Daytona CEO claims 51% monthly growth for AI agent sandbox platform
Daytona co-founder Ivan Burazin says the company is expanding rapidly with a self-serve model and no traditional sales team, serving major tech firms.
Key points
- Daytona reports nearly 51% month-over-month growth over the past 18 months.
- The platform provides sandboxed environments for AI agents, including rare Windows options.
- Co-founder Ivan Burazin predicts a future CPU shortage that will exceed current GPU constraints.
Daytona co-founder and chief executive Ivan Burazin told The Next Web that the company has achieved growth of just under 51% month over month for the last 18 months. This expansion has occurred without a traditional sales force, relying instead on a self-serve business model where customers sign up and pay for consumption.
The company currently employs about 30 people and serves notable clients including Meta, Cohere, and Handshake. Burazin, a Croatian citizen, founded Daytona with two co-founders, one of whom he has known for nearly two decades. Before launching Daytona, he co-founded Codeanywhere, a cloud development environment, and created the Shift developer conference, which was later acquired by Infobip.
Context and Background
Daytona originally automated developer environments for human engineers but has shifted focus to building computers for AI agents. Burazin explained to TNW host Cristian Dina on 24 September that the platform helps agents "go to school and go to work." The "school" phase involves post-training, where AI labs run agents in sandboxes to refine their capabilities.
About 70% of Daytona's revenue comes from this post-training segment. Most teams initially use Kubernetes for deployment, but Burazin stated that Daytona is at least 36% faster for these specific jobs. This speed advantage allows organizations to free up scarce GPU resources for other tasks, according to the CEO's comments to The Next Web.
Burazin emphasized that Daytona operates on the CPU side rather than the GPU side. He expressed a strong belief that a CPU crunch is imminent and will be larger than the current GPU shortage. Most sandboxes for AI agents are Linux machines with CPUs and no GPUs, which aligns with Daytona's initial market focus.
Who is affected
To address enterprise needs, Daytona built a Windows sandbox because many corporate applications only run on Windows. Burazin claimed Daytona is the first and only provider offering a Windows sandbox for AI agents. The company also launched a GPU sandbox for agents performing 3D rendering or design work, distinct from training or inference tasks.
Customers are using Daytona to run applications like AutoCAD and Blender. The company adapts daily to changing customer demand, which Burazin credits for sustaining growth. Daytona handles approximately 1,100 Slack channels with customers and only recently agreed to hire its first head of sales, despite its rapid expansion.
What should you do?
Organizations relying on Kubernetes for AI agent deployment should evaluate whether specialized sandboxes offer performance benefits. Security teams should verify that sandboxed environments properly isolate AI agents from sensitive corporate data. CTOs should monitor resource allocation, particularly CPU usage, as Burazin predicts a significant supply constraint in that area.
What to do and how to stay safe: Daytona
- Review current AI agent infrastructure to determine if Kubernetes is the optimal deployment method for post-training tasks.
- Verify that sandbox providers offer the necessary operating system support, such as Windows, for legacy enterprise applications.
- Monitor CPU resource utilization trends, as industry leaders predict a potential shortage in computing capacity.
- Assess vendor stability by checking employee counts and customer support channels like Slack for real-time engagement.
Step-by-step guide: Mobile Device Security: How to Protect Data Beyond Passwords
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What is Daytona's main business model?
Daytona uses a self-serve, consumption-based pricing model where customers sign up and add credit cards without needing a sales representative.
Who are Daytona's notable customers?
The company serves major technology firms including Meta, Cohere, and Handshake, according to CEO Ivan Burazin.
Does Daytona offer Windows support for AI agents?
Yes, Burazin stated that Daytona built a Windows sandbox and is the only provider currently offering this option for AI agents.
