Alexandr Wang

Chief AI Officer, Meta Superintelligence Labs — Meta Founder, Board Member — Scale AI

Listen — profile

0:00 / 3:00

Profile

At 19, Alexandr Wang dropped out of MIT and co-founded Scale AI in a Y Combinator batch with Lucy Guo, a former colleague from his teenage stint at Quora. The premise was almost boring: AI models need enormous volumes of carefully labeled data, and nobody wanted to run that business at scale. Wang did. By 2021 Scale was labeling data for OpenAI, Meta, Google, and the Pentagon, and Wang — at 24 — was the world’s youngest self-made billionaire. Not bad for a kid from Los Alamos whose parents were both physicists for the U.S. military.

What Wang actually built is less sexy than a frontier lab but arguably as strategic. Scale is the “data foundry” — his preferred metaphor, riffing on TSMC — for the entire AI supply chain. If you’ve trained a big model, there’s a real chance Scale’s contractors labeled some of your RLHF data, wrote some of your instruction-tuning examples, or red-teamed your outputs. That position gave Wang extraordinary situational awareness of what every serious lab was doing, and he used it. He cultivated policymakers, testified before Congress, and in January 2025 took out a full-page Washington Post ad addressed to President Trump titled “America must win the AI war,” laying out a five-point industrial policy for U.S. AI dominance.

Then came the pivot. In March 2025 Scale won the prime contract for Thunderforge, the Department of Defense’s flagship program to deploy AI agents across military planning and operations — partnering with Anduril (Palmer Luckey’s defense-tech company) and Microsoft. Three months later Mark Zuckerberg paid $14.3 billion for 49% of Scale and hired Wang away to become Meta’s first-ever Chief AI Officer, running the newly created Meta Superintelligence Labs. He stepped down as Scale CEO but stayed on the board. In April 2026 MSL shipped its first model, Muse Spark — the clearest signal yet that Meta’s AI stack is being rebuilt from the ground up under Wang.

For developers learning AI, Wang is worth understanding less as a researcher and more as a systems operator. He doesn’t build transformers; he builds the infrastructure that sits underneath them — the labeling pipelines, the eval harnesses, the government relationships, the data contracts. He also represents the sharp-elbowed hawkish wing of Silicon Valley AI: explicitly framing the field as a U.S.-China confrontation, leaning hard into defense work, and arguing that “personal superintelligence” is Meta’s answer to Sam Altman’s AGI pitch. Whether you find that posture bracing or alarming, it’s shaping where a lot of the money and contracts are flowing.

Key Articles & Papers

The AI War and How to Win It 2025 — Wang's open letter to President Trump — the five-point industrial-policy argument that marked his turn from data vendor to AI policy operator. Introducing Thunderforge: AI for American Defense 2025 — The announcement of Scale's prime contract with the DoD to bring AI agents into military planning and operations across combatant commands. Scale AI Announces Next Phase of Company Evolution 2025 — The official announcement of Meta's $14.3B investment and Wang's move to become Meta's Chief AI Officer leading Superintelligence Labs. Frontier Data Foundries: Unlocking AI's Future 2024 — Wang's clearest articulation of the 'data foundry' thesis — why labeled, human-generated data is the strategic chokepoint for frontier AI. Scale's Alex Wang on the US-China AI Race 2024 — Long-form interview laying out Wang's geopolitical framing of AI competition — the intellectual backbone of his later Trump letter.

Controversies

Data labeler working conditions. Scale AI and its subsidiary Outlier have faced a California class-action wage-theft lawsuit alleging contractor misclassification, unpaid overtime, and exposure of workers to graphic content on a Meta-linked self-harm safety project. The U.S. Department of Labor opened an investigation into Scale’s Fair Labor Standards Act compliance, active since at least August 2024. Scale has disputed the allegations and says pay meets or exceeds local living-wage standards. Reporting from The Washington Post documented gig workers in the Philippines training AI systems at below-minimum-wage rates.

Defense and dual-use posture. Wang’s aggressive push into Pentagon work — Thunderforge, partnerships with Anduril, and his pro-militarization advocacy — has drawn criticism from researchers concerned about deploying LLM-based agents in combatant-command planning before evaluation standards are mature.