Daytona Raises $60M Series B to Expand Its Agent Cloud

Forking a computer lets an AI agent pursue another approach without throwing away the work already done. Daytona has built that ability into an execution platform, giving application developers a supplier to call when their software needs machines that can start, pause, branch and resume along with the task.

The San Francisco company announced a $60M Series B on October 20, 2026, led by FirstMark Growth at a reported $1B valuation. Total funding now stands at $91M. The financing supports Daytona's broader Agent Cloud ambitions, with engineering, go-to-market and operations hiring under way in San Francisco.

The commercial opportunity sits inside a decision familiar to engineering leaders: how much supporting infrastructure should the team become responsible for? An application company giving its agents more freedom also needs somewhere for that work to happen, and somebody accountable for keeping those environments available. Daytona is asking to take on that operating relationship as agent products grow.

A second FirstMark round follows a change in customer

FirstMark's return puts the new financing alongside Daytona's $24M Series A, announced on February 5, 2026. That earlier round was led by FirstMark Capital, with Matt Turck joining the board. Daytona had previously announced a $2M pre-seed in November 2023 and $5M in funding in June 2024, which reconciles the company's disclosed history with its current $91M total.

The Series B brings strategic investors Qualcomm Ventures, Samsung Next and Okta Ventures into the round. Daytona also names Alumni Ventures, Evergreen Fund and Torch Capital among new investors, alongside angels Patrick Collison, Tobi Lütke, Aravind Srinivas, Eric Yuan, Olivier Pomel, Adam D'Angelo and Brendan Foody. Returning firms include E2VC, Upfront Ventures and Pace Capital; their repeat participation is disclosed without individual investment amounts.

Daytona's founding team comprises CEO Ivan Burazin, CTO Vedran Jukić and chief architect Goran Draganić. Founded in 2023, the business began with development environment management before moving toward execution infrastructure for AI-generated code. That history gives the financing an operating continuity: the team is still working on the conditions under which software gets built and run, while its customer increasingly needs to provision environments on behalf of agents.

When a product team inherits an infrastructure service

The WRITER customer case study published by Daytona shows the purchasing decision more clearly than a list of available operating systems. WRITER initially maintained its own sandbox infrastructure. As usage passed 20,000 sandboxes a day, co-founder and CTO Waseem Alshikh describes the work of keeping the platform running as an increasing demand on the engineering team.

The case study reports that WRITER provisioned 300,000 sandboxes in a single month with Daytona and redirected engineering capacity toward its own product. These are customer-attributed results published by the supplier. The daily figure describes the earlier operating burden and the monthly figure describes Daytona usage; they cover different periods and should not be read as a direct throughput comparison.

For an engineering buyer, that distinction leaves a useful question intact. Owning an internal sandbox platform means owning its continuing maintenance, capacity and adaptation as the agent application changes. Purchasing managed execution brings an external provider into that work, along with the need to assess its reliability, controls and responsiveness.

Daytona's sale therefore extends beyond provisioning speed. The supplier must stay useful as the customer adds workflows, changes requirements and encounters problems that were absent from the original demonstration. The relationship described in the WRITER case includes direct work between the two engineering teams, making support part of the operating arrangement.

The execution layer expands around the agent's task

Daytona's current documentation describes programmatically controlled sandboxes with their own computing resources, filesystem and network stack. Developers interact with the platform through SDKs, an API and a command-line interface. The documented capabilities include process execution, file operations, Git operations, snapshots and persistence, with Linux, Windows, macOS and GPU environments represented across the platform.

Those features answer different parts of a task. A persistent environment can carry work between sessions; a snapshot can preserve a useful state; parallel copies can let an agent explore alternatives. The value to an application developer depends on how these operations fit together inside a workflow, including the permissions and information that the agent brings with it.

That is why the Agent Cloud framing matters commercially. Daytona is extending the supplier's remit across the computer, browser and storage an agent uses. A broader remit can reduce the number of components a customer assembles, while also increasing the importance of understanding the provider's boundaries and the customer's own governance responsibilities.

DevCuration's coverage of Modal's AI execution platform provides related context for this expanding infrastructure category. The competitive decision starts with the buyer's alternatives, including continuing to operate an internal service. The WRITER account provides a specific example of that route. It offers no basis for declaring Daytona universally faster, safer or cheaper than every alternative, and the financing announcement supplies no comparative procurement study across the market.

Hiring for the customer relationship after the API call

Daytona reports approximately 25-fold revenue growth since January. The announcement does not provide an absolute revenue figure or audited financial statements, so the growth claim describes the company's own account of demand. Its named customer base spans model builders, training-data and evaluation businesses, and companies putting agents into end-user products.

Serving that range asks more of an infrastructure company as the supported work broadens. Different customers bring different operating systems, runtime requirements and expectations about how long an execution environment should persist. Each addition becomes part of the service Daytona must maintain after a developer has successfully made the first API call.

The company's current hiring plan spans engineering, go-to-market and operations. That combination puts customer acquisition beside the work of delivering the platform, an appropriate pairing for a business selling continuing execution capacity. The Series B provides resources for that expansion without specifying an allocation among those functions.

For the application teams considering Daytona, the purchase will keep unfolding inside ordinary engineering work: a new task, a paused environment, a branch worth preserving and a question for the supplier when the workflow changes. Burazin, Jukić and Draganić are building the company that has to remain reachable through those moments, as more of the customer's product comes to depend on the machines behind the agent.