Anthropic Exits Tech Industry Group, Backs Legislation to Restrict Chip Exports to China
nashnova research
Anthropic has left the Information Technology Industry Council after ITI lobbied Congress to block three bills restricting chip exports to China — a leading AI company is now openly breaking with Google, OpenAI, and Nvidia on chip controls.
What is the fight about?
ITI wrote to the Senate and House Armed Services Committees last week, urging them to strip the AI OVERWATCH Act, the Chip Security Act, and the MATCH Act from the annual defense policy bill.
ITI's argument: the bills would undermine U.S. dominance in global markets.
Anthropic takes the opposite view — it explicitly supports all three bills and quit the group over the disagreement.
This means → members of the same trade body can no longer agree on the central question: should the U.S. tighten chip controls on China?
What do the three bills actually do?
AI OVERWATCH Act: requires Commerce to notify Congress and open a review period before approving sensitive dual-use chip exports to adversary nations (the window was cut from two years to 18 months; ITI still calls it "an eternity").
MATCH Act: pushes U.S. allies to jointly restrict sales of chip-manufacturing equipment to China, aiming to slow China's domestic chip industry.
Chip Security Act: forces companies to track where AI chips end up, targeting smuggling of U.S. tech products.
In plain terms = the three bills tighten China chip controls from three directions — export approvals, equipment supply chains, and end-use tracking.
How far along are the bills?
All three have bipartisan support in the House Foreign Affairs Committee.
With White House technical input, they have been folded into the Senate's National Defense Authorization Act (NDAA) manager's package.
Conservative groups Heritage Action and American Compass have also endorsed them publicly.
This means → these are not aspirational proposals — the Senate reconvenes September 14, and the NDAA heads to a full floor vote. That is the nearest hard deadline.
Why does Anthropic keep breaking ranks?
This is not the first time: Anthropic previously split with OpenAI and Google on AI safety legislation.
It is also the only major AI company that did not co-sign the open letter supporting open-weight models.
This reflects a pattern — Anthropic is systematically positioning itself in the "safety-first" camp, even at the cost of leaving mainstream industry coalitions.
What to watch next?
Key date: September 14 — the U.S. Senate returns from summer recess and the NDAA enters the full-chamber vote process.
Anthropic's exit from ITI is a signal, but whether it shifts the legislative outcome depends on whether other major players follow.
ITI's roster still includes Google, OpenAI, and Nvidia — their stance is the real variable in this fight.
市场有风险，内容仅供研究参考，不构成投资建议。