AI Firms Pay Up to $500,000 for Ex-Spies to Guard Trade Secrets

Salaries of $200,000 to $500,000 now attach to counterintelligence roles at OpenAI, Anthropic and Nvidia, according to workforce data tracker Revelio Labs, as AI companies race to shield trade secrets from state-backed espionage.

The demand spike follows the scandal that ended Jensen Huang's tenure as CEO of Nvidia, the AI chip giant he cofounded. Last weekend he admitted a romantic affair with a woman U.S. officials suspect was recruited by China's Ministry of State Security. Surveillance showed she was feeding information to Beijing, according to reports. China has denied any espionage. Nvidia's board investigated and found no compromise of national security or company secrets — though the board, it should be noted, was assessing its own chief executive. Huang retained access to President Donald Trump and senior administration officials while U.S. intelligence raised alarms about the relationship.

The named party, Megaspeed executive Alice Huang, has not been charged, and the claims against her rest on unnamed officials. That gap between allegation and proof deserves a caveat every time the numbers are cited.

How fast is the hiring accelerating?

Revelio data compiled by Reuters shows steep growth at the largest AI labs:

• Anthropic hired 33 counterintelligence staff so far this year, up from 14 last year and nine in 2024. In 2025 it appointed former CIA deputy director David Cohen and former NSA cybersecurity director Dave Luber to its national security advisory council.
• OpenAI's counterintelligence hiring jumped to 22 in 2026 from nine in 2025. In 2024 it named former NSA director Paul Nakasone to its board.
• Nvidia hired 16 people for counterintelligence this year despite the scandal.

OpenAI is seeking a senior global security investigator with counterintelligence as a key skill; Anthropic is advertising a threat-intelligence security engineer; ASML is hunting for a Taiwan-based intelligence officer.

Christopher Falkenberg, a former U.S. Secret Service agent and president of Insite Risk Management, told Reuters: "The AI firms have snapped up a lot of security talent." He added: "There are some very sophisticated people out of government thinking about security for these businesses… Demand outstrips the supply of people with relevant experience." Scarce supply, of course, is what pushes pay toward the top of that range.

What does the Nvidia case teach?

A month before the Huang affair surfaced, an OpenAI engineer was arrested for planting a listening device in a company conference room. The Justice Department alleged he was communicating with a member of the Saudi royal family.

Eric O'Neill, a former FBI counterintelligence operative, told Fortune's Tatiana Sataua that approaches rarely begin with a request for secrets. They start at a conference, through a business introduction, over drinks, or on a dating app. Warning signs include a relationship that moves unusually quickly, persistent questions about sensitive matters, requests to meet influential colleagues, and encouragement to keep secrets.

"By the time a security professional raises concerns, the executive may already be emotionally invested and unwilling to listen," O'Neill said. His prescription: teach executives to recognize manipulation, and give others a way to intervene when the boss will not listen.

Is Washington moving the same direction?

The private sector is not acting alone. The Washington Post reported last month that the Pentagon plans a spying program deploying "nonofficial cover officers" posing as executives, part of a broader economic warfare campaign. Politico reported a Defense Department effort to covertly sabotage foreign technology programs: the Technology Support Office, established in April by merging two classified programs that have run smaller-scale sabotage since 2023. The Pentagon has denied both reports.

In September, CIA Deputy Director Michael Ellis said the U.S. is expanding espionage against China in AI, biotechnology and chips. This year the Pentagon also formed the Economic Defense Unit for "deterring, denying, disrupting, and helping to defeat adversaries through economic means," sources told DefenseScoop.

Denials aside, the direction is clear on both fronts. AI has become a national-security asset, and the market for people who can protect it now outstrips the people who can do the job.