# Emil Michael **Under Secretary of Defense for Research and Engineering** since May 2025 (confirmed 54–43). He is the Pentagon's chief technology officer, and in the second term its most visible voice on AI procurement. Before that he was [[Uber]]'s Chief Business Officer from 2013 to 2017 and [[Travis Kalanick]]'s closest lieutenant during Uber's most aggressive expansion, when it raised about $15 billion. In 2014 he suggested at a dinner that Uber could dig up dirt on journalists who criticized it, and later apologized. He left in 2017 when investors forced Kalanick out. In March 2026 he told interviewers he would "never forgive" them. ## Portfolio - **Research and engineering.** The USD(R&E) oversees the Pentagon's labs, DARPA and technology strategy. - **Defense Innovation Unit.** He served as acting director of the [[Defense Innovation Unit]] from August 2025, after its director resigned. - **Autonomy.** He is one of the principals behind the **Defense Autonomous Warfare Group**, the successor to the [[Replicator Initiative]], which carries a $54.6 billion FY2027 request. - **Maven.** In May 2026 a Pentagon official said the department would "never again" rely on a single AI provider. At an NDIA conference on September 11, 2026, Michael said the question was already moot for the [[Maven Smart System]] ([[Defense Tech Conflicts of Interest]]). ## The Anthropic fight Michael became the public face of the Pentagon's fight with [[Anthropic]], the AI company that had refused to let its models be used for mass domestic surveillance or fully autonomous weapons. | Date | Event | |---|---| | Feb 27, 2026 | Defense Secretary [[Pete Hegseth]] declares Anthropic a "supply chain risk," a label normally used against adversary-nation vendors, which bars military contractors from doing business with it. The designation is formalized in March. | | Mar 2026 | Anthropic sues. A federal judge in San Francisco temporarily blocks the designation, calling it "classic First Amendment retaliation." | | Aug 27–28, 2026 | Judge Rita Lin (N.D. Cal.) grants Anthropic summary judgment on one of the two statutory grounds, calling the designation "illegal and baseless." | | Sept 25, 2026 | A divided D.C. Circuit panel upholds the designation on the second statutory ground (Judges Katsas and Rao; Judge Henderson dissenting). Michael posts: "The hammer of justice has smashed @AnthropicAI arguments." Anthropic says it is "considering all options." | Fortune described him in early 2026 as "leading the war against Anthropic." Michael's stated objection is that a vendor's usage policies amount to "inserting their opinions in the chain of command." Critics read it as punishment for a company's safety terms ([[Anthropic Timeline]], [[Autonomous Weapons Debate]]). ## Why he matters here Michael combines the vault's revolving-door and founder-grievance patterns with the question of who gets to set the rules for military AI: the vendor or the customer. The Pentagon's answer so far is the customer, enforced with a designation built for foreign adversaries. The courts are split on whether that is lawful. ## See also - [[Anthropic]] · [[Anthropic Timeline]] · [[Autonomous Weapons Debate]] · [[Maven Smart System]] - [[Replicator Initiative]] · [[Defense Innovation Unit]] · [[Defense Tech Conflicts of Interest]] - [[Uber]] · [[Travis Kalanick]] · [[Pete Hegseth]]