Daytona Agent Cloud: Why a $60M Round Carries a $1B Valuation

Key Takeaways

• Daytona Platforms raised a 60 million dollar Series B round at a 1 billion dollar valuation, led by FirstMark Growth, bringing its total funding to 91 million dollars.
• The round was announced on 20 October 2026 and widely reported on 23 October 2026, following a 24 million dollar Series A in February 2026, also led by FirstMark.
• Investors named in the round include Qualcomm Ventures, Samsung Next, and Okta Ventures, plus prominent founder angels including Stripe’s Patrick Collison, Shopify’s Tobi Lütke, and Perplexity’s Aravind Srinivas.
• Daytona builds what it calls an Agent Cloud: secure, isolated computing environments, or sandboxes, where AI agents can run code, use tools, browse the web, and complete work.
• Founder Ivan Burazin’s pitch is that every agent needs its own computer, because agents spin up thousands of short, stateful jobs that traditional cloud infrastructure was not designed for.
• The valuation is a bet on agent infrastructure becoming a foundational layer, not a reflection of current revenue alone. Daytona reports a 25-fold increase in its revenue run rate between January and September, a company-reported figure that has not been independently audited in the coverage reviewed.

Daytona Platforms has joined the ranks of unicorns, the term for a private startup valued at 1 billion dollars or more, after raising 60 million dollars in a Series B round at a 1 billion dollar valuation. The round, led by FirstMark Growth, brings Daytona’s total funding to 91 million dollars and puts a striking price on a three-year-old company that most people outside the AI infrastructure world have never heard of. The question the valuation raises is not whether Daytona is a real company with real customers. It is, and its customer list includes some of the best-known names in AI. The question is why investors believe the infrastructure for AI agents deserves a billion-dollar valuation now, before agent use has become truly mainstream.

What Daytona actually builds

Daytona’s core product is a programmable sandbox, an isolated computing environment where an AI agent can execute code, call tools, browse the web, and manage workflows on its own. The company calls the broader platform the Agent Cloud. Chief executive Ivan Burazin has described the pitch in simple terms: every agent needs a computer, and Daytona is building millions of them. That phrase is a useful way to understand the product, because it reframes what a cloud is for.

Traditional cloud platforms were designed around human developers and long-running applications. A web service, a database, or an internal tool is launched, kept running, and scaled up or down as demand changes. Agents behave differently. They launch short-lived jobs in large volumes, often without advance planning, and they need environments that start quickly, hold state while they work, and can be paused, resumed, or copied to try several approaches at once. Daytona says its sandboxes provision in milliseconds, compared with up to minutes on conventional clouds, and offer persistent state and granular controls. Those differences sound technical, but they matter in practice: if an agent has to wait minutes for an environment to start, or loses its work when a session ends, it cannot operate at the speed and scale that agent workflows require.

Daytona’s platform covers sandboxed versions of Linux, Windows, and macOS, plus GPUs, on one platform, according to Burazin’s announcement post. That breadth matters because agents do not all work in the same environment. A coding agent may need Linux, a business automation agent may need Windows or macOS, and a training or evaluation workload may need GPUs. Providing all of those in isolated, programmable environments, with consistent controls, is the infrastructure problem Daytona is trying to solve. The company says its platform now covers training, evaluation, and production workloads for AI labs and enterprises, having started by building development environments for human programmers before moving to infrastructure for agents.

The round in detail

The Series B round was announced on 20 October 2026, with TechCrunch reporting the news on 23 October 2026. That gap between announcement and wider reporting explains why some coverage dates the round to 20 October and some to 23 October. Both dates refer to the same round, and this article uses 20 October as the announcement date and 23 October as the date of TechCrunch’s report.

FirstMark Growth led the round, having also led Daytona’s 24 million dollar Series A in February 2026. Roughly 50 million dollars of the new money came from existing investors, according to one report, which suggests that backers who already knew the company chose to increase their bets rather than simply maintaining them. That pattern is often read as a positive signal, because existing investors have more information about a company’s progress than new investors do, although it is not a guarantee of future success.

