The Pulse
Daytona Raises $60 Million at a $1 Billion Valuation
Daytona announced a $60 million Series B at a $1 billion valuation, bringing its total funding to $91 million. The startup sells virtual-machine sandboxes for AI agents and says it can configure one in 60 milliseconds.
AI.info Team · Written with AI · fact-checked ·
“The inference clouds is where the agent thinks, and our cloud is where the agent works.”
Ivan Burazin, cofounder and CEO of Daytona
A $1 billion bet on agent work
Daytona raised $60 million in a Series B at a $1 billion valuation, the company said on October 20. FirstMark led the round, which included strategic investments from Qualcomm Ventures, Samsung Next and Okta Ventures, as well as individual investments from founders of Stripe, Shopify, Perplexity, Quora and Zoom. Daytona cofounder and CEO Ivan Burazin told Fortune that about $50 million came from existing investors.
The financing brings the company’s total raised to about $91 million. Founded in 2023 by Burazin, Goran Draganić and Vedran Jukić, Daytona first focused on automating software development environments for enterprises. Burazin said the founders later rebuilt the business around computers for AI agents, with only about three months of revenue left when they relaunched.
Why Daytona sells isolated computers
Daytona provides virtual-machine sandboxes where AI agents can run code, browse the web and operate software. A sandbox can be limited to the programs and network access needed for a task; that isolation can reduce the risk of an agent taking unauthorized actions or exposing sensitive information. Daytona says AI data company Snorkel has provisioned more than one million sandboxes on its platform. The company also lists Anthropic, OpenAI, Nvidia, Cursor, Cognition, Meta and Cohere as customers.
The technical pitch centers on startup time. Fortune reports that provisioning a virtual machine through major cloud providers can take 30 seconds to several minutes, depending on its configuration. Burazin said Daytona can configure a sandbox in 60 milliseconds by keeping snapshots of prepared machines on server drives, then attaching processing power and memory when an agent needs one. Agents can also clone a running environment to test different approaches in parallel.
Growth figures come from Burazin
Burazin told Fortune that Daytona’s revenue run rate grew 25-fold between January and September. He put the company’s average monthly growth over that stretch at 50.9%.
Burazin also offered a back-of-the-envelope estimate for Meta’s Muse personal agent: serving 100 million users would require $2.8 billion worth of sandboxes, he said, based on his view that each chat has a computer underneath it. That is a hypothetical estimate, not a disclosed cost or a forecast from Meta.
The CPU argument behind the raise
Daytona’s pitch differs from that of GPU-focused data-center providers. GPUs run AI models; Daytona’s virtual machines need CPUs to run software and agent tasks. Burazin argues that infrastructure built for agents will require far more CPU capacity than conventional cloud services provide, and said the company plans to put most of the new capital toward hiring sales staff, along with some engineers.
That makes Daytona’s bet more specific than a claim that agents need more computing power: it is betting that companies will pay for isolated environments that start quickly and can be copied as agents branch into parallel work. The $1 billion valuation prices that bet before the company’s future CPU demand or revenue is certain. For now, its clearest operational claim is a 60-millisecond sandbox setup time. Fortune contrasts that with 30 seconds to several minutes to provision virtual machines through existing cloud providers.