The Fed Puts Marc Andreessen in Charge of Measuring What AI Does to Jobs

Federal Reserve Chair Kevin Warsh has appointed Marc Andreessen, co-founder of venture firm Andreessen Horowitz, to co-lead a new task force studying how AI reshapes productivity, employment, and economic growth. The “Productivity and Jobs” panel — announced July 9 — is one of five external task forces Warsh launched to modernize how the Fed reads the economy, and it must deliver concrete recommendations by the end of 2026.

Andreessen shares the co-lead role with Stanford economist Charles I. Jones, one of the most cited growth theorists alive, and Asha Sharma, CEO of Microsoft’s Xbox division. It’s a deliberately unusual troika: a VC with tens of billions riding on AI adoption, an academic who wrote the standard models of technological growth, and an operator running a division that just absorbed heavy AI-driven restructuring.

The mandate matters more than the names. The Fed sets interest rates against estimates of productivity growth and labor-market slack — and AI is currently scrambling both. If AI genuinely lifts productivity, the economy can run hotter without inflation, which argues for lower rates. If AI displacement is hollowing out white-collar employment faster than official statistics capture, the labor market is weaker than it looks. Warsh’s bet is that the Fed’s existing data pipelines, built for a pre-AI economy, can’t answer either question — so he’s importing people who see the frontier directly.

The conflict-of-interest critique writes itself. Andreessen Horowitz holds stakes across the AI stack — OpenAI, Mistral, Databricks, dozens of application-layer startups — and a16z spent 2025 lobbying aggressively against state AI regulation. Now its co-founder helps shape the analytical frame the world’s most powerful central bank uses to think about AI. Warsh and Andreessen are also 30-year friends, and Andreessen publicly championed Warsh’s nomination as chair. None of this is disqualifying for an unpaid advisory role. All of it will be quoted back when the panel’s recommendations conveniently favor faster AI deployment.

Still, the alternative was worse. A Fed that models 2026 labor markets with 2019 assumptions makes policy errors measured in percentage points of employment. Getting the people with the best proprietary view of AI adoption — VCs, hyperscaler operators — into the room is how you fix that, provided the room also contains people like Jones whose incentives point at accuracy.

The recommendations land by December. Watch whether the panel treats AI job displacement as a measurement problem or a talking point — that choice will tell you who actually ran it.

Sources

Related reading