Three months after the Chinese AI developer DeepSeek made waves in the tech community with a groundbreaking model, 28-year-old AI executive Alexandr Wang appeared before Capitol Hill to advise lawmakers on maintaining the United States’ competitive edge in artificial intelligence.
During an April hearing, Wang emphasized the necessity for the U.S. to create a “national AI data reserve,” ensure robust power supply for data centers, and avoid an unwieldy mix of state regulations. Lawmakers appreciated his insights, with Republican Representative Neal Dunn of Florida noting, “It’s good to see you again here in Washington. You’re becoming a regular up here.”
As the CEO of Scale AI, Wang may not have the same recognition as OpenAI’s Sam Altman, but his influence in both technology and policy circles has notably increased. Scale AI employs a vast network of contractors to annotate data utilized by tech giants including Meta Platforms and OpenAI, facilitating the training and refinement of their AI models. The company is also increasingly recruiting professionals with advanced degrees, such as PhDs and nurses, to enhance the sophistication of its offerings. At its core, Scale is a major player in the essential trifecta of AI: chips, talent, and data.
The startup’s influence is poised to expand further as reports indicate that Meta is negotiating a multi-billion dollar investment in Scale. This funding could surpass $10 billion, positioning it among the largest private funding rounds in history. Scale was previously valued at around $14 billion in 2024, bolstered by investment from Meta.
Scale’s trajectory bears a resemblance to that of OpenAI. Both companies emerged about a decade ago, anticipating a pivotal moment in AI development, which Wang dubbed an “inflection point.” The two CEOs, who are friends and briefly cohabited, have become prominent figures representing the AI sector in discussions with Congress. Likewise, OpenAI has secured substantial investment from a major technology firm.
Scale’s growth has both influenced and been influenced by the AI boom initiated by OpenAI. Early in its history, Scale concentrated on labeling images for self-driving car models. It has since shifted focus to curating extensive text data needed for developing large language models, which form the backbone of chatbots like ChatGPT.
This work has not come without criticism, particularly regarding the overseas contractors in countries like Kenya and the Philippines who support AI development. Scale has faced scrutiny for relying on a significant number of contractors who receive relatively low wages for processing vast amounts of online data, with some workers reporting psychological distress. However, in a 2019 interview, Wang claimed that these workers earn “good” pay—within the 60th to 70th percentile of their local wage scales.
In a recent statement, Scale AI’s spokesperson, Joe Osborne, pointed out that the U.S. Department of Labor has concluded an investigation into the company’s labor practices without raising concerns regarding compliance with fair labor regulations.
The business of Scale AI is evolving in response to changing market needs. More technology companies are now exploring the use of synthetic, AI-generated data for training purposes, which could diminish the demand for Scale’s traditional services. However, leading AI research labs continue to struggle with acquiring sufficient high-quality training data necessary for creating advanced AI capable of complex task management.
To address this demand, Scale is increasingly sourcing higher-paid contractors with advanced degrees to enhance its AI systems. These specialists engage in a method known as reinforcement learning, which incentivizes correct answers and penalizes mistakes.
The experts working with Scale are responsible for developing complex problems for the AI models to tackle. According to a source familiar with the operations—who requested anonymity due to the sensitive nature of the information—12 percent of contributors engaged in refining these models possess PhDs, while more than 40 percent hold advanced degrees in fields such as law or business.
One area of focus includes improving AI models’ ability to address questions around tax law, which varies significantly across jurisdictions.
These strategic moves are contributing to substantial growth for Scale. The company reported approximately $870 million in revenue in 2024 and anticipates generating $2 billion in revenue this year. Following the emergence of DeepSeek, there’s been heightened interest in Scale’s network of experts as companies invest in developing models that can simulate human reasoning and undertake more intricate tasks.
Furthermore, Scale has strengthened its ties with the U.S. government through various defense contracts. Wang, a known critic of China, has cultivated relationships with lawmakers concerned about Chinese advancements in AI. Additionally, Michael Kratsios, a former executive at Scale, now serves as a key technology advisor in the Trump administration, playing a role in shaping U.S. AI policy.
A partnership with Scale could help Meta not only catch up with AI competitors like Google and OpenAI but also strengthen its connections with the U.S. government amid a greater focus on defense technology. For Scale, an alliance with Meta could prove advantageous, providing access to substantial resources.
Reflecting on his entrepreneurial journey, Wang once shared that when asked by a venture capitalist about his inspiration for starting a company, he humorously cited the film The Social Network, which chronicles the origins of Facebook.
© 2025 Bloomberg LP
(This story has not been edited by NDTV staff and is auto-generated from a syndicated feed.)
