Google plans to terminate its partnership with Scale AI, an artificial intelligence data services company. This decision stems from Google’s concerns that Meta Platforms’ massive investment in Scale AI could lead to the leakage of its own trade secrets. As a leading enterprise in the field of AI data annotation, Scale AI provides data annotation and human feedback services required for model training to nearly all major AI companies through its global network of contractors. Google originally planned to pay approximately $200 million to Scale AI this year to purchase manually annotated training data, which is crucial for developing advanced artificial intelligence models such as Gemini.
Sources stated that after Meta announced the acquisition of a 49% stake in Scale AI, Google has held talks with several competitors of Scale AI this week, preparing to transfer most of its workload away from Scale AI. Scale AI’s valuation soared to $29 billion after Meta took a stake, compared to only $14 billion before that. Nevertheless, losing Google as a key client could deal a severe blow to its business, as its core revenue is highly dependent on a small number of major clients. Three sources indicated that Google began seeking diversification of data service providers over a year ago, but Meta’s stake announcement this week accelerated Google’s determination to fully terminate key contracts with Scale.
As part of Meta’s share acquisition plan, Scale AI CEO Alexandr Wang will join Meta along with a small number of employees, while Scale AI will continue its existing business operations. A spokesperson for Scale AI said in a statement that the company maintains partnerships with multiple large enterprises and government agencies, its business remains strong, and it is committed to protecting customer data, but declined to comment on specific matters related to Google.
It is reported that Scale AI achieved $870 million in revenue in 2024, with Google accounting for approximately $150 million of that. Other tech giant clients, such as Microsoft, are also gradually reducing their cooperation. A source said that Elon Musk’s xAI also plans to withdraw. Another insider stated that OpenAI began to downplay its cooperation with Scale a few months ago, although its spending is far lower than Google’s. In response, OpenAI’s chief financial officer recently stated that it will continue to regard Scale as one of its data suppliers.
Due to commercial sensitivity and technical security considerations, many AI laboratories are now worried that cooperating with Scale, which is partially owned by Meta, may indirectly expose their own R&D priorities, roadmaps, and even technical secrets to competitors. Jonathan Siddharth, CEO of Turing, a competitor of Scale, said: “The deal between Meta and Scale is a watershed.” Manu Sharma, CEO of another competitor Labelbox, told the media that by the end of this year, the company expects to gain hundreds of millions of dollars in new revenue from clients “fleeing” Scale.
The change in the relationship between Google and Scale AI not only reflects the intensified competition in the data service field within the AI industry but also highlights enterprises’ high emphasis on data security and trade secret protection. With the rapid development of AI technology, data has become one of the core assets of enterprises. How to ensure the security of data in the process of use and cooperation will become an important factor affecting enterprises’ strategic decisions.
