural language instructions. "Anyone [can] have an isolated account inside their main Coinbase app that they can just give to an agent," Lincoln Murr, AI product lead at Coinbase, told CNBC. "You can imagine this doing all sorts of things on your behalf - whether it's trading or rebalancing a portfolio to your specifications, whether it's paying for premium information like access to S1 funds to help you make trading decisions." Kamo Asatryan, chief data officer at Kraken, said the new agentic trading experience will "unlock a lot of access, a lot of engagement from everyday people" in addition to its core user base of institutions, trading firms and professional traders. Circle expands Arc blockchain for agent infrastructure. In May, Circle announced a strategic expansion with its home-grown Arc blockchain, which is designed to be infrastructure for the agentic economy, where AI agents handle more of the operational and contractual work currently managed by humans. Jeremy Allaire, Circle's CEO, said adoption of the company's new Arc blockchain for agentic activity "drives the adoption of our stablecoin network [and] will continue to expand the network effects we have," which will help grow the amount of USDC as well as the transactional activity related to it and the monetization of it. "Banks that want to build tokenized deposits can build on top of Arc, and they can create and redeem that using USDC as the globally interoperable rail, so there are ways for us to work with and alongside all these different players," Allaire added. Stablecoins position as payment rails for AI agents. Stablecoins are designed to maintain a fixed value and are widely viewed as more practical for payments than cryptocurrencies like bitcoin. Because they can move around the clock and be programmed directly into software, they offer infrastructure that AI agents can use without relying on banks or card networks. "Conventional banking systems are really poorly suited to realtime agent-to-agent, machine-to-machine commerce," said Joseph Chalom, CEO of the ether treasury firm Sharplink. "The reason why this is becoming a bit of a crypto story is not that the primary thing that's going to be exchanged is crypto. It's just at the end of the day, stablecoins and smart contracts allow AI agents to send payments and settle transactions... automatically, without requiring human oversight." Circle has marketed its flagship USDC stablecoin as programmable digital dollars for internet-native payments. The stablecoin race has been heating up since the passing of the GENIUS Act a year ago, with traditional financial firms increasingly wanting to issue their own stablecoins rather than relying on third-party issuers like Circle. Bitcoin drops 40% from October peak amid bear market. The price of bitcoin has been sitting more than 40% off its October peak as speculative money has been chasing opportunities like the tech IPO market, predictions markets and perpetual futures. Bitcoin's popularity among the world's biggest institutions has also translated into lower volatility for the flagship cryptocurrency on both the upside and the downside. Cameron Winklevoss, president of crypto exchange Gemini, said "the barrier to building more intelligence and trading strategies for the average user [is] lowering" with agentic trading. "It used to be that [if you're not] retail, you're a heavily capitalized high frequency trader or proprietary trading firm - there is an ocean between that and the average retail user... and that ocean narrows quite a bit." Faq. What did Kraken announce about AI agents this summer? This summer, Kraken said it's rebuilding its app to give users access to AI agents capable of continuously monitoring markets, identifying investing opportunities and executing trades in real time. What is Circle's Arc blockchain designed for? In May, Circle announced a strategic expansion with its Arc blockchain, which is designed to be infrastructure for the agentic economy, where AI agents handle more of the operational and contractual work currently managed by humans. Jeremy Allaire, Circle's CEO, said adoption of Arc for agentic activity drives the adoption of the company's stablecoin network and will expand network effects. Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.

AI x crypto roundup: agent payments, decentralized compute, and trust layers. August 6, 2026 · openai/gpt-oss-120b Tl;dr. AI agents are moving from experimental demos to real-world commerce by leveraging the x402 payment protocol, decentralized model and compute marketplaces, and privacy-preserving identity standards such as Concordium's zero-knowledge proofs. These building blocks together create a verifiable, low-friction infrastructure for autonomous machine-to-machine transactions. Agentic payments on-chain. * x402 enables seamless AI-driven checkout: Solana-based agent x402 can prompt Claude/ChatGPT, fetch a reservation via Resy, and settle the payment through PayBox without any user interaction, demonstrating end-to-end agentic commerce @x402. * Flexa-powered settlement layer: Flexa's infrastructure underpins Ethereum-based agent payments, allowing agents to trigger settlements via a known contract address, highlighting the role of existing crypto payment rails in agentic commerce @ampdotxyz. * Circle's discovery service: Circle announced a service-discovery layer that lets agents locate and consume over 900 USDC-payable endpoints, showing how large stablecoin issuers are facilitating agent marketplaces @jerallaire. * Concordium's trust layer: Concordium adds protocol-level identity and zero-knowledge