Alexandr Wang, Meta’s young AI leader, bets billions on new AI models

Alexandr Wang, then 28, was chosen by Mark Zuckerberg to lead Meta’s new artificial intelligence strategy. As co-founder of Scale AI, he helped reorganise the company’s research and now heads the lab responsible for the Muse models.

In April 2025, during a work meeting, Wang received a WhatsApp message from Zuckerberg. The conversation lasted about 44 minutes, and two days later, Wang sent a memo outlining his proposals for the company’s future.

The appointment placed Wang at the centre of a billion-dollar investment. For 2026, Meta plans to spend between $130 billion and $145 billion on infrastructure, including the AI systems he now needs to deliver.

From Scale AI to Meta’s lab

Born in Los Alamos, New Mexico, to Chinese immigrant parents, Wang learned to code at a young age and enrolled at MIT. He left university at 19 and, in 2016, founded Scale AI with Lucy Guo. The startup grew by providing annotated data to train AI models for other companies.

Scale AI worked with autonomous vehicle manufacturers, major tech companies, and government clients. In 2021, at age 24, Wang became a billionaire when the company reached a valuation of $7.3 billion.

In 2025, Meta paid $14.3 billion for a 49% stake in Scale AI. Wang stepped down from his executive role but remained on the board. The payment was not personal compensation but a combination of equity, resources for Scale, and liquidity for its shareholders.

Strategy to build an elite team

Zuckerberg was dissatisfied with the performance of the Llama model family. Wang proposed forming a small team of exceptional professionals and giving them autonomy to rebuild the work. The ability to attract these talents was decisive in the decision.

The core research team at Meta Superintelligence Labs consists of about 100 people, supported by Meta’s larger engineering and infrastructure teams. The concentration of specialists is part of the strategy, while necessary resources remain distributed across the company.

Muse puts the bet to the test

On October 12, Meta announced the Muse Wonder and Muse Wonder Full models, expanding the family that began with Muse Spark. The new tools are designed for agents performing personal and professional tasks, with features to retrieve information from previous conversations, maintain multi-step activities, and work with programming.

Muse Wonder is available through Meta Model API and Muse Code, while the Full version was released for Muse Code subscribers. Integration with the Muse app will be implemented later, starting with paying users.

The success of this transition from research to application will be the ultimate test of Zuckerberg’s decision. Meta expects its systems to be able to receive a task, gather the necessary information, and execute it reliably. For Wang, the challenge now is to ensure the products gain the trust of their users.

Frequently Asked Questions

Who is Alexandr Wang? He is the co-founder of Scale AI and leads Meta Superintelligence Labs, Meta’s advanced AI laboratory.

How old is Wang? He was 28 when recruited in 2025 and is now 29.

Did Wang receive $14.3 billion to work at Meta? The amount was paid by Meta to acquire 49% of Scale AI, not as personal salary.

What is Meta Superintelligence Labs? It is the laboratory responsible for developing Meta’s most advanced AI models, including the Muse family.

How much does Meta plan to invest in 2026? The projection is $130 billion to $145 billion in capital investments, covering the company’s entire infrastructure.