Applied Compute Raises $350 Million at $3.25 Billion Valuation

The AI infrastructure startup, launched less than 18 months ago, reports annualized recurring revenue topping $100 million.

Updated on Nov. 4, 2026 in Artificial Intelligence

Applied Compute Raises $350 Million at $3.25 Billion Valuation

Applied Compute raised $350 million in a funding round led by Elad Gil that valued the company at $3.25 billion. The round also drew participation from Nvidia Corp., Shopify Ventures, HubSpot Ventures and more than a dozen other investors.

Why it matters

The raise underscores continued investor appetite for companies building infrastructure to train and run large language models in the United States. Applied Compute's reported revenue suggests demand for its platform has grown quickly since its launch.

Applied Compute's platform runs data generation and weight adjustment simultaneously through asynchronous reinforcement learning, and its autoscaling engine adjusts hardware allocation to model traffic. Exact hardware specifications and performance benchmarks were not disclosed in the source material.

The players

Applied Compute

An artificial intelligence infrastructure company offering a cloud platform for training and running large language models.

Elad Gil

A prominent angel investor and startup adviser who led the funding round.

Nvidia Corp.

A major semiconductor company whose graphics processors are widely used for AI training and inference, and which participated in the round.

The details

Applied Compute operates a cloud platform for training large language models and running inference, offering a model router called Shift-1 and supporting self-distillation that removes the need for teacher models. The company announced the round on Monday, November 3, 2026.

- Monday, November 3, 2026: Applied Compute announces the $350 million funding round.

Tech Race

Applied Compute's asynchronous reinforcement learning approach extends established post-training methods by running data generation and weight updates concurrently. The move positions the startup among a growing group of infrastructure providers competing with larger cloud and chip companies to serve AI developers in the United States.

Developers and businesses building AI tools may see new options for training and serving models through cloud platforms like Applied Compute's. The company's model router and autoscaling engine aim to reduce the hardware overhead of matching capacity to user traffic.

The takeaway

Investor interest in AI infrastructure remains strong, and developers evaluating model training and inference platforms may want to watch how startups like Applied Compute compare on cost and efficiency.