On April 1, the Department of Commerce launched its Call for Proposals for the American AI Exports Program, responding to President Trump’s executive order aimed at promoting the American AI tech stack globally. Applications from “Pre-Set Consortium” offering a complete technology stack will be accepted until June 30. Eligible tech stacks must include AI-optimized hardware, data pipelines and labeling systems, AI models and systems, security and cyber security measures, and AI applications for sector-specific functional use cases. Applications for “On-Demand Consortium” in direct response to a specific, identified export opportunity will be released at a future date.
On April 22, the House Foreign Affairs Committee is slated to markup additional legislation including the Multilateral Alignment of Technology Controls on Hardware (MATCH) Act, which would restrict the sale of semiconductor manufacturing equipment (SME) to foreign adversaries and the Deterring American AI Model Theft Act of 2026, which seeks to “prevent foreign adversaries from threatening the national security of the United States by extracting key technical features of closed-source, American-owned artificial intelligence models.” The Committee has been taking an active role in AI legislation so far this year, including approving the Chip Security Act in late March. That legislation would “prevent advanced American chips from falling into the hands of adversaries like Communist China” and require location verification for advanced chips, enforce mandatory reporting, and require the Department of Commerce to study additional safeguards.
The previewed release of Anthropic’s new model, Claude Mythos, is creating concern among lawmakers and administration officials over its purported ability to identify – and potentially exploit – cyber security vulnerabilities. Anthropic briefed government officials before the launch, prompting Federal Reserve Chairman Jerome Powell and Treasury Secretary Scott Bessent to call a special meeting with the CEOs of major U.S. banks to discuss the tool and its implications for cyber defenses. Now, the White House is looking to give agencies access to the new model “in the coming weeks,” as the fallout from Anthropic’s battle with the Pentagon continues to play out (more on that below). In addition, Anthropic CEO Dario Amodei met with White House Chief of Staff Susie Wiles last week – his second high-stakes meeting with a senior Trump official this year – in what sources describe as an effort to pave the way for a broader deal.
