CScale Raises $145M for AI Interconnects
CScale is targeting optical interconnects, one of the networking bottlenecks inside large AI data centres.
CScale’s stealth exit highlights a growing reality of AI infrastructure: connecting chips together is becoming as important as the chips themselves.
What happened
Palo Alto-based CScale exited stealth with $145 million in Series C funding to develop optical interconnect technology for AI scale-up networking. The round was co-led by Atreides Management, Valor Equity Partners and Premji Invest, with Nvidia and Intel Capital joining as strategic investors.
The company is focused on high-bandwidth, low-latency connections for large AI systems.
Why it matters
Modern AI workloads depend on huge clusters of accelerators working together. If the connections between chips and systems are too slow or inefficient, expensive compute capacity can be underused.
Optical interconnects are one way to improve bandwidth and energy efficiency as AI clusters grow. CScale’s investor base also signals that major semiconductor players see networking as a core infrastructure layer.
The bigger picture
AI infrastructure is becoming a full-stack hardware challenge. GPUs get the most attention, but memory, networking, optics and power systems are increasingly strategic. CScale is part of the wave of startups tackling the less visible bottlenecks behind AI scale.