Meta is planning to invest over 10 billion dollars (about 13 trillion won) in the American startup Scale AI to enhance artificial intelligence (AI) technology. This is Meta’s largest investment in an external company and is a noteworthy example of large-scale funding for a privately held company.
Scale AI is a startup specializing in ‘data labeling,’ which is one of the core aspects of machine learning technology. It processes unstructured data like images, texts, voices, and videos into a form that can be analyzed. This process is a preparatory step essential for AI models to learn and is crucial for advancing AI technology.
Scale AI categorizes data manually and combines it with an automated system to enhance accuracy. Its major clients span across various industries, including autonomous vehicles, healthcare, finance, and e-commerce. The key platform, ‘Scale Annotator,’ is designed to efficiently handle large volumes of data.
Scale AI was founded in 2016 by Alexander Wang, who was 19 at the time. The company achieved a revenue of approximately 870 million dollars in 2024 and aims to surpass 2 billion dollars in 2025. In May 2024, it secured 1 billion dollars from Amazon, Meta, and others in a Series F investment round, which valued the company at 13.8 billion dollars. It is also considering a public tender offer to raise its current valuation to 25 billion dollars.
Meta has focused on internal AI research and open-source model development until 2023. However, as competitors increasingly collaborate with external AI startups, a strategic shift became necessary. Microsoft invested over 13 billion dollars in OpenAI, and both Amazon and Alphabet, Google’s parent company, have invested billions into the AI startup Anthropic.
Consequently, Meta has declared AI as its top management priority from 2024 and plans to invest 65 billion dollars in AI projects by 2025. The construction of a 2-gigawatt large data center in Louisiana, USA, is part of this plan. Meta is also undergoing organizational restructuring and expanding its recruitment of machine learning specialists.
Meta and Scale AI have previously collaborated. Both companies developed the ‘Defense Llama,’ a large language model for defense purposes based on Meta’s AI model architecture, ‘Llama3.’ It was designed to assist in military operation planning and threat analysis. This collaboration is considered a successful example of establishing technological credibility for the sake of U.S. national security.
Meta’s AI-focused strategy is impacting its business results as well. Its advertising revenue for the fourth quarter of 2023 increased by 24% from the previous year to 38.7 billion dollars, thanks to enhanced precision in targeted advertising achieved through AI technology. The company is also rapidly expanding its AI assistant service, with monthly active users surpassing 700 million by the end of 2024.
The negotiation with Scale AI is interpreted as Meta’s attempt to boost its technology and market competitiveness by partnering with external expertise, acknowledging the limitations of relying solely on internal technology development. If the deal is successful, Meta’s position within the AI ecosystem is expected to strengthen considerably.