Scale AI Cuts 14% of Staff: The Turning Point for the 'Behind-the-Scenes' Company of the Generative AI Era
Data labeling is an essential 'behind-the-scenes' function that has supported the development of generative AI. However, in July 2025, Scale AI, a leading data labeling startup, announced the layoff of 200 employees—representing 14% of its workforce—and the termination of contracts with 500 contractors worldwide. The backdrop to this is the rapid expansion of the company's core business and a shift in its business structure following a massive $1.43 billion (approximately 230 billion yen) contract with Meta.
In this article, we will analyze the changes in the generative AI industry, the limitations of B2B AI companies, and the future direction of strategic shifts through the latest developments at Scale AI.
1. What is Scale AI: The Unsung Hero of AI Labs
1-1. Core Business: Providing 'Training Data' Essential for AI Development
Scale AI is a U.S. startup founded in 2016 that specializes in creating structured data necessary for machine learning. For example, when developing image recognition AI for autonomous vehicles, it is necessary to accurately label images as 'this object is a human' or 'this is a traffic light.' Scale AI's core business is streamlining this 'annotation' work using a combination of massive human labor and systems.
The company is known for providing services to AI giants such as OpenAI, NVIDIA, and Meta, and has grown rapidly as the 'oil of AI'—a source of data supply.
2. Background of the Layoffs: The Rebound from Rapid Expansion and Loss of Customers
2-1. CEO Change and Shift in Business Policy
According to a Bloomberg report, the layoffs were revealed in an internal memo from Scale AI's interim CEO, Jason Droge. He stated, 'We expanded our core data labeling business too rapidly,' and emphasized the restructuring of the organization and a focus on sales to enterprises and governments.
This statement suggests that the existing scale-type B2B service model has reached saturation, forcing a shift toward government projects and the enterprise market, where higher unit prices and long-term contracts can be expected.
2-2. The Side Effects of the $1.43 Billion Contract with Meta
Just one month before the layoffs, Meta signed a massive $1.43 billion contract with Scale AI, effectively poaching its founder, Alexandr Wang. This could be considered a de facto 'reverse acquihire,' a strategy by Meta to absorb both talent and technology.
However, the public announcement of this deal likely caused other major clients of Scale AI—such as OpenAI and Anthropic—to worry about competitive conflicts and terminate their contracts. As a result, the core business shrank significantly, forcing a reorganization of the company.
3. Comparison with Other Companies: Similar Patterns Seen in Inflection AI
Scale AI's case shares similarities with Inflection AI, which was acquired by Microsoft in 2024. Although Inflection was also a startup with its own independent AI platform, it ultimately ended up having its talent and IP (intellectual property) absorbed by a major tech company.
Such cases highlight the structural challenge that, amidst the generative AI boom, startups tend to converge on 'talent value over technology,' and that data supply models are difficult to differentiate.
4. Future Outlook: Where is Scale AI Heading?
4-1. Focus on AI Solutions for the Government
Scale AI already provides AI solutions to the U.S. Department of Defense and is expected to focus on similar government projects in the future. Especially in the areas of national security and defense, pre-labeled, highly reliable datasets remain highly valuable.
4-2. From Product-Based to Relationship-Based
It has become inevitable to shift the focus from the traditional product-based business of 'selling structured data' to partnerships with customers. The scale business model based on one-off contracts has reached its limit, and BtoG (government-facing) business, where customizability and reliability are key, will likely become the main battlefield.
Conclusion: Urgent Need to Reconstruct Positioning in the AI Ecosystem
Scale AI's layoffs symbolize a structural turning point for the 'behind-the-scenes' companies that have supported the generative AI boom. Now that the foundational function of data labeling is becoming difficult to sustain as a standalone business, Scale AI must redefine 'which ecosystem it will engage with and in what capacity.'
As Meta and Microsoft continue to 'hoard AI talent,' it will be interesting to see how long independent startups can maintain their autonomy.
Recommended Articles
Next Big Wave (Growth Stocks, Seeds of Ideas, Deep Dives into Trends)
Unraveling the secrets of growth, from seed startups to legendary companies
