Simular Raises $21.5M to Build Autonomous Computer Agents

Led by Felicis Ventures, with participation from NVentures, Basis Set Ventures, Flying Fish Partners, South Park Commons and angel investor Lenny Rachitsky

San Francisco — Simular, a research-driven startup founded by former DeepMind scientists, today announced a $21.5 million Series A funding round to accelerate the development of its computer-using agents. The round was led by Felicis Ventures, with participation from NVentures (NVIDIA’s venture arm), Basis Set Ventures, Flying Fish Partners, and South Park Commons. Angel investor Lenny Rachitsky also joined the round.

Previous backers include Samsung NEXT, Xoogler Ventures, Lilian Weng and Abhinav Gupta. 

"Simular’s team brings a rare combination of world-class AI research expertise and business depth. Their focus on building computer-use agents reliable enough for enterprises instantly convinced us that they'll be a key driver for the future of knowledge work," said Peter Deng, general partner at Felicis.

“This is a uniquely talented team with a clear, ambitious vision. They’ve been ahead of every major technological wave, and I’m excited to see what they create in the decade to come,” said Lan Xuezhao, managing partner at Basis Set, which led Simular’s $5 million seed round.

Nearing Human Performance

Simular builds AI agents that can operate software on behalf of a user, navigating interfaces, completing workflows and interacting across applications like a real knowledge worker. 

The company’s open agentic framework, Agent S, recently achieved 69.9% success on OSWorld, the leading benchmark for evaluating an agent’s ability to complete real computer tasks. This performance is now approaching human accuracy at 72%.

On the heels of the fundraise, the company is debuting Simular 1.0, a native desktop agent that can browse in the background, operate across applications and complete long, complex workflows.

“Agents must be able to reliably use a computer in real environments, not just in simulations,” said Ang Li, CEO and co-founder of Simular. “Simular 1.0 holds two keys: It can repeat its own success and collaborate seamlessly with humans as a capable teammate.”

Meet Your Teammate: Simular 1.0

Simular 1.0 is designed to bring AI agents into real, everyday workflows. It’s trained with humans in the loop, allowing users to redirect or correct the agent at any step using natural language. Over time, it learns from this supervision just like a real teammate would, becoming increasingly reliable at repeating successful workflows.

Simular 1.0 also supports contextual task triggers that users can configure so the right workflow launches automatically based on real-time activity on their computer. Lastly, this version of the agent introduces a Bigfoot persona that surfaces tips and flags issues, making interacting with an AI feel more approachable and human.

About Simular – The Autonomous Computer Company

Simular is building the AI infrastructure for autonomous computers -- machines that see, click, type, and learn across desktops, browsers, and smartphones. Powered by a hybrid agent system that combines neural exploration, symbolic execution, and continual learning, Simular delivers production-grade autonomy across platforms.

Our open-source agent, Agent S, ranks top on the OSWorld benchmark. Our user-facing agents have executed millions of steps for users, automating tasks in insurance, recruiting, travel, and more. Founded by AI experts from DeepMind, Baidu, LLNL and UCSB, the team has published 100+ papers in top AI venues with 10,000+ citations, including a Best Paper Award at ICLR 2025 Agentic AI workshop.

Simular's global team photo @ December 2025