Daytona Raises $60 Million, Reaches Unicorn Valuation

Daytona, the technology company led by Croatian entrepreneur Ivan Burazin, has raised $60 million in a Series B funding round that values the business at $1 billion. The round was led by FirstMark Growth and brings total capital raised to $91 million. The company is building infrastructure that enables autonomous AI agents to execute complex digital tasks in their own computing environments.

The investor group includes Stripe’s Patrick Collison, Shopify’s Tobi Lütke, Perplexity’s Aravind Srinivas, Zoom’s Eric Yuan, Datadog’s Olivier Pomel, Quora’s Adam D’Angelo and Mercor’s Brendan Foody. Qualcomm Ventures, Samsung Next and Okta Ventures joined as strategic investors, alongside new and existing backers.

Daytona is targeting a structural difference between conventional cloud workloads and agent-driven computing. Traditional cloud services are largely engineered for applications and services that remain active for extended periods. AI agents, by contrast, may generate thousands of short-lived tasks, each requiring an isolated environment with access to files, browsers, development tools and persistent state.

Burazin describes the company’s approach as an Agent Cloud: infrastructure for executing an agent’s work, rather than merely running code or monitoring models. Daytona is developing environments that can start in milliseconds, pause without losing state, branch into parallel copies and be removed when a task ends. The objective is to reduce the latency and operating overhead associated with large volumes of automated work.

The platform supports isolated sandbox environments for Linux, Windows, macOS and GPU workloads. This allows agents to work with code and Git repositories, test native Mac applications, operate Windows desktops or perform computationally intensive tasks such as CUDA kernel optimization. Daytona positions the breadth of operating-system support as a key differentiator, although its claims about competitive uniqueness are the company’s own.

Matt Turck, a partner at FirstMark, argues that computing infrastructure is shifting from services designed primarily around human users towards systems built around how agents operate. He highlights rapid provisioning, persistent state and tools for safe code execution. Whether this becomes a distinct category of cloud infrastructure will also depend on cost efficiency, reliability and security under production workloads.

Daytona says its revenue has grown approximately 25-fold since January, while average monthly net revenue retention has been around 135 percent. Net revenue retention measures changes in spending among existing customers, including expansion, but it does not establish the company’s absolute revenue or profitability. Daytona also says that customers acquired six months earlier are collectively spending almost three times as much today.

The customer base spans several parts of the AI market. Daytona names Meta and Cohere among model developers, while Handshake, Snorkel, Surge AI, Turing, AfterQuery and Trajectory work on data and environments used to train advanced models. Writer, Clay, n8n, Postman, Mintlify and bolt.new represent another group embedding agent capabilities into software and workflows.

Writer illustrates the operational pressure that can push companies away from maintaining their own agent infrastructure. It previously ran an internal sandbox platform until activity exceeded 20,000 environments per day. Daytona says Writer subsequently provisioned 300,000 sandboxes on its platform in one month. AfterQuery, meanwhile, reportedly migrated from its in-house Google Kubernetes Engine cluster to Daytona within a day.

Quinn Li, senior vice president at Qualcomm Technologies and global head of Qualcomm Ventures, says AI agents require secure environments to execute code, invoke tools, access websites and train through reinforcement learning. For infrastructure providers, that introduces requirements around workload isolation, permissions and predictable resource consumption, beyond the availability of raw compute power.

The financing places Daytona among the small number of technology businesses founded by Croatian entrepreneurs to achieve a valuation of at least $1 billion. The broader industry question is whether agent execution becomes an independent layer of cloud infrastructure or a capability absorbed into established platforms. The new funding gives Daytona additional resources to develop its technology as companies move agent systems from prototypes into business operations.