Alexandr Wang

Overview

Alexandr Wang is an American entrepreneur, born in January 1997 in Los Alamos, New Mexico.^[1] He is known for two roles in particular: co-founding Scale AI, a company that supplies labelled data and model evaluation to AI developers, and, from mid-2025, holding senior artificial-intelligence positions at Meta.

Education and the founding of Scale AI

Wang attended MIT briefly, then left to co-found Scale AI with Lucy Guo in 2016.^[1] The company's business sits at a part of machine learning that draws less public attention than model launches but is hard to do without. Training a model usually requires examples that people have marked up: images tagged with what they contain, text paired with the correct answer, or outputs sorted by quality. That is what data labelling means. Model evaluation is the companion task — running a model over test material and scoring whether its answers are correct, consistent or safe. Both are labour-intensive, and both are needed before a model can be relied on in a product.

Growth and the billionaire milestone

Scale AI's valuation reached $7.3 billion in 2021, briefly making Wang a billionaire; he is described as having become the world's youngest self-made billionaire that year, at age 24.^[1] A valuation is the price investors assign to a company when they put money into it, and a founder's fortune is usually estimated from the value of their stake, so it moves with those valuations rather than with cash in hand. That is also why the milestone is described as brief: it rests on a number that can change.

Move to Meta

In June 2025, Meta agreed to invest $14.3 billion in Scale AI in exchange for a 49% non-voting stake.^[2] Wang stepped down as Scale's chief executive on 12 June 2025, stayed on its board, and joined Meta.^[2] A non-voting stake means the investor holds an economic interest in the company without voting rights over how it is run, which is what separates a deal of this kind from an outright acquisition.

On 30 June 2025, Wang was named Meta's chief AI officer and leader of Meta Superintelligence Labs, which he co-leads with Nat Friedman.^[3] Within that organisation he leads TBD Lab, the group that builds Meta's large language models.^[4] The first model released under his leadership was Muse Spark, on 8 April 2026.^[5]

Why it matters

Wang's path runs from the supply side of AI to the models themselves. Scale AI built its business by organising human judgement into training and testing data at scale; at Meta he now leads a lab whose product is the model rather than the data behind it. Data work and evaluation have long been treated as backstage labour, and this career line puts them in view: how far a model gets depends not only on methods and computing power but on whether high-quality human judgement can be obtained reliably. For anyone learning how the field is organised, that is a useful reminder of a part that is easy to overlook — who supplies the reference answers, and who decides whether a model's answers are right.

^[1]: Wikipedia, "Alexandr Wang". ^[2]: CNBC, "Scale AI founder Wang announces exit for Meta as part of $14 billion deal". ^[3]: CNBC, "Mark Zuckerberg announces creation of Meta Superintelligence Labs. Read the memo". ^[4]: Wikipedia, "Meta Superintelligence Labs". ^[5]: TechCrunch, "Meta debuts the Muse Spark model in a 'ground-up overhaul' of its AI".