AI Policy Wiki
Dashboard

Howard Lutnick

medium confidence · updated 2026-07-21

US Secretary of Commerce in the second Trump administration; emerged in mid-2026 as a central figure in AI export-control policy after signing the order restricting foreign access to Anthropic's most advanced models.

Howard Lutnick is the US Secretary of Commerce in the second Trump administration. Through the Commerce Department's Bureau of Industry and Security and its Center for AI Standards and Innovation (CAISI), the department holds jurisdiction over AI chip export controls and frontier-model evaluation, placing Lutnick at the center of federal AI policy as those tools became the administration's primary means of shaping the industry.

Role in AI export-control policy

Lutnick was reported in October 2025 to be on the outs within the administration, described by sources as going off message in television interviews and lacking a firm handle over his department's jurisdiction (Source: axios.com). His standing recovered through the June 2026 confrontation between Anthropic and the administration. A letter signed by Lutnick imposed export controls on Anthropic that effectively created a licensing regime, contributing to the June 12, 2026 takedown of the company's Fable 5 and Mythos 5 models for all users after the US government limited foreign access to them; officials said the framework could eventually extend to other AI labs (Source: axios.com; reuters.com). One senior administration official said of Lutnick, "He's fixing a problem. He's not being a problem. And he's doing a great job" (Source: axios.com).

By mid-June 2026, with AI-and-crypto czar David Sacks having stepped back from day-to-day involvement and Senior White House AI Policy Advisor Sriram Krishnan preparing to leave, Axios reported that influence over AI policy had shifted to a broader group in which Lutnick was a leading figure. He led meetings on the sidelines of the June 2026 G7 summit on expanding "trusted partner" access to advanced US AI models and stood next to President Trump during the summit press conference. Chris Fall at Commerce's CAISI held parallel technical meetings in Washington (Source: axios.com; reuters.com).

The June 12, 2026 directive took the form of an Is-Informed Letter (IIL) requiring Anthropic to obtain a license before any export, reexport, or in-country transfer of its Mythos and Fable models to any foreign person worldwide, including foreign nationals employed by Anthropic in the United States. Anthropic said it was not technically feasible to block access only by foreign persons on short notice and disabled the models for all users. According to public reporting, the directive followed outreach to senior government officials by a US company raising concerns that researchers had identified a method of bypassing safety guardrails, enabling access to unrestricted cybersecurity capabilities such as identifying zero-day vulnerabilities and generating working exploit code; The Wall Street Journal reported that talks between Amazon's CEO and US officials triggered the action (Source: mayerbrown.com; wsj.com). Bloomberg published the letter itself (Source: bloomberg.com).

The IIL asserted two grounds for Commerce authority: Section 4817(b)(1) of the Export Control Reform Act of 2018, which authorizes interim controls on emerging or foundational technologies essential to national security, and Section 744.22(b) of the Export Administration Regulations, which lets the Bureau of Industry and Security require a license where there is unacceptable risk of diversion to a military-intelligence end use or end user in countries of concern. Legal analysts at Mayer Brown described both applications as novel: it was the first time Commerce treated an AI model itself, rather than its weights or source code, as controlled technology, and treating remote API access as a controlled "release" departs from three BIS advisory opinions (2009, 2011, and 2014) holding that remote access to cloud-based software is not an export — a gap the pending Remote Access Security Act would address by statute (Source: mayerbrown.com). An Anthropic customer challenged the directive as exceeding statutory authority in Legion LegalTech, Corp. v. United States (No. 1:26-cv-02225).

On June 26, 2026, Lutnick issued a second letter exempting "certain trusted partners" and their foreign-national employees, along with Anthropic's own foreign-national employees, from the license requirement applicable to Mythos 5; the letter did not address exemptions for Fable 5 (Source: mayerbrown.com). Commerce withdrew the controls on both models on June 30, 2026, Lutnick writing that Anthropic "has agreed to proactively detect and address security" issues (Source: insideaipolicy.com).

Chip export policy and enforcement

Lutnick's department has simultaneously administered a loosening of chip export controls toward China: in early 2026 the administration approved exports of higher-tier Nvidia chips to China and suspended further export restrictions, a recalibration that drew congressional objections (Source: eastasiaforum.org). Members of Congress pressed Lutnick in both directions: a January 12, 2026 congressional letter opposed Nvidia H200 chip sales to China (Source: amo.house.gov), with John Moolenaar, chair of the House Select Committee on China, writing repeatedly against H200 exports (Source: eastasiaforum.org); Sen. Tom Cotton and Rep. Bill Huizenga urged stronger AI-chip export controls (Source: cotton.senate.gov); and Rep. Sam Liccardo wrote to Lutnick on June 18, 2026 concerning the frontier-model export controls (Source: liccardo.house.gov).

On enforcement, Lutnick told congressional appropriators in April 2026 that BIS needs additional funds in fiscal 2027 to police export controls and prevent diversion of semiconductors to China and other adversaries, endorsing significantly higher fines for export-control violations and expanded cooperation with industry and allied governments (Source: rollcall.com; wttlonline.com).

Relationships