Signs of Strain in Meta’s Partnership with Scale AI

Since June, Meta has poured $14.3 billion into Scale AI, appointing CEO Alexandr Wang and other distinguished figures to lead the Meta Superintelligence Labs (MSL). However, there are early indications that this collaboration may face challenges.

Wang has brought in Ruben Mayer, the former Senior Vice President of GenAI Product and Operations at Scale AI, to join MSL. Sources from TechCrunch suggest that Mayer’s stint at Meta lasted only two months.

Mayer joined with almost five years of experience at Scale AI, where he managed AI data operations teams under Wang’s direction. However, he was not part of TBD Labs, the central initiative aimed at achieving AI superintelligence, which has attracted several leading researchers from OpenAI.

Mayer has not responded to numerous requests for comments from TechCrunch.

Reports indicate that TBD Labs is collaborating with various external data providers, in addition to Scale AI, for future AI models. Companies like Mercor and Surge are believed to be involved in this effort.

Typically, AI labs work with multiple data vendors—Meta has engaged with Mercor and Surge even before TBD Labs was formed—making reliance on a single provider unusual. This has raised concerns among insiders, who claim that researchers at TBD Labs find Scale AI’s data lacking, prompting them to seek alternatives from Surge and Mercor.

Initially, Scale AI utilized a crowdsourcing approach with a large, cost-effective workforce for basic data labeling. As AI models have evolved, the focus has shifted toward specialized experts, such as doctors and scientists, to create and refine high-quality data that enhances model performance.

As Scale AI aims to attract these experts through its Outlier platform, competitors like Surge and Mercor have quickly established a market presence by enhancing their recruitment strategies to attract top talent from the outset.

A representative from Meta downplayed concerns regarding the quality of Scale AI’s data, while Surge and Mercor declined to comment. In response to questions about Meta’s reliance on competing data providers, a spokesperson from Scale AI characterized the initial investment announcement as a sign of a growing business partnership.

Meta’s partnerships with various third-party data vendors illustrate its intent to diversify data sources while also making significant investments in Scale AI. Conversely, Scale AI has faced its own difficulties: shortly after Meta’s major investment, both OpenAI and Google revealed plans to terminate their partnerships with the data provider.

After losing these clients, Scale AI laid off 200 employees from its data labeling team in July, citing changes in market demand, according to new CEO Jason Droege. He stated that Scale AI plans to recruit in other areas, particularly following the acquisition of a $99 million contract with the U.S. Army.

Initially, there were speculations that Meta’s investment in Scale AI was aimed at securing Wang—a key figure with substantial AI expertise since the company’s inception in 2016—thus enhancing Meta’s appeal to elite AI talent.

Beyond Wang, the overall significance of Scale AI for Meta remains uncertain.

A member of MSL noted that many leaders from Scale brought on by Meta are not part of the central TBD Labs team, much like Mayer. Additionally, it appears that Meta is not wholly reliant on Scale AI for its data labeling requirements.

Meanwhile, the atmosphere within Meta’s AI division has reportedly grown more turbulent since Wang’s arrival and the influx of several leading researchers, according to two former employees and one current MSL member. New hires from OpenAI and Scale AI have expressed concerns over the bureaucratic nature typical of larger organizations, implying a diminishing influence of Meta’s former GenAI team.

These factors suggest that Meta’s largest AI investment to date confronts a unique set of difficulties as the company seeks to address its AI development challenges. Following the lukewarm launch of Llama 4 in April, sources report that Meta CEO Mark Zuckerberg expressed dissatisfaction with the AI team’s progress, according to insights from one current and one former employee shared with TechCrunch.

In light of these challenges and to maintain competitiveness against entities like OpenAI and Google, Zuckerberg has implemented decisive measures to foster partnerships and initiated an aggressive recruitment strategy aimed at attracting top AI talent.

Alongside Wang, Zuckerberg has successfully recruited distinguished AI researchers from OpenAI, Google DeepMind, and Anthropic. Meta has also acquired AI voice startups like Play AI and WaveForms AI, forming alliances with Midjourney, a firm specializing in AI image generation.

To further its AI goals, Meta has publicly announced plans for substantial data center expansions across the U.S., including a $50 billion facility in Louisiana named Hyperion, which draws inspiration from a titan in Greek mythology recognized as the father of the Sun God.

Wang, who lacks formal academic credentials in AI, is viewed as an unconventional choice to head an AI lab. Reports suggest that Zuckerberg considered more traditional candidates for the role, including OpenAI’s chief research officer, Mark Chen, and founders of startups established by Ilya Sutskever and Mira Murati—none of whom accepted the offers.

Recent reports from Wired indicate that several AI researchers brought in from OpenAI have already left Meta. Additionally, numerous longtime members of Meta’s GenAI division have departed amid recent organizational changes.

Rishabh Agarwal, a researcher at MSL, recently announced his resignation, stating on X this week that he will be leaving the organization.

“The pitch from Mark and @alexandr_wang to build in the Superintelligence team was incredibly compelling,” Agarwal noted. “However, I ultimately adhered to Mark’s advice: ‘In a world that’s changing so fast, the biggest risk you can take is not taking any risk.’”

When asked about his experiences at Meta and his decision to leave, Agarwal opted not to provide a comment.

Recently, Chaya Nayak, Director of Product Management for Generative AI, and Rohan Varma, Research Engineer, also announced their exits from Meta. The pressing question now is whether Meta can stabilize its AI operations and retain vital talent for future success.

MSL is currently focused on creating its next-generation AI model, with Business Insider reporting an anticipated launch by the end of the year.