Funding

CScale Exits Stealth and Raises $145M Series C for AI Optical Interconnect

mm
Add Unite.AI to your preferred sources on Google

CScale exited stealth on September 30, 2026, announcing a $145 million Series C round aimed at speeding the development and commercialization of its optical interconnect technology for AI scale-up. The company said the round brings its total funding to $188 million.

The round was co-led by Atreides Management, Valor Equity Partners, and Premji Invest. Sutter Hill Ventures, which has backed CScale since its inception, continued its support alongside existing investor Maverick Silicon. The company said NVIDIA and Intel Capital joined as its first strategic investors. CScale, based in Palo Alto, California, describes its focus as resilient optical interconnect for AI scale-up.

The Scale-Up Interconnect Problem

CScale describes AI scale-up as the links that allow large numbers of accelerators to function together as one much larger machine, requiring very high bandwidth and extremely low latency. The company said that in the gigawatt-class AI data centers expected by the end of the decade, a single scale-up domain will encompass thousands of closely linked accelerators spread across dozens of racks. Such systems, it said, require consistently high bandwidth, latency that stays low and predictable, and uninterrupted communication.

When a fleet reaches hundreds of thousands of accelerators, link failures that would be rare in isolation turn into a constant, fleet-wide condition that can impair parts of what the company calls the AI factory, CScale said. The company said it is designing its interconnect so that optical failures are contained while compute continues to run.

Martin Lund, CEO of CScale, said in the company’s announcement that optical interconnect becomes essential as scale-up domains stretch across dozens of racks. “We’re designing the interconnect for continuity. Lasers will fail. Compute shouldn’t.” Lund said dependable, predictable communication becomes fundamental to system economics at that scale, and he distinguished easier part replacement, which improves serviceability, from continuity.

CScale is building what it describes as an integrated light engine intended to provide dependable optical connectivity for AI infrastructure at gigawatt scale. On its homepage, the company says AI factories keep getting bigger and that with thousands of accelerators operating together, failures are a constant; it says its optical interconnect is designed to isolate those failures so that compute keeps running.

Investor Statements

Gavin Baker, Managing Partner and Chief Investment Officer at Atreides Management, said AI infrastructure has become a systems problem that extends beyond compute alone. He said bandwidth is meaningless without system reliability as scale-up systems approach gigawatt-class deployments. “Every optical failure is a compute failure,” Baker said, adding that failures at that scale are inevitable rather than hypothetical. He said CScale is designing reliable scale-up optics for next-generation AI factories that deliver the full performance of optics without compromises.

Sandesh Patnam, Managing Partner at Premji Invest, said CScale’s strength lies in its architectural judgment, which he described as knowing which technical choices matter to the system as a whole. He said the team pairs expertise across photonics, electronics, and software with the deployment experience to make optical capability dependable in production AI infrastructure. Patnam said Premji Invest backs ambitious teams with the discipline to build lasting companies, described CScale as such a team, and said the firm is supporting the company as it brings its technology to market.

Founding and Leadership

CScale was founded in 2023 and is headquartered in Palo Alto, California, with approximately 85 employees globally.

According to the company, Lund previously grew Broadcom’s switching business into a billion-dollar operation and held senior executive roles at Microsoft and Cadence. Most recently, he led Cisco’s Common Hardware Group, overseeing silicon, hardware systems, and optics, including Silicon One. The company said his career spans three decades of silicon and network engineering leadership across Cisco, Microsoft, Cadence, Broadcom, and Intel.

Founder and CTO Sanjai Kohli co-founded SiRF, which helped take GPS mainstream, and received the 2010 European Inventor Award. He later founded Inovi, which Facebook acquired in 2014. The company said Kohli has founded and sold multiple startups in the communications industry.

Theo Nash is an AI-generated specialist at Unite.AI, covering AI infrastructure, compute, and the hardware systems that power modern artificial intelligence. His work focuses on the technical foundations behind large-scale AI workloads, including data centers, accelerators, networking, and the software stacks that tie them together.

With an analytical and engineering-driven perspective, Theo examines how advances in GPUs, custom silicon, memory architectures, and distributed systems enable new generations of AI models. He pays particular attention to performance trade-offs, energy efficiency, scalability, and the practical constraints that shape real-world deployment of AI infrastructure.

Articles authored by Theo Nash are AI-generated and reviewed by Unite.AI’s editorial team to ensure technical accuracy, clarity, and responsible coverage of the rapidly evolving AI compute landscape.