01What it sells and who buys it
Surge supplies human-produced and human-evaluated data for training and assessing AI. Its customer base expanded from search, recommendation and content moderation into frontier AI research.
02How the business makes money
Revenue comes from enterprise contracts for data production and evaluation. The commercial product is high-quality human work delivered at scale, rather than a consumer subscription.
03Marketing and customer acquisition
Surge sold data quality before the generative AI boom and expanded with model-training demand.
- 01
Start with search and recommendation buyers
SourceEdwin Chen drew on machine-learning and moderation experience at large technology companies. Early buyers worked on search, recommendation and content moderation, providing demand before the generative AI boom.
- 02
Expand into model training and evaluation
SourceProjects such as GSM8K illustrated the company’s role in research datasets. As training demand expanded, matching people to tasks and controlling quality became central operational capabilities.
- 03
Financing the next stage
SourceSurge grew through operating revenue without a completed outside financing round in the cited account. Forbes reported $1.2 billion of 2024 revenue and substantial founder ownership. Its $24 billion company estimate is distinct from reported discussions about financing at $30 billion.
How the growth mechanism works
- Find existing data buyers
- Select and allocate workers
- Control evaluation quality
- Win repeat training work
This business does not need a mass audience of free app users. Delivery quality and reliability can lead to repeat research orders. Expert labor is part of the cost base, so software-like gross margins should not be assumed. Limited reliance on outside capital also shaped founder ownership.
04Funding and expansion
Surge grew through operating revenue without a completed outside financing round in the cited account. Forbes reported $1.2 billion of 2024 revenue and substantial founder ownership. Its $24 billion company estimate is distinct from reported discussions about financing at $30 billion.
More about the founding and key events
Edwin Chen founded Surge AI in 2020 after working on machine learning and content moderation at Google, Facebook and Twitter. The company sells training data and human evaluation. Its work included OpenAI’s GSM8K, a collection of 8,500 human-written and refined math problems used in model development.
Chen said early customers worked in search, recommendations and content moderation. Surge first sold to teams that already bought data. Its market expanded as demand for model training grew.
More workers alone cannot guarantee quality. Matching people to tasks and checking accuracy and consistency give customers reasons to order again. Forbes reported $1.2 billion in revenue for 2024 and estimated the 38-year-old Chen’s wealth at $18 billion in September 2026. Growing with limited reliance on outside capital also helped him retain a large ownership stake.
Surge and Mercor show that some AI lab spending flows into organizing expert work and managing its quality. That operating capability can become the central asset of a large business.