E2B says it crossed $100 million ARR selling cloud computers to AI agents

E2B, the San Francisco company that gives AI agents their own isolated cloud computers, said on Thursday, Nov. 5, 2026, that it has passed $100 million in annualized run rate, up from $8 million a year earlier, in its first public revenue disclosure.

Every serious coding agent eventually needs a machine it can break without breaking yours. E2B turned that boring requirement into a self-serve cloud and is disclosing a nine-figure run rate before most of its peers admit they have revenue. If agents are the new apps, sandboxes are the new servers — and someone has to meter them.

In the body of its Nov. 5 Business Wire release, E2B said it has passed $100 million in annualized run rate. Annualized run rate, ARR, takes recent revenue and stretches it across a year. It is a pace the company is reporting. The headline of that same release, and a graphic that went with it, said "$120 million" and "$120M+". The $100 million figure is the one in the body. Both wordings are E2B's. This is a revenue pace, not a funding round.

E2B said it was at $8 million ARR a year ago, and that this year alone it is projected to add over $150 million of net new revenue. A projection is the company's forecast, not cash already collected. E2B says the growth is almost 100 percent from what it calls ALG, Agentic-Led Growth: inbound and self-serve, rather than outbound sales. Inbound means the customer showed up. Those lines are E2B's.

Close to 200,000 teams use E2B. Ninety-four of the Fortune 100 have signed up. Manus, named in the release, found E2B on its own and scaled with no sales calls or demos. Those lines are E2B's.

E2B crossed 1 billion agent computers started in June 2026, passed 2 billion in October, and said it will hit 3 billion in the next 30 days. An agent computer, here, is one isolated machine started for a task. The first two counts are what the company says has already happened. The 3 billion line is what it says comes next. Those figures are E2B's.

Each agent task runs on its own isolated machine, a full operating system, with its own networking and firewall, instant snapshot copies, and browser and password-vault tools. Isolated means one task cannot see another customer's machine. A snapshot is a saved copy the task can start again from. E2B says it is SOC 2 Type II compliant and HIPAA-ready. SOC 2 Type II is an outside review of how a company protects customer data over a stretch of time. HIPAA-ready means E2B says the setup can be used in a way that meets U.S. health-privacy rules. Those lines are E2B's.

E2B has raised $38 million to date, including a $21 million Series A led by Insight Partners in July 2025. Other investors include Decibel, Kaya VC, and Sunflower Capital. A Series A is an early venture round. Those lines are E2B's. The $100 million figure above is revenue pace. It is not money investors just put in.

E2B was co-founded by chief executive Vasek Mlejnsky and chief technology officer Tomas Valenta. The company started in Prague and later moved to San Francisco after OpenAI president Greg Brockman shared an early demo on X in 2023. Those lines are E2B's.

Vasek Mlejnsky, co-founder and chief executive, said: "The future of agentic infrastructure is increasingly more automated. We see agents managing more complex systems that go beyond just running a single computer in a cloud." That quotation is his. Agentic, here, means software that takes a series of steps, not only answers a question.

The picture is E2B's agent sandbox screen: a terminal creating an isolated sandbox and writing a checkout app, with session i4h2k9x in us-east-1. The credit is E2B.

In plain terms, E2B says the cloud computers it rents to AI agents are now running at a $100 million annual pace, up from $8 million a year ago, and that almost all of the growth came inbound. Close to 200,000 teams use it, including 94 of the Fortune 100. The company has raised $38 million in funding. The $100 million is revenue pace.