A handful of engineers now run GPU fleets that used to require entire platform teams, and that shift is exactly what convinced investors to back a one-year-old infrastructure startup with $11 million. Aranya announced the funding this week, split between a $9 million seed round First Round Capital led and a $2 million pre-seed round Asylum Ventures led, with several other firms joining both rounds.
The company’s pitch centers on a real operational problem: inference now makes up roughly two-thirds of all AI compute, roughly double what it was three years ago, and most data centers never built their infrastructure with that workload in mind. Aranya’s answer is ClusterdOS, an open-source system built on Kubernetes that turns raw bare-metal servers into managed, production-ready clusters, typically in under 48 hours rather than the weeks such setups usually take.
One customer working with infrastructure provider Hydra Host saw cluster setup times drop from six weeks to under two days after adopting Aranya’s system, and outages fell by 90% once it was running. Aaron Ginn, Hydra Host’s co-founder and CEO, said the partnership gives customers a faster path to production Kubernetes without needing to build or hire a dedicated platform team.
That’s the core problem Aranya is trying to solve across the industry. Standing up a GPU cluster at scale typically means paying hyperscaler prices, wrestling with custom hardware configurations, or effectively building your own cloud infrastructure from scratch. Aranya’s system lets teams manage large compute clusters through plain-language commands instead of configuration files or command-line tools, spinning up inference endpoints or adjusting node counts without deep infrastructure expertise.
Before reaching its first year, the company says it now manages more than $500 million worth of GPU hardware for AI companies running some of the industry’s most demanding inference workloads. Todd Jackson of First Round Capital said the team built precisely the kind of infrastructure layer this moment calls for, one that makes operational complexity disappear rather than manages around it.
With the new funding, Aranya plans to expand its engineering and go-to-market teams while pushing forward its multicluster interface, aiming to make large-scale compute management accessible to teams that don’t have specialized infrastructure expertise on staff.
