“`html
Meta Platforms recently shook the artificial intelligence world by announcing a massive investment of 14.3 billion dollars to acquire a 49% stake in Scale AI, a startup specializing in data annotation for training AI models. This operation, which values Scale AI at 29 billion dollars, is accompanied by the arrival of Alexandr Wang, CEO and co-founder of Scale AI, to lead a new Meta initiative focused on artificial general intelligence (AGI). However, this strategic move raises growing concerns about competition and data privacy, as well as questions about possible evasion of antitrust scrutiny.
A strategic investment to strengthen Meta’s position in AI
Meta, already in the spotlight for its AI ambitions with projects like the language model Llama, seeks to consolidate its position in the global race for AGI. Founded in 2016, Scale AI plays a key role in the AI ecosystem by providing high-quality annotated data, essential for training powerful AI models. Among its clients are giants like Microsoft, OpenAI, and even direct competitors of Meta. By taking a non-voting minority stake of 49%, Meta avoids automatic regulatory review, unlike a complete acquisition.
Furthermore, the integration of Alexandr Wang, often described as a Silicon Valley prodigy, into Meta’s team marks a turning point. Wang will lead a new division focused on developing AGI, a field where Meta lags behind actors like OpenAI or Google. According to Reuters, this operation reflects an acqui-hire strategy, combining talent acquisition and access to strategic technological resources.
Why is Scale AI so coveted?
Scale AI generated 870 million dollars in revenue in 2024, primarily thanks to a small group of major technology clients. The startup excels in manual data annotation through independent workers, a process crucial for refining generative AI models. However, Meta’s investment immediately raised concerns among Scale AI clients, who fear that their proprietary data could be accessible to a direct competitor like Meta. This situation led to a customer exodus, with actors like Google, Microsoft, OpenAI, and xAI (Elon Musk’s company) reassessing their partnerships with Scale AI.
A structure designed to avoid antitrust scrutiny?
One of the most controversial aspects of this deal is its structure. By opting for a non-voting minority stake, Meta avoids the requirement to submit the transaction to automatic antitrust review in the United States. According to David Olson, professor of antitrust law at Boston College Law School, this approach offers “significant legal protection,” although the Federal Trade Commission (FTC) could still investigate if serious concerns emerge.
This strategy is not isolated. Experts note that major technology companies increasingly structure their investments to circumvent regulators. For example, the Department of Justice (DOJ) is currently examining a partnership between Google and Character.AI to determine whether it was designed to avoid antitrust scrutiny. The Meta-Scale AI transaction fits into a broader trend where tech giants secure access to key AI tools and talent without triggering full regulatory reviews.