Anthropic Explores In-House AI Chip Design Amid Rising Demand and Hardware Shortages: Report

Artificial intelligence research lab Anthropic is exploring the possibility of designing its own custom AI chips, according to reports from sources familiar with the matter. The San Francisco-based startup, which develops the Claude chatbot, is considering this move to address the global shortage of high-performance hardware essential for training and deploying advanced AI systems. While the plans are in their early stages, the move signals a strategic shift toward self-reliance in a hardware market currently dominated by a few major players.

The exploration of custom silicon comes at a time of explosive financial growth for the company. Anthropic recently disclosed that its run-rate revenue has surpassed USD 30 billion in early 2026, a significant increase from USD 9 billion at the end of 2025. Despite this growth, a spokesperson for the company declined to comment on the specific details of the hardware project. Anthropic Mythos Dangers: Check All About Risks Involved With New Artificial Intelligence Model by Anthropic.

Current Infrastructure and Strategic Partnerships

Anthropic currently relies heavily on third-party hardware to power its AI models. The company utilizes Tensor Processing Units (TPUs) designed by Google, as well as specialized chips from Amazon. Earlier this week, Anthropic strengthened these ties by signing a long-term agreement with Google and Broadcom, the firm that assists in TPU design, to bolster its computing capabilities.

This new deal is part of Anthropic’s broader commitment to invest USD 50 billion in United States computing infrastructure. However, the high cost of cloud credits and the restricted supply of leading-edge GPUs have prompted the company to evaluate whether developing proprietary silicon would be more cost-effective and efficient in the long term.

Technical Challenges and Industry Trends

Designing advanced AI chips is a capital-intensive and technically complex undertaking. Industry experts estimate that developing a single high-end AI chip can cost approximately USD 500 million, excluding manufacturing. The process requires a dedicated team of specialized engineers and significant investment to ensure that the manufacturing phase—often handled by external foundries—remains defect-free.

Anthropic’s deliberations mirror those of its primary competitors. Both OpenAI and Meta have reportedly explored or initiated their own custom chip projects to reduce dependency on external vendors like Nvidia. By designing hardware tailored specifically to the architecture of their own large language models (LLMs), these companies hope to achieve better performance-per-watt and lower operational costs for their AI services.

Early Stage Deliberations and Future Outlook

Sources indicate that Anthropic has not yet committed to a final design or assembled a full-scale engineering team for the project. There remains a possibility that the company may ultimately choose to continue purchasing chips from established partners rather than venturing into the semiconductor design space. Anthropic Expands Claude AI Integration to Microsoft 365 for All Users.

As the demand for AI services like Claude continues to scale, the decision will likely hinge on the availability of hardware and the stability of the global supply chain. If Anthropic proceeds, it would mark a significant transition from a software-focused research lab to a vertically integrated technology firm, joining the ranks of major cloud providers that control both the silicon and the software layers of their infrastructure.

(The above story first appeared on LatestLY on Apr 10, 2026 10:38 AM IST. For more news and updates on politics, world, sports, entertainment and lifestyle, log on to our website latestly.com).