Daytona raises $60 million to build cloud infrastructure for AI agents

A new infrastructure problem

AI agents are beginning to create a unique set of infrastructure challenges because they do not behave like standard software. While traditional software often runs as long-lived services, agents frequently generate thousands of short, stateful jobs that were not planned in advance. Daytona Platforms is betting that this specific technical gap is large enough to support a dedicated cloud provider.

The startup recently secured $60 million in a Series B funding round led by FirstMark Growth. This investment brings Daytona's total funding to $91 million and establishes a reported valuation of $1 billion. The round attracted a notable group of technology leaders and venture firms, including the founders of Stripe, Shopify, and Perplexity.

The core issue Daytona addresses is that current cloud systems were not designed for the way agents operate. Agents require the ability to open files, run specific software, and preserve context over time. They must do this without sharing an environment with other processes, which creates a need for high levels of isolation and specific resource management.

Giving agents their own computers

Daytona describes its platform as providing every individual agent with its own computer. This environment is designed to be a workspace where an agent can perform complex tasks independently. It includes sandboxed versions of major operating systems, such as Linux, Windows, and macOS, to ensure compatibility with various types of software.

Beyond just providing an operating system, the platform also offers access to GPUs and other necessary services. This setup allows an agent to complete a task over a period of time while maintaining its own unique state. By providing a dedicated space, the infrastructure ensures that the agent's actions do not interfere with other tasks or users.

This distinction is vital for the reliability of AI systems. Because agents are increasingly expected to handle research, shopping, and everyday tasks, they need a stable place to work. Daytona's infrastructure provides the underlying

Moving beyond chat windows

The current trend in AI adoption is shifting away from models that simply generate answers within a chat window. Instead, the focus is moving toward software that can safely carry out work across different systems. This transition means that the infrastructure layer is becoming the primary product for many developers building agentic tools.

For companies building these products, the competitive landscape will soon involve questions about where agents run and what they can access. Reliability and trust become the primary metrics for success as agents take on more autonomous roles. Daytona is positioning its workspace around the agent itself, rather than treating the agent as a simple feature inside an existing app.

As software becomes more capable of acting across systems, the underlying cloud becomes less visible to the end user. However, it remains the most important component for ensuring that an AI can perform a task without errors. This shift highlights a move toward functional AI that can interact with the digital world in a structured, isolated way.

Consumer and enterprise impact

The move toward agent-based infrastructure has clear implications for how consumers interact with large companies. For example, Walmart has explored allowing AI agents to shop on behalf of customers. This demonstrates the consumer-facing version of the change where software acts as a proxy for human actions across various platforms.

For businesses, the ability to deploy agents that can run in isolated environments provides a layer of security. Companies can grant agents the power to perform specific tasks—like managing files or running internal software—without risking the integrity of the broader system. This isolation is a prerequisite for trusting AI with sensitive data or complex workflows.

By providing a standardized way to host these agents, Daytona aims to simplify the development process. Developers no longer have to build complex sandboxing solutions from scratch for every new agent they create. They can instead rely on a specialized cloud that understands the unique requirements of stateful, short-lived AI jobs.

The future of agentic work

The next phase of AI adoption will likely depend on whether an agent can safely execute work rather than just generating a response. This means that the infrastructure layer will become a critical gatekeeper for the next generation of software. As more companies integrate agents into their workflows, the demand for isolated, stateful environments will grow.

Daytona's funding and the involvement of major tech figures suggest a significant amount of confidence in this direction. The goal is to create a workspace where agents can operate with the same level of independence as a human user on a laptop. This could lead to a world where software is defined by its ability to complete goals rather than its interface.

As the cloud becomes more specialized to meet these needs, the distinction between