Anthropic Offers $850,000 Salary to Train AI to Design Chips

Surging computing demand from generative AI has made custom application-specific integrated circuits, or ASICs, a strategic priority for technology companies seeking lower costs, better performance and less reliance on general-purpose GPUs. Anthropic’s move to assemble an internal chip team signals that competition among AI developers is expanding beyond models and data centers into semiconductor architecture and AI-assisted hardware development.

As of Aug. 9, 2026, Anthropic is recruiting a research engineer to help teach AI systems how to design chips, offering annual compensation of as much as $850,000. The package stands well above typical pay for conventional chip-engineering roles and highlights a new front in the AI race: using artificial intelligence to automate parts of the custom ASIC design process while building semiconductor expertise in-house.