Daytona Hits $1B Valuation With Computers Built for AI Agents

Daytona has raised a $60 million Series B at a $1 billion valuation, becoming the latest unicorn built on the belief that AI agents need their own kind of cloud. The round, announced on October 20, 2026 and led by FirstMark Growth, brings Daytona's total funding to roughly $91 million, with about $50 million of the new money coming from existing investors doubling down on the bet. The San Francisco startup, founded in 2023 by Croatian co-founders Ivan Burazin, Goran Draganić and Vedran Jukić, sells what it calls an Agent Cloud: secure, isolated computing environments designed for autonomous agents rather than human developers.

The investor roster reads like a founder summit. Strategic backers Qualcomm Ventures, Samsung Next and Okta Ventures joined the round alongside founders including Stripe's Patrick Collison, Shopify's Tobi Lütke, Perplexity's Aravind Srinivas, Quora's Adam D'Angelo, Zoom's Eric Yuan, Datadog's Olivier Pomel and Mercor's Brendan Foody, according to reports. Institutional investors include Alumni Ventures, Evergreen Fund, Torch Capital, Pace Capital, Upfront Ventures and E2VC, and several existing backers increased their positions substantially. The Series B follows a $24 million Series A, also led by FirstMark, that closed in February 2026.

Every agent needs a computer

Daytona's core product is a programmable sandbox: an isolated virtual machine where an AI agent can execute code, call tools, browse the web and manage workflows on its own. Chief executive Ivan Burazin has framed the pitch as every agent needing a computer, and the company describes its long-term goal as building millions of them. The environments provision in about 60 milliseconds, compared with 30 seconds to several minutes on conventional cloud platforms, because Daytona pre-loads snapshots onto server drives. They also carry persistent state, so an agent can pause a sandbox, resume it later, or fork it into parallel copies to try several approaches at once.

The company argues that traditional cloud platforms were designed around human developers and long-running applications, while agents behave differently: they launch huge numbers of short-lived jobs and need environments that start fast and disappear when the task is done. Daytona supports Linux, Windows and macOS sandboxes, plus GPU-based environments, so agents can handle workflows ranging from optimizing CUDA kernels to testing native Mac applications or operating Windows desktops. The platform covers training, evaluation and production workloads, and Daytona says its customers include organizations developing foundation models, coding agents and enterprise software, as well as teams that train and evaluate AI systems. Dealroom lists Anthropic, OpenAI, Nvidia, Cursor, Cognition, Meta and Cohere among them, and data company Snorkel alone has provisioned more than one million Daytona sandboxes.

The agent infrastructure land grab

Daytona is far from alone in betting that running agents safely at scale becomes a major cloud market. Modal raised $355 million at a $4.65 billion valuation in May, and Sail Research raised $80 million in June, both building infrastructure aimed at the same agent workload. Security-focused entries have raised too, including Rein Security's $25 million round for AI agent runtime security. The demand picture underpinning these bets is aggressive: Gartner forecasts that spending on enterprise cross-functional AI agents and assistants will exceed $23 billion in 2030, a compound annual growth rate of 59 percent. That figure covers the agents themselves rather than the infrastructure beneath them, which suggests the compute layer could be larger still.

The security framing is central to Daytona's story. Burazin has described the coming enterprise reality in workforce terms: companies will soon manage thousands of digital employees, and each one needs a safe place to fail so that a single rogue agent cannot cascade into company-wide chaos. That pitch resonates in a year when agent security has moved to the center of enterprise attention, from Microsoft treating agents as insider threats to research showing dormant agents keeping live credentials. Investors also asked Burazin why the market needed Daytona when large providers like AWS can run similar workloads; the company's answer is that agent workloads need fundamentally different infrastructure from the kind built for human-operated software.

What the new money buys

Daytona plans to use the funds to expand the Agent Cloud beyond its initial sandbox product, scale its compute infrastructure and grow its global team. The company has 35 employees and plans to hire across San Francisco, Croatia and Singapore, and it said in February that it was constrained by hardware. Demand appears to be accelerating: Daytona reports that its revenue run rate grew 25-fold between January and September, averaging about 50.9 percent growth each month, with average month-over-month net revenue retention around 135 percent.

FirstMark's thesis, articulated by partner Matt Turck, who joined Daytona's board with the Series A, is that the next infrastructure shift moves from human-centric cloud primitives to agent-native ones, with instant startup, persistent state and the tooling agents need to write code, use Git and execute safely at scale. Burazin has also argued that CPU demand from agentic AI will eventually exceed GPU spending, a view echoed by AMD's raised server CPU forecasts. Daytona's $1 billion valuation sits in the top 8 percent of US enterprise software Series B rounds, according to Dealroom, which makes it a priced-for-growth bet that agent-specific infrastructure has become consensus rather than speculation. For a company that began by automating development environments for human programmers before pivoting to agents, the round is a marker of how completely the industry's center of gravity has shifted toward the machines that give agents somewhere to work.

Sources

Reporting for this article was drawn from Today's Startup News on Daytona's $1B valuation round, Dealroom's coverage of the $60M Series B, Tech.eu on the agent cloud raise, Pulse2's round details, AlleyWatch's 10/20/2026 funding report and Fortune's exclusive on the round via Alpha Maven. Related reading: Reduck's $3M raise to build browser automation for agents, Blacksmith's $66M Series C for trust in agent-built code and Microsoft's Project Perception treats AI agents as insider threats.