AI-infrastructure startup Delos Data has closed more than $100 million in fresh funding led by Matrix Partners, with Playground, Socratic Partners, Capricorn's Technology Impact Fund, Matter Venture Partners and IAG joining the round. The Palo Alto company is stepping out of stealth alongside its first product, Delos Nonstop AI, aimed squarely at what CEO Ed Doe calls “the most expensive idle asset in a data center” — a GPU waiting on the network.
Networking Silicon For A GPU-Starved World
Delos Nonstop AI is a combined silicon and software stack that Delos says drops inter-node latency by 10x and improves overall cluster efficiency by 10x under fluctuating agentic workloads. CTO Dan Daly frames the pitch as building a data interface “that's fast and doesn't break under load,” a specific answer to how modern agentic pipelines burst across distributed compute, accelerators and storage. Delos is going straight after frontier labs, hyperscalers, neoclouds and sovereign AI programs.
Where The $100M Goes
Proceeds will fund silicon and software engineering hires and accelerate product delivery and go-to-market. Delos is positioning itself against the growing conviction that the AI capacity crunch is not just a chip-shortage story — it's a data-movement story, and hyperscale customers are increasingly willing to pay for interconnect silicon that recovers stranded GPU cycles.
Adjacent To A Silicon Buying Spree
The raise slots into a broader AI-hardware capital wave. See our related coverage of EUCLYD's $231M Series A for European AI silicon, Alchip's 2nm ASIC platform, and the Ambarella–ZEDEDA cloud-to-edge partnership.
Reporting based on coverage from PR Newswire, FinSMEs, Electronics Weekly and Tech Edge AI.