Applied Compute raises $350 million at $3.25 billion valuation

Applied Compute, a San Francisco start-up founded by three former OpenAI researchers, has raised $350 million at a valuation of $3.25 billion. The company announced the round on Monday, 2 November 2026, and said it had passed $100 million in annualised revenue run rate less than 18 months after it was founded. The round was led by the angel investor Elad Gil, with Nvidia, Shopify Ventures, HubSpot Ventures and more than a dozen others taking part.

Most companies that use AI today rent general-purpose models, the same ones that millions of others use. Applied Compute's pitch is different: it trains a model for each customer, on that customer's own data, so the model is tuned to that company's work. Its founders worked on OpenAI products including the Codex programming assistant and the o1 reasoning models, and the company uses reinforcement learning, a training method in which a model improves by being rewarded for good results and penalised for bad ones.

The round shows how quickly investors now reward AI start-ups that can show real revenue. Reaching a $100 million run rate within 18 months is fast by any standard for business software. An annualised run rate takes recent monthly revenue and multiplies it to a full year, so it shows the current pace of sales, not audited annual revenue, and both the run rate and the valuation are the company's own figures.

The investors are also telling. Nvidia supplies the chips on which such models are trained, and Shopify and HubSpot are large software companies investing through their venture arms. Backing from both chip and software companies shows how widely custom AI for businesses is being watched.

For Indian founders and IT services firms, the model is worth watching. Building and tuning AI for a client's own processes is close to the work India's technology services industry does today. A start-up that does this with a small team and its own training methods points to how that work may change, and to where global investors see the openings.

Peepals calculation: At a $3.25 billion valuation and a $100 million run rate, Applied Compute is valued at about 32 times its current annualised revenue (3,250 ÷ 100 = 32.5). Adding its reported rounds of $20 million before launch, $80 million at launch and $350 million now gives about $450 million raised in total (20 + 80 + 350 = 450). Our calculation from the reported figures.

Short-term: The company will add computing capacity and staff to take on more enterprise customers. Rival firms that customise AI models for businesses may face pressure to show similar growth to raise money.

Long-term: If companies come to prefer models trained on their own data, the market for custom AI could grow beside general AI assistants, changing how businesses buy AI and how IT services firms sell it.

Who is affected: Large companies building AI into their operations, AI start-ups and IT services firms that customise models for clients, including those in India, and investors in AI infrastructure.

Questions and Answers