Cover Lucy Guo (right, photographed with Kate Goodlad) co-founded Scale AI with Alexandr Wang. Though she left Scale in 2018, her continued stake in the company has made her the youngest self-made female billionaire in the world, surpassing musician Taylor Swift (Photo: Jack Taylor/Getty Images for SXSW London)
Lucy Guo (right, photographed with Kate Goodlad) co-founded Scale AI with Alexandr Wang. Though she left Scale in 2018, her continued stake in the company has made her the youngest self-made female billionaire in the world, surpassing musician Taylor Swift (Photo: Jack Taylor/Getty Images for SXSW London)
1 / 1
The multibillion-dollar company Scale AI has become essential to the global AI boom—while raising big questions about ethics and trust.
Once a modest start-up launched by two former Quora employees, Scale AI has quietly become one of the most influential players in the global artificial intelligence race. Founded in 2016 by Alexandr Wang and Lucy Guo, the San Francisco-based company powers the world’s most advanced AI systems—from OpenAI’s ChatGPT to Google’s Gemini and Meta’s generative models. It also supplies critical data to industry heavyweights like Microsoft, Toyota and the US Department of Defense.
Behind the scenes, Scale is facing growing scrutiny: a new US$14.3 billion deal with Meta has sparked concerns about data security and corporate trust, especially after a recent data breach exposed confidential documents involving top clients. As AI continues to shape everything from how we work to how we live, understanding the companies behind it—like Scale—has never been more urgent.
See also: From singularity to artificial intellgence agents: Sam Altman on how AI will be in 2025 and beyond
What does Scale AI really do?
Above Short video of life at artificial intelligence data company Scale AI
Scale AI’s core business is the provision of high-quality training data for machine learning systems. Its services—data labelling, annotation and curation—are fundamental to how AI models are trained and refined. Because this work is labour intensive, Scale heavily depends on a global network of gig workers, particularly through its Remotasks platform, to manually classify vast volumes of data that algorithms require to function effectively.
OpenAI uses Scale to train its language models, while Google relies on Scale’s annotations for Gemini. Meta, Microsoft, Toyota and Uber also depend on Scale’s carefully labelled datasets. For the US Department of Defense, Scale has supported projects involving satellite imagery analysis and conflict zone assessments.
The company has also become instrumental in model safety and alignment. Its Safety, Evaluation and Alignment Lab (SEAL) helps clients ensure that AI systems behave according to human values—an increasingly crucial function as generative AI becomes embedded in everyday applications, from virtual assistants to content creation tools.
A US$14 billion Meta deal—and a leadership change
Scale reached unicorn status by 2019 following a US$100 million funding round. By 2021, its valuation had climbed to US$7 billion as it expanded beyond annotation into enterprise and defence sectors. That figure rose to US$13.8 billion in 2024 and, as of 2025, has reached US$29 billion. Its co-founder Lucy Guo became the youngest self-made woman billionaire in the world, achieving the milestone at age 30.
Also read: What Studio Ghibli thinks of your AI generated cartoon photos
But Scale’s rapid ascent has not been without controversy. Its work with the Pentagon has prompted concerns about the potential use of AI in military operations. While the company has clarified that its defence projects focus on decision-making and planning, critics fear that such technology could eventually be applied to autonomous weapons.
Scale has also been criticised for its labour practices. Numerous complaints from Remotasks contractors cite underpayment, late payment and non-payment issues. In response, the US Department of Labor launched an investigation into the company’s wage practices. In January 2023, Scale laid off around 20 percent of its workforce, citing broader challenges in the tech sector.
The $14B deal with Meta
Above Scale AI's CEO and co-founder Alexandr Wang discussing artificial intelligence at the Web Summit 2025 in Qatar (Photo: Web Summit Youtube channel)
Scale AI's CEO and co-founder Alexandr Wang discussing artificial intelligence at the Web Summit 2025 in Qatar (Photo: Web Summit Youtube channel)
1 / 1
On 12 June 2025, Scale announced a major development: Meta has invested US$14.3 billion into the company, expanding their strategic partnership to accelerate Meta’s AI efforts. As part of the deal, Scale CEO Alexandr Wang will step down and join Meta’s AI division, though he will remain on Scale’s board. An interim CEO will lead the company in the meantime.
However, this partnership may come at a cost. Google—formerly Scale’s largest customer—has begun exploring alternative vendors due to concerns over Meta’s influence. Microsoft and OpenAI may also reevaluate their engagements. The move has sparked industry-wide questions about data privacy, client trust and potential conflicts of interest, especially as proprietary data shared with Scale could indirectly benefit its rivals.
Security breach exposes internal documents
The most recent issue surrounding Scale is the exposure of sensitive data. In a report published by Business Insider last June 24, an investigation revealed that Scale AI inadvertently exposed sensitive information through publicly accessible Google Docs links. The breach involved at least 85 documents spanning thousands of pages, containing confidential project guidelines and internal evaluations related to AI training work for its clients such as Meta, Google and Elon Musk’s xAI.
The security lapse extended beyond project materials to include contractors' personal information. According to the report, publicly viewable spreadsheets contained workers' private email addresses, details about payment disputes and confidential performance evaluations. Some of the files that were publicly accessible contained details about Scale’s work on Google’s AI chatbot, training data for Meta and a confidential xAI project. In response, Scale has locked down the files and is launching a thorough investigation of the matter.
Scaling up
The company's market position remains strong despite recent competitive challenges. While Google and other major customers are reportedly backing away following the Meta deal, Scale AI's diversified client base and essential role in AI development suggest continued growth potential.
The data annotation as a whole might experience growth as other companies turn to Scale’s rivals. There might also be a shift towards in-house data labeling in some AI companies. Only time will tell if Scale’s deal with Meta proves to be profitable as for the company as a whole.
