Daytona Hits $1B Valuation Renting Secure Cloud Sandboxes to AI Agents
Daytona raised $60M at a $1B valuation to rent secure cloud sandboxes where AI agents can execute code without breaking company systems.
Funding · Source: Tech.eu
What happened
Daytona just secured a $60 million Series B funding round at a massive $1 billion valuation. FirstMark led the investment round. The San Francisco based startup builds cloud infrastructure specifically designed for AI agents. They provide isolated virtual computers, known as sandboxes, where autonomous systems can write code, browse the web, and operate software. The company has now raised $91 million in total funding since being founded in 2023 by Ivan Burazin, Goran Draganic, and Vedran Jukic.
The customer list reads like a roster of the biggest players in the artificial intelligence industry. OpenAI, Anthropic, Meta, and NVIDIA all rely on Daytona for their infrastructure needs. Fast growing AI coding startups like Cursor and Cognition are also using the platform. The demand is translating into massive financial growth. Between January and September, Daytona saw its revenue run rate grow 25-fold. They are reportedly growing at an average rate of 50.9 percent every single month.
The cap table is packed with heavy hitters from Silicon Valley. Stripe founder Patrick Collison, Shopify CEO Tobi Lutke, Zoom founder Eric Yuan, and Perplexity CEO Aravind Srinivas all invested in this round. Existing investors showed massive conviction too. Pace Capital doubled its investment, Upfront Ventures tripled its stake, and E2VC quadrupled its commitment. Daytona will use the fresh capital to scale its computing infrastructure and hire new talent across San Francisco, Croatia, and Singapore.
Key facts
- $60 million — Series B funding raised by Daytona
- $1 billion — New valuation for the AI infrastructure startup
- 25-fold — Increase in revenue run rate between January and September
- 50.9% — Reported average monthly growth rate
- 96% — Reduction in evaluation cycles for customer Snorkel
Why it matters
Building AI agents is no longer just about calling an API to generate text responses. Agents need to execute code, interact with applications, and perform real work autonomously. That requires serious computing power. Traditional cloud providers like Amazon Web Services, Microsoft Azure, and Google Cloud were built for predictable human workloads. AI agents operate differently. They create sudden, massive bursts of demand that can overwhelm standard systems. Daytona solves this by provisioning secure environments in milliseconds across Linux, Windows, and macOS. Builders can now spin up thousands of sandboxes instantly. They can scale resources up or down, pause tasks, and resume previous work without losing progress.
The infrastructure bottleneck in artificial intelligence is actively shifting from GPUs to CPUs. The industry spent the last two years hoarding expensive graphics processors to train massive models. Now, those models are being deployed as autonomous agents that need standard computing power to execute daily tasks. Security is the real wedge driving this shift. As Daytona CEO Ivan Burazin noted, agents do their jobs without morals or fear of consequences. Recent agent security incidents are just previews of what is to come. Giving an agent unrestricted access to your company systems is a disaster waiting to happen. Dedicated agent clouds will become a mandatory layer of the enterprise software stack.
For builders
Faster evaluation cycles for AI models
AI data startup Snorkel used Daytona to provision over one million sandboxes on the platform. This massive scale reduced their evaluation cycles by 96 percent. If you are testing agents at scale, purpose built infrastructure saves massive amounts of time and engineering resources.
Instant environment startup times
Another customer named Trajectory cut their environment startup times from 30 minutes down to mere seconds using Daytona. Builders paying for idle cloud compute can switch to millisecond provisioning. This directly lowers infrastructure costs and speeds up deployment for agentic workflows.
Secure isolation prevents system chaos
Agents will inevitably break rules or execute malicious code by mistake while trying to complete a task. Daytona isolates these failures in secure sandboxes so one broken agent does not spill into company wide chaos. Enterprise founders must adopt this architecture or risk catastrophic security breaches.
My take
The artificial intelligence gold rush has a new shovel seller. Everyone is obsessed with training smarter models, but Daytona proved that models are just the show. Infrastructure is the actual business. They realized that millions of digital employees need a safe place to fail, and they are taxing the entire ecosystem to provide it.
Original reporting: Tech.eu. This is my rewrite and opinion.