Proximal raises $15 million seed round, claims $200 million in annualized revenue
The AMW Read
A new training-data entrant's claimed revenue meaningfully updates the player map, while its segment-level impact depends on verification of the operating figures.
Proximal raises $15 million seed round, claims $200 million in annualized revenue
Proximal has emerged from stealth with a $15 million seed round led by General Catalyst, according to The Information. The company uses AI to generate and evaluate coding tasks for model training. The round was reported at a $300 million valuation. Proximal's CEO says the business has surpassed $200 million in annualized revenue and is profitable. The available account does not specify the period used to annualize revenue, the revenue definition, or the customer mix.
Proximal sits in the training-data and evaluation layer of the AI market, where model developers need coding tasks that can both teach and test model behavior. Its combination of task generation and evaluation puts the quality of those assessments at the center of its value proposition: more examples help only if the evaluations reliably distinguish better outputs. The claimed revenue would be striking for a company emerging from stealth, but the account does not establish how much is recurring or whether its evaluations predict performance on coding work beyond its own tasks.
For builders buying coding data, the concrete test is whether Proximal's tasks improve models on independent coding evaluations. For investors, the CEO's annualized figure calls for a bridge to recognized revenue, repeat purchases, customer concentration, and margins before it can be treated as a durable run rate. The seed round and reported valuation provide financing context; neither verifies the operating claims.
