Surge AI seeks $1B at $15B+ valuation in first-ever raise as Scale AI loses customers and CEO to Meta

Surge AI is raising its first-ever round of funding—and it’s swinging big.

The data-labeling AI startup, known for staying quiet while racking up serious revenue, is now in talks to raise up to $1 billion, according to an exclusive report from Reuters, citing sources familiar with the matter. The raise, which would include both primary and secondary capital, could push Surge’s valuation north of $15 billion.

“Surge AI, a data-labeling firm that competes with Scale AI, has hired advisors to raise as much as $1 billion in the first capital raising in the firm’s history,” Reuters reported.

Founded by former Google and Meta engineer Edwin Chen, Surge has operated under the radar since 2020. Chen bootstrapped the company from the start, and it’s been profitable—an outlier in an industry built on venture capital. Last year alone, Surge pulled in over $1 billion in revenue, outperforming Scale AI, which brought in $870 million during the same period.

If that wasn’t enough, the timing couldn’t be better. Surge is picking up customers at a time when Scale AI is losing them fast.

Meta’s recent $28.6 billion valuation of Scale AI came with a major twist: it snapped up a 49% stake and took its CEO, Alexandr Wang, to lead Meta’s newly launched Superintelligence Labs. That move spooked some of Scale’s biggest clients, including Google and OpenAI, who are now reportedly backing away from the platform. The fear? That their data and research priorities might end up feeding Meta’s next big model.

Scale insists its business remains strong and that customer data is protected. But that hasn’t stopped competitors like Surge from capitalizing on the moment.

Surge has quietly built a reputation for high-end data labeling work, used by major AI labs like Google, OpenAI, and Anthropic. As reinforcement learning from human feedback (RLHF) becomes a bigger part of AI training, companies are leaning harder on well-structured, high-quality datasets. Surge’s model, which relies on a vetted network of skilled contractors instead of crowdsourced labor, has drawn interest from those who need precision over volume.

Now, with investors circling, the company is about to test how much appetite there is for the data-labeling space. Some investors still see it as a long-term necessity for training and fine-tuning AI systems. Others worry that as models get smarter, the need for human-labeled data could shrink—or get automated away.

Surge declined to comment.

But make no mistake: if this raise goes through, it could mark a shift in where serious AI infrastructure money is flowing. Not just to model builders, but to the companies behind the scenes, feeding them the data they need to work.