AI agents are starting to create a new infrastructure problem: they do not behave like ordinary software.
Daytona Platforms is betting that problem is large enough to build a cloud around. The startup has raised a $60 million Series B led by FirstMark Growth at a reported $1 billion valuation, bringing its total funding to $91 million.
The round also includes a striking list of technology founders and venture arms. Daytona says participants include 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, alongside Qualcomm Ventures, Samsung Next and Okta Ventures.
AI agents need somewhere to work
Daytona’s pitch is that today’s cloud was designed for long-lived services, while agents create thousands of short, stateful jobs that were not planned in advance. In practical terms, an agent may need to open files, run software, preserve context, use a particular operating system and complete a task over time without sharing an environment with another process.
The company describes its platform as giving every agent its own computer. That environment can include sandboxed versions of Linux, Windows and macOS, as well as GPU access and other services. The distinction matters because the next phase of AI adoption will depend less on whether a model can generate an answer and more on whether it can safely carry out work.
The infrastructure layer is becoming the product
This is where the funding story becomes more interesting than another billion-dollar AI valuation. Daytona is positioning the workspace around the agent, rather than treating the agent as a feature inside an existing application.
That direction connects with the broader shift toward agents handling shopping, research and everyday tasks. Walmart’s move to let AI agents shop on customers’ behalf shows the consumer-facing version of the same change: software is increasingly expected to act across systems, not simply respond inside a chat window.
For companies building agent products, the competitive question will therefore include where agents run, what they can access and how their work is isolated. The cloud may become less visible to users, but more important to the reliability and trust of every AI service built on top of it.