Meta is paying $ 14.3 billion to get AI and its CEO, Alexander Wang, on a scale of 49 %, to improve its troubled AI efforts.
The two companies announced Thursday that as part of the deal, Wang Meta’s CEO will report Mark Zuckerberg directly and will remain in the Scale Board of Directors. Sources say that Wang will lead a new AI lab in Metaa in which the construction of the “Superintendent” will be entrusted. Meta spokesman Ashley Zandi says the company “will share some more about the attempt to join the team in the coming weeks.”
According to those familiar with the matter and other press reports, Zuckerberg is actively recruiting a new team of researchers from rival firms to join the Wang team. The founder of Meta has reached the potential recruitment directly- usually through a cold email or WhatsApp message- and some of them removed from companies like Google from seven and eight data compensation packages.
The Scale works with Google, Openi, Entropic, and others to help humans train their model by interpreting and labeling their data. Most of this is done through low -cost labor outside the United States, and it has become an important component of AI’s development. With joining Wang Meta, Scale is named Jason Druge, his former Chief Strategy Officer, named his new interim CEO.
“We have grown as more than 1,500 people and have become a trusted partner for model builders, businesses, and governments, who became famous as the world’s youngest billionaire billionaires at the age of 21, Wang said in a memo for jobs.
“Scale is now one of the most impressive companies in the world, which can be the most important technology in human history by accelerating development. Today, we are announcing a large -scale investment from Meta. It is an important milestone and you have a powerful verification of the hard work you’ve worked in this scale mission.”
Wang added that “the amount from Metaa investment will be distributed among you who owns shareholders and equity holders, while as an ongoing equity holders maintain the opportunity to participate in our future development.”
By not receiving a fully scale, Meta Big is following the playbook to hire AI talent services, while also investing at the beginning. This process is designed to reduce the risk of anti -trust checking. However, it is possible that the contract governments may still be examined with a scale, though Meta will get only non -voting shares in the company.
This year, the scale signed an agreement with the Defense Department for the AII Agent Program of its kind to plan the US Army. Since then, it has extended its business with other governments, recently signed a five -year contract to provide AI tools to Qatar. The company’s work with governments in Asia and Europe could result in a large part of its total sales in the next two constituencies.
Meanwhile, Meta went to trial only to defend itself from the breakdown of the US government. The result of this matter is still uncertain.
Since the disappointing start of Lalma 4 earlier this year, Mark Zuckerberg, CEO of Meta, has been anxious to meet rivals such as Google, Openi, Entropic and Deep Sek. The release of Lama 4 was delayed several times, and then Mata was caught by gaming a public leader board to make the model really look better. The company has not yet released Lama 4 Behimot, which is the largest and most expensive version of it, which was flown in April.
Last month, Zuckerberg said that the two top priorities of Meta for 2025 are making its chat GPT rivals, Meta AI, “well -known personal AI” and “construction of complete general intelligence.” He recently said that Meta AI has reached one billion monthly users, though the importance of Assistant in Instagram, WhatsApp, and Facebook affects this number.
In April, Meta released a standstone app for Meta AI, which introduces a social feed that shows how people are using it. The app briefly reached the top of the App Store but does not maintain its popularity level.
