Bootstrapped Surge AI quietly surpassed Scale AI with $1B+ revenue and zero outside funding

Summary

Surge AI, founded by Edwin Chen in 2020, has become the largest data labeling company by revenue, generating over $1 billion last year without raising any outside funding. The company surpassed Scale AI, which generated $870 million in revenue during the same period, despite operating with just 110 employees across New York and San Francisco offices versus Scale's thousand-plus headcount.

Surge has been profitable from inception. Scale, which raised $1.5 billion in venture capital, was not profitable in 2024, though it retained roughly $1 billion in cash reserves and was not in financial distress. Chen's bootstrapped operation outgrew the venture-backed alternative by nearly $200 million in annual revenue.

Meta's $14.3 billion acquisition of a 49% stake in Scale brought CEO Alexander Wong into a senior role at Meta. OpenAI has stopped using Scale for data annotation over concerns that Scale's alignment with Meta, a competitor in open-source AI, creates a conflict of interest. Other foundation model labs are likely to follow. Surge, untethered to any corporate parent, stands to capture that displaced business.

Chen's path to building without capital

Chen studied linguistics and math at MIT, then worked as a machine learning engineer at Facebook, Dropbox, Google, and Twitter. He saw firsthand how data labeling broke down at scale. At Facebook, his team needed 50,000 accurately labeled businesses for a Yelp competitor prototype. An outside vendor took six months and delivered corrupted data, with restaurants labeled as coffee shops and coffee shops as hospitals. That failure showed him that data labeling quality was broken and no one was solving it well.

When Chen left Twitter in 2020 to start Surge, he recruited contractors he already knew and self-funded the company using his savings. He made a crucial strategic bet on language modeling just as that field accelerated. Scale started in visual data for autonomous vehicles and had to pivot later.

Chen's positioning was deliberate. Surge charges two to five times what Scale bills, justifying the premium through quality. A former Scale employee confirmed that Surge often outperformed Scale in customer audits. Within months, OpenAI hired Surge to fine-tune models on harmful-response detection. By 2022, Anthropic became a customer. Both companies later published research papers featuring Surge's work, all while the startup remained largely invisible to the tech press and venture ecosystem.

Margin structure and operational discipline

Surge operates as a marketplace. It takes contracts from foundation model labs, then pays a network of contractors through a subsidiary, Data Annotation Tech, which advertises $20 per hour starting wages. The actual margin profile is opaque. If a $1,000 contract results in $800 flowing to contractors, Surge keeps $200. Even at that split, $200 million in gross margin on $1 billion in revenue is a substantial and self-sustaining business.

Scale spent heavily on headcount and marketing. Wong became a conference circuit regular and X personality. Surge stayed quiet. Chen had no investors demanding quarterly updates, no board pressuring for top-line growth at any cost, and no need to build hype around future funding rounds.

If Surge were to raise capital or go public at a valuation comparable to Meta's $14.3 billion investment in Scale, Chen would become a billionaire many times over on paper. So far, he hasn't needed to.