Table of Contents

• You might want to know


• Main Topic


• Key Insights Table


• Afterwards...

You might want to know

Did the timing and sequence of the official’s divestitures and government actions create potential conflicts of interest?

How do ethics rules and oversight apply when senior technology officials hold or recently held stakes in private AI firms?

Main Topic

This article summarizes and analyzes recent disclosures showing that Emil Michael, serving as the Department of Defense’s undersecretary for research and engineering and its chief technology officer, sold a stake in Perplexity AI in June for an amount reported within a broad range. According to the financial filings reviewed by press outlets, the sale was reported as falling between $5 million and $25 million. That disposal is one of several AI-related financial moves Michael made during the year that together have drawn media attention and prompted questions from ethics observers.

Earlier in the same year, Michael divested holdings in xAI, a company associated with Elon Musk. The xAI stake, which had been reported at an initial range of $500,000 to $1 million upon his entry into government, was sold on January 9 for an amount also reported between $5 million and $25 million, representing a substantial realized gain compared to the value disclosed at the time of his appointment. Reporters and analysts have noted this as a very large percentage increase and have used it to underscore questions about private-sector ties carried into public service.

In addition to the Perplexity and xAI transactions, Michael’s filings show proceeds from other technology-sector interests earlier in the year, including at least $5 million from a stake in Brex sold in April. These sequential divestments have been publicly documented and are now part of his official disclosures, along with an itemization of a prior personal loan he took from Perplexity before joining government.

That loan was listed as between $250,000 and $500,000 at an interest rate of 4.57%, according to Michael’s 2025 financial disclosure. In his ethics agreement when entering public service, Michael committed not to profit from unvested shares in Perplexity; however, the disclosures and public reporting do not clearly state whether the June sale involved vested shares, unvested shares, or a mix. This ambiguity has been central to scrutiny from ethics experts and media commentators.

Separately, Michael previously served on Perplexity’s advisory board and resigned from that role at the beginning of 2025, prior to joining the administration in May of that year. His prominent position in the Pentagon has also made him a visible advocate for the department’s policy positions regarding certain AI vendors and suppliers.

One notable policy dispute this year involved Anthropic, a developer of large language models that directly competes with Perplexity and xAI for government contracts. In late February, presidential direction instructed federal agencies to cease using Anthropic’s services. Shortly thereafter, the Department of Defense designated Anthropic as a "supply chain risk," and contractors were barred from engaging with the company. The move affected procurement pathways and prompted litigation from Anthropic.

Last week, a U.S. district judge ruled that the Pentagon’s blacklist of Anthropic was improperly retaliatory and therefore unlawful, vacating the designation and issuing a permanent injunction. The court found the action had penalized the company for refusing to permit certain uses of its models — specifically, their use in mass surveillance or weapon systems — and that the government’s approach overstepped legal boundaries. By the time of the court’s ruling, the Pentagon had already reallocated at least one contract originally intended for Anthropic to OpenAI, and had signed classified-network agreements with a number of other AI firms, including xAI.

Ethics observers have expressed concern about the timing and sequencing of the official’s divestitures and the department’s vendor decisions. Richard Painter, a former White House ethics official, has been quoted saying the official "should have sold all interest in the company before he started working," arguing that potential appearances of conflicts matter even if formal rules are followed. The Pentagon has responded by stating that Michael and other officials are in compliance with applicable ethics laws and that the department employs a layered ethics review and divestiture process designed to prevent conflicts.

The Perplexity sale also intersects with the company’s fundraising prospects. Public reports indicate Perplexity has been in discussions with major chipmaker Nvidia about a funding round that could value the company at more than $30 billion, a figure notably higher than the roughly $20 billion valuation reported in September 2025. Such upward valuation momentum helps explain why holdings in AI startups can produce rapid and sizable gains for early investors and advisors, and why divestiture timing is consequential for both the official and public perception.

Taken together, the disclosures of multiple high-value exits, prior advisory ties, and the loan arrangement combine to create a fact pattern that draws scrutiny even where officials assert they complied with the letter of ethics laws. Key issues include the clarity of disclosure language about vesting status, the timing of divestiture certificates and actual sales, and how procurement or policy actions align chronologically with those private-sector holdings and exits.

It is important to distinguish between legal compliance and the appearance of conflict. Ethics regulations are intended both to prevent actual conflicts and to preserve public trust in decision-making. Even if formal procedures were followed — such as obtaining divestiture certificates or executing pledges not to profit from unvested shares — observers note that the sequence of events can still undermine public confidence unless transparency and careful separation are evident.

Finally, these events illustrate broader challenges at the intersection of government technology leadership and private-sector ties in a fast-moving industry. As AI companies rapidly increase in value and win government contracts, the potential for rapid personal financial gain by individuals who soon enter public service grows. That dynamic has prompted renewed calls for stricter rules, clearer disclosure requirements, and more conservative approaches to divesting or avoiding private stakes before assuming influential policy roles.

Key Insights Table

Afterwards...

Looking forward, questions remain about whether current ethics frameworks and disclosure practices adequately address the rapid appreciation and commercial influence of AI firms. Potential responses include tightening divestiture timelines, clarifying reporting on vested versus unvested holdings, and considering pre-appointment divestment as a standard for senior technology positions. Continued transparency, independent review, and clearer public disclosure will be central to maintaining trust when private-sector expertise intersects with national-security technology decisions.

Policymakers, ethics officials, and industry stakeholders will likely continue to debate the right balance between attracting private-sector talent into public service and preventing real or apparent conflicts. The sequence of divestitures, loan arrangements, advisory roles, and procurement choices in this case will probably be referenced in those discussions as regulators and agencies refine guidance for future appointments.