In an attempt to surpass competitors, Meta intends to invest $15 billion in Scale AI.

The transaction would give the parent company of Facebook a 49% ownership in the start-up, making it one of the largest of its kind.

In one of the largest agreements of its kind, Meta intends to invest roughly $15 billion in data-labelling start-up Scale AI and hire the group's co-founder and leading researchers in an effort to catch up to competitors.

According to people with knowledge of the situation, the acquisition, which may be disclosed as early as Wednesday, would give Meta a 49% share in Scale AI and value the start-up at about $28 billion.

According to one of the persons, Meta's goal is to establish a "superintelligence" lab that will surpass OpenAI, Anthropic, and Google, which are also creating models they say will surpass human intelligence. This includes investing in Scale and trying to recruit its best employees.

A request for comment from Scale was not answered, and Meta did not reply.

Since Llama 4, Meta's most recent large language model, did poorly on critical reasoning and coding benchmarks, its release was mostly viewed as a failure.

In the meantime, rivals like Google, OpenAI, and Anthropic have each introduced a new generation of potent "reasoning" models that solve issues by dissecting them piece by piece. Additionally, open source rivals like China's DeepSeek, which have produced strong models at a significantly lower cost, are putting pressure on Meta.

With a market valuation of almost $2 trillion, Meta has made significant investments in generative artificial intelligence. However, Mark Zuckerberg, the CEO, has repeatedly reorganized the initiatives since development has been stalling. Joelle Pineau, vice-president of AI research at Meta, resigned, the company said in April.

According to The New York Times, Alexandr Wang, a 28-year-old paper billionaire who co-founded Scale in 2016, will be joining Meta's "superintelligence" lab. The Information and Bloomberg were the first to report on Meta's investment details.

The primary function of Scale is to manually identify the data required to train sophisticated AI models in order to guarantee its accuracy.

In addition to positioning Scale to support businesses creating autonomous vehicles and, more recently, generative AI models, Wang has developed connections with some of Silicon Valley's most prominent investors and technologists, such as Sam Altman of OpenAI.