Meta targets Scale AI in big AI push

Social media company Meta is reportedly engaged in talks to make a multi-billion dollar investment into Scale AI, a key player in the data labeling sector for AI models. According to a report from PYMNTS, the potential deal could exceed $10 billion, making it Facebook parent company’s most substantial external AI commitment so far.

Scale AI provides data labeling services that enable companies like Microsoft and OpenAI to train their AI models. This process involves preparing and annotating vast quantities of images, text, and other data, making it usable for AI algorithms. The company’s operations heavily rely on contract workers for this data preparation. Scale AI is positioned as a key infrastructure provider within the rapidly expanding generative AI ecosystem.

The company has demonstrated decent financial growth, reporting $870 million in revenue last year and projecting sales to exceed $2 billion this year. Scale AI’s most recent valuation, following a $1 billion Series F funding round in 2024 (which included Meta), stood at approximately $13.8 billion. Earlier this year, reports indicated discussions that could elevate its valuation to $25 billion.

Meta is also expanding its internal AI infrastructure, including the development of a 2-gigawatt data center and holding approximately 350,000 Nvidia H100 chips for its AI projects. The company is also developing its own hardware for AI advancement, reinforcing its ambition to lead the global AI race. A successful investment could provide Meta with greater access to Scale AI’s evolving data classification tools, potentially strengthening its competitive standing in the global AI landscape, especially after the perceived underwhelming reception of its Llama 4 launch earlier this year. This strategic move could grant Meta a critical advantage by securing access to specialized datasets and refining its own data pipeline, thereby improving the training processes for its AI models.

Historically, Meta has primarily focused on in-house AI research and a more open development approach for its AI technology, exemplified by its Llama large language models. The Meta AI chatbot, for example, reportedly serves one billion users on a monthly basis. This potential investment in Scale AI contrasts with its previous internal emphasis and mirroring the significant investments made by other major tech firms like Microsoft (over $13 billion in OpenAI) and Amazon and Alphabet (billions in Anthropic). Unlike its peers, Meta does not operate a public cloud business, which influences the structure of its AI investments, potentially involving more direct cash outlays rather than cloud computing credits.

Both Meta and Scale AI have engaged with the defense sector. Meta recently disclosed a new partnership with defense contractor Anduril to develop tools for the US military, including an AI-powered helmet with integrated virtual and augmented reality features. Meta has also granted approval for U.S. government agencies and defense contractors to utilize its AI models. Scale AI has been working with the US government on AI services for defense purposes, including a contract with the Department of Defense to develop AI agents.