After putting his child to sleep, Alex Ratner was preparing to work for a few more hours when he suddenly learned that Meta Platforms Inc (META.US), the parent company of Facebook, would invest $10 billion in Scale AI, a leader in the AI data labeling field. Within minutes, the head of Snorkel AI received several calls from the board discussing how to win over clients concerned about Scale AI's operational independence, especially super clients like Google, Microsoft, and OpenAI.
Then came the inquiries from potential clients—he mentioned that there were dozens in just one day, and potential deals worth &#34;tens of millions of dollars&#34; had formed over the past week, currently in various stages of negotiation. &#34;Every responsible large language model (LLM) developer will take significant actions to diversify their data service providers,&#34; Ratner stated, &#34;There will be major shifts and significant opportunities in the market.&#34;
Undoubtedly, OpenAI is one of them. This AI unicorn, which has consistently ranked among the top global AI model developers due to the global popularity of ChatGPT, indicated that it is gradually halting its business collaboration with Scale AI. Although Scale AI accounts for only a small portion of OpenAI's data labeling contracts, this move highlights the new challenges Scale AI faces following Meta's massive investment.
Last week, Meta announced an investment of up to $14.3 billion in the AI unicorn Scale AI, acquiring a 49% stake, which pushed the valuation of this AI startup to over $29 billion. A major highlight of this significant deal in the AI sector is that Scale AI's CEO Alexandr Wang will officially join Meta's core R&amp;D team to lead the &#34;Super Intelligence&#34; team focused on general artificial intelligence.
This massive investment has sparked a surge in demand and interest from major clients in the cloud computing and AI technology sectors for other AI data labeling service providers, as they worry that Meta will increase the technical visibility of its AI development process, leading to a significant rise in demand for data labeling services from companies like Labelbox and Turing.
Scale AI's Competitors Welcome &#34;Windfall&#34;
For years, Scale AI has been one of the most well-known companies helping large tech firms and cloud computing giants label and annotate AI training data, attracting major clients like Meta and OpenAI, and ranking among the highest-valued AI startups. However, last week, Meta acquired a 49% stake in Scale AI for up to $14.3 billion and poached the company's CEO Alexandr Wang to lead Meta's new &#34;Super Intelligence&#34; department, a deal that could reshape the competitive landscape of the AI data labeling field.
Now, Scale AI's strongest competitors, including Snorkel AI, Turing, Invisible Technologies, Labelbox, and tech companies like Uber Technologies Inc., which leads in ride-hailing and Robotaxi services, are all striving to provide data-related services
To meet the insatiable data demands of AI large models and AI application developers. Several companies have reported a surge in customer interest and demand following Meta's investment in Scale AI, with some clients concerned that Meta will gain deeper insights into their AI development processes and detect proprietary technologies through data labeling.
&#34;The demand is unprecedented,&#34; said Manu Sharma, founder and CEO of Labelbox. The company has received investments from SoftBank Group led by Masayoshi Son and institutions like Andreessen Horowitz. &#34;We have collaborated with the largest AI labs, and now the remaining AI super players are actively reaching out to us; we expect to gain more business.&#34;
Scale AI spokesperson Joe Osborne stated in a statement that the company remains independent, and customer trust is still one of its &#34;most valuable assets.&#34; He added, &#34;Our commitment to protecting customer data and ensuring customer success has not changed at all.&#34; Scale AI declined to comment on OpenAI's gradual termination of cooperation.
Before receiving Meta's substantial investment, Scale AI was already facing uncertainty. According to insiders, Scale's revenue for 2024 is expected to be around $870 million, falling short of the $1 billion annual target. However, Osborne stated that the company anticipates a 160% year-on-year revenue growth in 2024, with new business exceeding $1.5 billion last year.
As AI training/inference systems rapidly evolve, some companies, including the superstar clients listed on Scale AI's official website, are beginning to reassess their reliance on the company's data labeling services following Meta's significant investment.
An OpenAI spokesperson stated that to support increasingly advanced AI large models, OpenAI has long sought more specialized and diversified data service providers. Even before Meta's investment, OpenAI was gradually reducing its reliance on Scale AI's data labeling services.
What exactly is Scale AI?
Founded about ten years ago, Scale AI initially relied on a large number of contractors to label text and images for early AI developers. Today, the company increasingly hires higher-paid labelers with PhDs and advanced degrees in fields such as nursing, biotechnology, and physics to help develop more complex AI large models, but competition in this field is becoming increasingly fierce.
Scale AI was founded by Alexandr Wang in 2016, initially providing data labeling services for large machine learning/AI systems. Its leading global data labeling technology supports the development of AI large models and AI developer ecosystems for large clients like Microsoft (MSFT.US) and OpenAI, and it has emerged as a key player in the global generative AI wave since 2023. The company previously achieved an overall valuation of approximately $14 billion in its latest round of financing in 2024, with major investors including Meta and Microsoft
