Anthropic in Talks To Rent Microsoft AI Chip Servers as Donald Trump Cancels Executive Order Regulating Frontier Models
Anthropic is exploring a partnership to rent servers powered by Microsoft AI chips for complex workloads. This collaboration could enhance Microsoft's position against rivals Alphabet and Amazon by offering alternatives to dominant Nvidia chips. Meanwhile, US President Donald Trump cancelled a scheduled AI executive order, aiming to maintain the country's technological lead over China.
Artificial intelligence startup Anthropic is in early-stage discussions to rent servers powered by Microsoft custom-designed AI chips to handle its complex workloads. The potential agreement, which was first reported by The Information, could signal a major shift in the cloud computing landscape as tech giants race to secure infrastructure for next-generation AI models. Shares in Microsoft rose by approximately 1.5% in premarket trading following the news.
A successful partnership would mark a substantial victory for Microsoft as it seeks to establish its hardware as a viable alternative to Nvidia, which currently dominates the AI chip market. By developing its own silicon, Microsoft aims to emulate rivals Alphabet and Amazon, both of which have invested heavily in proprietary chip infrastructure to reduce dependency on third-party suppliers and lower operational costs for enterprise clients. Anthropic to Finalising USD 1.5 Billion AI Joint Venture with Blackstone and Goldman Sachs: Report.
Anthropic Microsoft Chip Partnership
The negotiations between the Claude software maker and Microsoft are still in their infancy, and sources close to the matter indicate that the talks may not ultimately result in a formal agreement. Representatives from both Microsoft and Anthropic did not immediately respond to requests for comment regarding the financial terms or the scale of the server infrastructure under discussion.
The push for enhanced computational power comes at a time when Anthropic is developing highly sophisticated systems. The company recently created an unreleased model codenamed Mythos, which has reportedly triggered intense interest within Washington national security circles due to its unprecedented cyber capability.
AI Executive Order Cancelled
In a separate development impacting the broader technology sector, US President Donald Trump abruptly cancelled the signing of a highly anticipated artificial intelligence executive order. The proposed policy would have mandated that leading tech firms, including Google, Meta, OpenAI, and Microsoft, submit their most powerful AI models for comprehensive government evaluation before releasing them to the general public.
The sudden cancellation occurred just hours before a scheduled Oval Office event, reportedly after several high-profile technology chief executives were unable to attend on short notice. The White House had extended invitations to the leaders of OpenAI, Google, Anthropic, Meta, and Microsoft just 24 hours prior, leading to scheduling conflicts that forced several firms to send replacement executives instead.
US China Tech Race
President Trump later indicated that a recent state visit to China heavily influenced his decision to delay the regulatory framework. He expressed strong concerns that implementing stringent pre-release evaluation pipelines could inadvertently hinder domestic innovation and compromise the technological advantage held by American firms in the global arena.
"I think it gets in the way of – you know, we’re leading China, we’re leading everybody, and I don’t want to do anything that’s going to get in the way of that lead," Trump told reporters. The decision represents a significant, immediate relief for the American tech sector, allowing companies to deploy rapidly evolving models without facing regulatory compliance bottlenecks. Anthropic Claude Sonnet 4.6 Released: New Model Features 1 Million-Token Window and Enhanced Coding Capabilities; Check Details.
The cancelled order would have granted the Office of the National Cyber Director and various intelligence agencies a months-long window to audit frontier models. Security anxieties had intensified after Anthropic's Mythos model demonstrated an 83.1% success rate at reproducing known vulnerabilities during internal testing, systematically unearthing tens of thousands of undiscovered zero-day flaws across major operating systems.
(The above story first appeared on LatestLY on May 22, 2026 04:17 PM IST. For more news and updates on politics, world, sports, entertainment and lifestyle, log on to our website latestly.com).