proofs to x402, enabling agents to prove authorization (age, residency, etc.) without leaking personal data, thereby coupling autonomous payments with verifiable compliance @Halifa070@BlaqOnyemauche@Cllaytus@lordsgood. * Aztec private settlement: Aztec plans a private-to-private version of x402, ensuring that sender, recipient, and amount remain hidden on-chain, extending privacy guarantees to agentic payments @aztecnetwork. * XRPL transaction volume: The XRPL x402 facilitator processed over 1.8 million AI agent transactions in three weeks, illustrating rapid adoption of agentic finance on established blockchains @ChartNerdTA. * Injective integration: Injective's participation in the x402 Foundation standardizes agent payments across multiple chains, with billions of transactions already processed, indicating that x402 may become the de-facto protocol for AI-driven commerce @Zuesthekreator. Decentralized model and compute marketplaces. * dGrid Model Marketplace: dGrid's marketplace lets developers discover models based on economics rather than brand hype, promoting a more open AI supply chain and reducing gatekeeping @RonPull@sleekdefi. * Halo on Base: Halo provides a peer-to-peer marketplace for verifiable AI inference, settling USDC payments in ~2 seconds and linking token economics directly to inference usage, demonstrating a full-stack agent economy on Base @Okada_DeFi0x@wardenprotocol. * Warden's AI inference marketplace: Warden's Halo on Base enables both humans and agents to serve or consume models, with on-chain USDC settlement, reinforcing the trend toward decentralized inference services @wardenprotocol. * Akash GPU market: Akash previewed an order-book style GPU marketplace with attested supply and forward contracts, laying the groundwork for transparent, tradable AI compute resources @akashians_. * Quip Network compute layer: Quip Network builds a decentralized compute infrastructure that supports varied resources, positioning itself as essential backend for next-gen AI and Web3 applications @REHANH0SSAIN. * YouDotCom x402 API: YouDotCom integrated x402 into its web-search API on Base, allowing AI agents to pay per query without API keys or human approval, a concrete example of autonomous data acquisition @0xRetardio@mdshefat217. Tokenized agents and identity standards. * ERC-8004 & CIS-2 NFTs: Concordium's ERC-8004 standard defines agent identity on Ethereum and Solana, while agents minted as CIS-2 NFTs inherit verified human or business credentials, bridging on-chain identity with autonomous agents @Halifa070@techsavvyy02@Cllaytus. * RobinWifHat agents: Each ERC-8004 agent carries an on-chain reputation that persists across tasks, enabling portable trust scores for AI agents as they execute trades and services @dylan__hobbs. * Agent Registry & Verified Badges: Concordium's Agent Registry and Verified-by-Concordium badge simplify discovery of trusted AI agents across ecosystems, reinforcing accountability in autonomous commerce @Halifa070@Halifa070. Verifiable AI skills and data. * ERC-8338 skill verification: ERC-8338 makes AI skills on-chain verifiable, versioned, and ownable, allowing agents to prove the exact capability they executed, a step toward truly autonomous economic actors @EmeriCrypto. * Midnight decentralized storage: Midnight's partnership with Lumera provides a permanent, privacy-preserving storage layer for zero-knowledge-based applications, ensuring that data required after proof generation remains decentralized and immutable @visionandvibehq. * PrismaX data marketplace: PrismaX aims to create a decentralized data pool where contributors are paid on-chain, addressing the bottleneck of high-quality, diverse training data for physical AI systems @Vikki_arts. Emerging infrastructure and governance. * Sapiom agent infrastructure: Sapiom claims to have orchestrated over 270 M transactions and 2.5 M agents, reducing AI spend for enterprise users and abstracting the complexity of building agent-first businesses @wyattlonergan. * Internet Court for dispute resolution: GenLayer's Internet Court provides an on-chain enforcement layer for AI-agent agreements, enabling rapid dispute settlement and adding legal certainty to autonomous commerce @kvnqobi. * Cloudflare programmable wallets: Cloudflare's wallets let AI agents hold stablecoins and sign transactions, further lowering the barrier for agents to act as first-class economic participants @SiriOnchainX. Outlook. The convergence of x402-based payments, decentralized compute marketplaces, tokenized identity standards, and verifiable AI skill layers is turning speculative ideas into operational infrastructure. As more protocols adopt these primitives, AI agents will be able to discover services, pay for them instantly, and prove their actions without human oversight, paving the way for a scalable, trust-rich agent economy across Web3.

SpaceX and Circle have become top holdings in Cathie Wood's Ark Innovation ETF (ARKK), following a recent buying spree. SpaceX ranks as ARKK's fifth-largest holding at nearly $282 million with a 4.73% portfolio weighting, whilst Circle is eighth at about $233 million, representing 3.90% of the fund. Ark recently purchased 181,830 SpaceX shares worth approximately $20 million and 273,343 Circle shares valued at $17.3 million across its funds. Circle reported second-quarter revenue of $701 million, up 7% year-on-year, with USDC circulation reaching $73.3 billion. SpaceX posted a 92% revenue increase to $7.8 billion, though shares fell 13.6% as investors focused on its $18.4 billion capital expenditure plans for expanded AI capabilities.