Scale CEO Alexandr Wang may not be as well-known as OpenAI's Sam Altman or Anthropic founder Dario Amodei, but his company has become the absolute leader in the data sector among the three pillars of AI—chips, talent, and data. This startup provides data annotation services necessary for AI model training to tech companies like Meta and OpenAI through a large outsourced service team, and assists in developing customized AI applications.
Large tech companies and cloud computing giants typically collaborate with multiple data and related annotation service providers to build and fine-tune their AI large models. For instance, insiders say that OpenAI needs more specialized data to support the development and application of different AI large models and has begun to turn to other companies like Mercor.
In the initial phase after the announcement of the agreement, OpenAI CFO Sarah Friar stated that the company plans to continue collaborating with Scale AI. &#34;We don't want to completely freeze the ecosystem, as mergers and acquisitions are inevitable,&#34; she said.
However, an OpenAI spokesperson indicated that over the past 6 to 12 months, OpenAI has determined that Scale AI is not the best choice, as the specialized data it requires exceeds what Scale AI can provide, and OpenAI is more focused on diversifying service providers to mitigate single risks. OpenAI is shifting towards building more advanced AI large models, such as those capable of simulating human reasoning processes and agent-based models that can complete tasks with very limited input. Forbes reported that OpenAI has been reducing its use of Scale AI's data annotation services for several months.
OpenAI's major shareholder Microsoft is also listed on Scale AI's official client list, but insiders say that Microsoft currently has no large contracts with Scale AI, and these small transactions will not change due to Meta's investment. Microsoft declined to comment.
Scale AI is also gradually diversifying its services, such as directly helping companies build customized AI application software and collaborating more closely with the defense industry. For Meta, which aims to expand in this field, defense business may also hold long-term value.
While the impact of Meta on Scale AI's business remains to be seen, Scale AI and its competitors generally agree on one point: this deal will bring more attention to another long-overlooked corner of the AI market.
Compared to new AI agents and scarce AI computing power chips, data work has historically been &#34;overlooked,&#34; said Olga Megorskaya, co-founder and CEO of AI data annotation company Toloka AI. The company was spun off from cloud service provider Nebius, with major clients including Microsoft, Amazon AWS, and Anthropic. &#34;This deal symbolizes that the industry is beginning to recognize the importance of human data training for AI large models,&#34; she said
Investing in Scale AI is a stroke of genius, likely to help Meta's stock price embark on a long-term bull market trajectory
Analyst Geneva Investor from the investment research platform Seeking Alpha recently stated that the substantial investment by Meta, the parent company of social media platforms Facebook and Instagram, in the AI startup Scale AI may significantly enhance its exposure to AI-related businesses and provide a &#34;positive catalyst&#34; for the stock price of this tech giant led by Mark Zuckerberg to enter a long bull market.
The acquisition of Instagram and WhatsApp over a decade ago has already proven Meta's keen investment insight. Jonathan Weber, an opinion leader from the investment community Cash Flow Club, stated, &#34;The management has performed excellently in identifying and investing in quality targets in the past, including the acquisition of Instagram over a decade ago, so I believe the risk of overpaying is not significant.&#34;
Digital advertising is the core revenue engine for Meta, and its 3 billion users serve as a cornerstone. Meta's AI advertising assistance tools and Meta AI have consistently helped the company's advertising revenue exceed expectations for several quarters. In the field of digital advertising, where Meta relies heavily, the powerful open-source AI large models launched by Meta, along with various generative AI-assisted software tools, enable advertisers to reach a broader potential user base, providing a new AI-based advertising recommendation experience for Meta's advertisers and users. This is also an important logic behind Wall Street analysts' general expectation that Meta's stock price will continue to rise.
Therefore, for Meta, in the context of global companies vying for dominance in AI applications, Scale AI may help Meta create &#34;killer&#34; AI applications. As a leader in global data labeling and model evaluation, Scale AI holds the &#34;data production materials&#34; of the AI era. Scale AI can provide industry-leading data labeling and model evaluation platforms, while Meta is eager to commercialize the Llama series of large models on a large scale and deeply embed Meta AI into its social, advertising, and hardware ecosystems. Scale AI can be seen as the final key piece of Meta's &#34;AI moat&#34;: the integration of computing power (NVIDIA AI GPUs + self-developed ASICs + large-scale data centers covering the globe), large models (Llama family), and data (Scale AI)

Zhitong

Meta Platforms

OpenAI

GraniteShares 2x Long NVDA Daily ETF

SoftBank

Alphabet

Alphabet - C

Amazon

YieldMax NVDA Option Income Strategy ETF

NVIDIA

Microsoft

Meta's massive investment in Scale AI triggers a chain reaction: surging demand in the AI data labeling market