Who uses Daytona

Daytona says it provides infrastructure to Anthropic, OpenAI, Nvidia, Cursor, and Cognition, while its customer base also includes Meta, Cohere, Mercor, Handshake, Surge AI, SambaNova, Postman, Bolt.new, n8n, and Clay. That list spans AI labs, developer tools, and automation platforms, which is consistent with a product that serves both training and production use. The company also reports growing adoption for AI training and evaluation, with customers including Snorkel and Trajectory using its infrastructure to reduce processing times.

For readers who build or buy AI tools, the practical question is simpler: does agent infrastructure change what your team can do. If an agent can be given a safe, isolated computer for each task, with the ability to run code, use tools, and be stopped or copied without affecting other work, then agents become safer to deploy in production, not just in demos. That safety point is central to Burazin’s argument. He has said that companies are going to manage thousands of digital employees, and each one needs a safe place to fail, so that one rule-breaking agent does not spill into company-wide chaos. Isolation is therefore not just a technical feature. It is a governance feature, and governance is what enterprises need before they trust agents with real systems and data.

What agent-cloud infrastructure is, and why it is different

Agent-cloud infrastructure is a distinct layer from general cloud computing, and understanding the difference helps explain the valuation. General cloud computing provides virtual machines, storage, and networking that humans configure and manage. Agent-cloud infrastructure provides programmable, short-lived, stateful environments that agents themselves can create, pause, resume, and discard, with controls that limit what each agent can access and do.

Three differences matter most. The first is startup speed. Agents that spin up thousands of jobs cannot wait minutes for each environment, so millisecond provisioning is a functional requirement, not a performance bonus. The second is state. Agents need to remember what they were doing across steps, and to resume after a pause, which requires persistent state tied to the sandbox rather than to a single long-running server. The third is isolation and control. Because agents act autonomously and can make mistakes, each agent’s environment must be sandboxed so that a failure, a bad instruction, or a malicious input does not spread to other agents or to company systems. Granular controls let operators set limits on what an agent can access, how long it can run, and what it can change.

Why investors are paying a billion dollars

There are at least three reasons investors may find that thesis persuasive. First, agent adoption is already visible in coding, customer operations, research, and automation, and Daytona’s customer list suggests it is serving that demand now, not only in the future. Second, the technical problem is genuinely hard, involving fast provisioning, state management, isolation, and multi-operating-system support, which creates a barrier for new entrants. Third, the investor roster itself creates momentum: when prominent founders and strategic investors back a company, it becomes easier to hire, to win customers, and to raise the next round, which can become self-reinforcing.

What Daytona plans to do with the money

Daytona says it intends to use the funds to expand its agent cloud platform beyond its initial sandbox product, to scale its computing infrastructure, and to grow its global team, with hiring planned across San Francisco, Croatia, and Singapore. That plan is consistent with the thesis: the sandbox is the starting point, but the value lies in a broader platform for training, evaluation, and production, backed by enough compute to serve large customers. Hiring across three locations also reflects Daytona’s origins, with Croatian co-founders Ivan Burazin, Goran Draganić, and Vedran Jukić, and its current base in San Francisco.

For businesses evaluating AI agents more broadly, our guide to how AI can help your business offers a practical framework for separating workflows that work today from those that are still experimental. Agent infrastructure belongs in that evaluation, because an agent is only as safe and reliable as the environment it runs in. A powerful model in an unsafe environment is a risk, while a capable model in a well-isolated, well-controlled environment is a tool a business can actually deploy.

For a wider look at where today’s AI tools sit on the road to more capable systems, see our article on AI to superintelligence.

The fair summary is this: Daytona has raised 60 million dollars at a 1 billion dollar valuation to build computers for AI agents, backed by FirstMark and a notable roster of strategic and founder investors, with a customer list that includes major AI labs and developer platforms. The valuation reflects a bet that agent infrastructure will become a foundational layer, not a claim that it already is one. If agents scale as investors expect, and if Daytona’s platform proves to be the safest and fastest way to run them, the bet may look prescient. If either condition fails, the valuation will be revisited. For now, Daytona is a company to watch, not because of its valuation alone, but because of what its product says about where AI work is heading: from models that answer questions to agents that do work, each needing its own safe place to do it.

Sources: Agent cloud startup Daytona hits $1B valuation backed by FirstMark and a who’s who of angels and Daytona Raises $60M in Series B Funding and Daytona hits $1B valuation with $60M Series B for AI agent cloud