CScale raises $145M Series C for optical interconnects in AI data centers
CScale, a Palo Alto-based startup, has secured $145 million in Series C funding to develop optical interconnect solutions for gigawatt-scale AI data centers. The round, led by Atreides Management and Valor Equity Partners, brings the company’s total funding to $188 million. Strategic investors NVIDIA and Intel Capital joined alongside existing backers like Sutter Hill Ventures and Maverick…
Key points
- CScale raises $145M Series C, total funding now $188M, with NVIDIA and Intel Capital as strategic investors
- Startup targets optical interconnects for gigawatt-scale AI data centers to reduce latency and failures
- CEO Martin Lund says design prioritizes continuous operation despite optical component failures
The company’s focus is on creating resilient, high-bandwidth optical links that enable seamless data exchange between thousands of GPUs and accelerators. Current AI systems struggle with latency and failures in these interconnects, leading to wasted compute cycles. CScale’s CEO, Martin Lund, emphasizes that optical failures are inevitable but argues that compute operations should remain uninterrupted. The startup’s architecture aims to isolate faults without disrupting overall system performance, a critical requirement for next-generation AI data centers. With approximately 85 employees, CScale combines expertise from industry veterans, including Lund (former Cisco executive) and CTO Sanjai Kohli (co-founder of SiRF and Inovi).
CScale Raises $145M Series C to Build Resilient Optical Interconnect for Gigawatt-Scale AI
Unite.AI · 30 September 2026
Loading the full article…
This text was published by Unite.AI. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
2sourcesThe headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Hardware & Compute
All →- Cerebras CEO Feldman debates AI scaling limits at Disrupt 2026 · 1 src
- Deepseek releases open-source TileLang for Huawei Ascend chips · 2 src
- Artificial Analysis releases open-source tool to benchmark local AI agents on laptops and workstations · 1 src
- Ollama cheat sheet guides local AI model management · 1 src
- OpenAI VP details Jalapeño ASIC’s AI-assisted design and efficiency focus · 4 src
Comments
via GitHub Discussions