CScale emerged from stealth on Sept. 30, 2026 with $145 million in Series C funding to build optical interconnects for AI scale-up — the high-bandwidth, low-latency links that let thousands of accelerators behave as a single, larger machine. The round brings the Palo Alto, Calif. company's total funding to $188 million since its 2023 founding.

Atreides Management, Valor Equity Partners and Premji Invest co-led the round. Sutter Hill Ventures, a backer since inception, and Maverick Silicon re-upped. NVIDIA and Intel Capital joined as CScale's first strategic investors — a signal that both accelerator vendors see interconnect reliability as a gating factor for future AI factory deployments.

CScale's pitch rests on a failure-math argument. In the gigawatt-class AI data centers the company expects by the end of the decade, a single scale-up domain will span thousands of tightly coupled accelerators across dozens of racks. Across hundreds of thousands of accelerators, link failures that are statistically rare on any one connection become a continuous fleet-level condition, degrading portions of the AI factory unless they can be contained. CScale says it is architecting its interconnect to contain optical failures without interrupting compute.

"As AI scale-up domains extend across dozens of racks, optical interconnect becomes essential," said CEO Martin Lund. "At that scale, reliable, predictable communication is fundamental to system economics. Easier part replacement improves serviceability, not continuity. We're designing the interconnect for continuity. Lasers will fail. Compute shouldn't."

The reliability framing is the company's central differentiator. Interconnect bandwidth alone, the argument goes, counts for little if a failed laser strands expensive compute. "AI infrastructure is no longer just a compute problem. It is a systems problem," said Gavin Baker, Managing Partner and Chief Investment Officer at Atreides Management. "As scale-up systems move toward gigawatt-class deployments, interconnect bandwidth means nothing without system reliability. Every optical failure is a compute failure. At this scale, failures are not hypothetical — they are inevitable."

Sandesh Patnam, Managing Partner at Premji Invest, pointed to the team's cross-domain depth. "The team combines deep expertise across photonics, electronics, and software with the deployment experience to turn optical capability into dependable AI infrastructure," he said.

The team is the startup's most tangible asset. Lund previously built Broadcom's switching business into a billion-dollar franchise and held senior executive roles at Microsoft and Cadence. Most recently he led Cisco's Common Hardware Group, responsible for silicon, hardware systems and optics, including the Silicon One silicon family. Founder and CTO Sanjai Kohli co-founded SiRF, which brought GPS to the mass market and earned the 2010 European Inventor Award, and later founded Inovi, acquired by Facebook in 2014.

CScale is building an integrated light engine to deliver what it calls reliable optical connectivity for AI infrastructure at gigawatt scale. The company has roughly 85 employees globally.

The company has disclosed no product specs, shipping dates or customer commitments — the Series C funds development and commercialization, so the burden now shifts to demonstrating that failure containment in scale-up optics can move from architectural claim to deployed hardware. With NVIDIA and Intel Capital on the cap table, CScale's progress will be a data point in how the optical interconnect market for multi-rack AI scale-up domains takes shape through the end of the decade.