TRM Labs reports that AI‑enabled crypto fraud activity rose about 500% over the past year, with generative tools accelerating phishing, identity theft, and laundering. Chainalysis says these operations earn 4.5 times more per scheme and generate 9 times more daily transactions than non‑AI scams. Recent incidents include large holder losses and U.S. seizures linked to pig‑butchering rings.

On 26 February 2026, Ripple took part in a $5 million seed round for San Francisco‑based t54 Labs, which builds identity and risk infrastructure for autonomous AI agents across XRPL, Solana and Base. The financing, co-led by Anagram, PL Capital and Franklin Templeton, will fund engineering growth and developer outreach.

On February 25, 2026, Thrive Capital committed roughly $1 billion to OpenAI. In December, the firm agreed to a deal at a $285 billion valuation, and OpenAI is now raising up to $800 billion, implying Thrive's position could approach triple its entry. Thrive also closed its Thrive X fund at over $10 billion, nearly twice its prior vehicle.

MoonPay has launched MoonPay Agents, a non-custodial software layer that lets AI agents open wallets, hold digital assets and transact with stablecoins directly onchain. The infrastructure is being introduced as financial institutions explore stablecoin rails and as forecasts project the AI agent market could reach $236 billion by 2034.

Anthropic introduced Claude Cowork on January 30 and, during an online event on Tuesday, said Claude can now operate inside popular business apps. The company highlighted integrations with Microsoft Excel and PowerPoint, Google Drive, Gmail, and Calendar, plus DocuSign and LegalZoom, as stocks reacted with sharp declines in early February and a Tuesday rebound for FactSet (+3.8%) and Thomson Reuters (+8.8%).

AMD and Meta unveiled a five-year AI chip partnership valued at up to $60 billion to deploy as much as 6 gigawatts of AMD Instinct GPUs. AMD shares were up as much as 15% in premarket trading; as of 8:16 AM EST on February 24, the stock traded at $214.35 after a $196.60 close on February 23.

DeepSeek plans to unveil a new AI model as soon as next week, and U.S. officials say it was trained on Nvidia’s Blackwell chips that are barred from export to China. A senior official said Monday the setup could breach U.S. export controls and suggested the hardware may be clustered at a DeepSeek site in Inner Mongolia.

Shares of leading United States cybersecurity firms fell sharply after Anthropic launched its Claude Code Security vulnerability scanner on Feb. 20. Palo Alto Networks, CrowdStrike and Fortinet posted declines of up to 18% in the days following the debut of the AI-powered tool, which is built on Anthropic’s Claude Opus 4.6 model.