Delos Data Raises Over $100 Million for AI Infrastructure

Sep 15, 2026
Delos Data raised more than $100 million to expand its engineering teams and develop its Nonstop AI infrastructure products.

Delos Data raised more than $100 million to expand its engineering teams and develop its Nonstop AI infrastructure products, the company said in a press release. Investors included Matrix, Playground, Socratic Partners, Capricorn's Technology Impact Fund, Matter Venture Partners, IAG and industry leaders across compute, hyperscale, networking silicon, infrastructure software and optical connectivity.

The Palo Alto company also introduced its Nonstop AI Reference Architecture and Nonstop AI Data Interface. Delos says the products provide 10 times lower latency and 10 times greater efficiency for inference, while allowing customers to combine different hardware, models and network topologies within one domain.

The Data Interface joins the company's existing Nonstop AI Clusters and Nonstop AI Server. Delos Nonstop AI rests on the company's premise that networking is the next major bottleneck in AI systems. Delos says agentic inference workloads run persistently across GPUs, XPUs, accelerators, CPUs, memory and storage and must keep operating through failures at scale, and it cites projections of roughly 300 times growth in token demand by 2030, most of it from agentic inference.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Silicon Brief, AI Funding Brief or Daily AI Brief.

Free newsletter

AI Funding Brief

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more
Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.