Tech News

Reports point to new frontier AI models in testing at Anthropic and OpenAI, though neither company has confirmed the details. Reports circulating this week point to new large language models in development at two of the industry's most closely watched labs.

Mizuho is said to be leading the Japanese tranche as the AI company prepares for a US public offering. Anthropic is reportedly preparing to raise as much as $3.5 billion from Japanese investors ahead of a planned listing on the Nasdaq exchange in the United States.

A growing list of issuers has pushed back against OKXICE's effort to offer tokenized versions of their shares. OKX's tokenized stock initiative, referred to as OKXICE, is facing mounting resistance from the companies whose shares it aims to represent. Fifty issuers have now registered objections to the product, according to reporting on the dispute.

A deal would mark OpenAI's first major shift toward Google's custom Tensor Processing Units for large-scale AI training and inference. OpenAI is reportedly exploring a purchase of Google's artificial intelligence chips, according to reports from The Cryptonomist and Decrypt.

The company disclosed a holdup accessing its Bitcoin treasury on the Biconomy platform, triggering a steep selloff in its shares. Datavault AI, a publicly traded firm that has positioned itself in the data and artificial intelligence sector, saw its stock price collapse after disclosing trouble accessing its own Bitcoin holdings.

The bank's analysts say the rise of autonomous AI payment agents could pressure the stablecoin issuer's core business model. Mizuho has told clients that Circle could face a revenue risk of up to 20% tied to the growth of AI agents operating in payments and commerce.

The AI developer's paperwork for a highly anticipated public listing details security risks tied to its Claude models. Anthropic, the company behind the Claude family of AI models, has filed for an initial public offering. Reports describe the filing as a blockbuster event, reflecting the company's scale and the intense investor appetite for AI-related listings.

The Ethereum co-founder also raised concerns about risks tied to KYC verification in AI systems Vitalik Buterin, the co-founder of Ethereum, has expressed support for an AI model that deliberately limited its own hacking capabilities. The model, identified in reports as Kimi AI, reportedly had its offensive cybersecurity skills curtailed as part of its design.

The stablecoin issuer's open automation platform, known as MDK, is gaining industry support as AI tools enter Bitcoin data centers. Tether, the issuer of the USDT stablecoin, has announced that 15 Bitcoin mining companies are backing its automation toolkit known as MDK.

The asset manager points to tokens linked to computing and data demand, including one it says has gained 580% this year. Grayscale Investments has published research naming three crypto projects it considers well placed to capture value from the ongoing artificial intelligence infrastructure expansion.

Swedish officials have restricted access to OpenAI's and Anthropic's chatbots while allowing continued use of Microsoft's AI assistant. Sweden's Government Offices has moved to restrict which artificial intelligence tools its staff can use. ChatGPT, made by OpenAI, and Claude, made by Anthropic, are now blocked. Microsoft Copilot has been kept available for continued use.

Developers say faster-moving AI tools are prompting quicker public disclosure of some software bugs, raising pressure on node operators to update promptly. Bitcoin Core developers are changing how they handle bug disclosure. Some vulnerabilities will now be made public sooner than under previous practice.

Binance founder Changpeng Zhao says artificial intelligence won't break Bitcoin, but leaked personal data could still hurt users Changpeng Zhao, the founder of Binance widely known by his initials CZ, has weighed in on growing concerns about artificial intelligence and crypto security.

The new share structure lets Mark Zuckerberg raise capital for Meta's AI buildout without diluting his voting power. Meta has created a new class of non-voting shares for Mark Zuckerberg, according to Cryptopolitan. The structure is intended to help the company raise capital to fund its artificial intelligence ambitions.

The new AI agent is positioned as a proactive rival to OpenAI and Meta, with built-in safeguards for payment data. Anthropic has rolled out a new AI agent called Claude Duet, according to reporting from Cryptopolitan. The agent is built to handle purchases independently, using a virtual card rather than a user's actual payment credentials.

The stablecoin issuer backs a company building payment infrastructure for autonomous AI agents. Tether has led a $15 million funding round for Mural, a startup building payment infrastructure for artificial intelligence agents. CoinDesk and crypto.news both reported the investment on October 21, 2026, describing Mural as an AI-focused payment company.

Cryptopolitan and CryptoBriefing agree on Anthropic's $80.7B net loss and other core IPO-filing figures but give different total sizes for its cloud/infrastructure commitments. Cryptopolitan and CryptoBriefing agree on Anthropic's $80.7B net loss and other core IPO-filing figures but give different total sizes for its cloud/infrastructure commitments.

Server maker's early sales figures lift investor sentiment amid broader AI infrastructure demand Super Micro Computer shares rallied following the company's disclosure of strong preliminary sales results. The report, covered by CryptoBriefing, did not specify exact figures but indicated the numbers were well received by markets.