Scale-up interconnect startup CScale launches with $188M in funding
CScale Inc., a startup that develops optical interconnects for data centers, launched today with $188 million in funding. The company raised about three-quarters of the capital through a recently closed Series C round. It was co-led by Atreides Management, Valor Equity Partners and Premji Invest. The funds were joined by several other backers including Nvidia […] The post Scale-up interconnect…
CScale, an interconnect startup specializing in optical technology for data centers, has secured $188 million in funding to fuel its growth. The substantial capital injection was primarily raised through a closed Series C round, with notable backers including Atreides Management, Valor Equity Partners, Premji Invest, Nvidia Corp., and Intel Capital.
Founded in 2023 by electrical engineer Sanjai Kohli, CScale's mission is to address the limitations of current AI infrastructure by developing high-speed interconnect solutions. Kohli, who previously led SiRF Technologies, serves as the company's CTO, while long-time data center executive Martin Lund heads the executive team, bringing expertise from Cisco Systems and Broadcom.
The startup's technology is specifically designed to interconnect accelerators within AI clusters. Unlike traditional setups where data moves slower between different domains, CScale's optical interconnects enable faster data transfer within a single domain, optimizing performance. Nvidia's recent efforts to consolidate multiple domains into one large domain aim to further enhance data flow speed.
CScale's optical interconnects use optical links to transfer data between chips, providing superior speed and reliability compared to traditional copper wiring. In the event of an optical link failure, CScale's technology is said to "contain" the issue, minimizing its impact on the entire cluster and boosting uptime.
The company plans to utilize its new funding to refine its technology and hasten its commercialization efforts. Each CScale system is expected to incorporate a system-on-chip with an onboard central processing unit, potentially running network automation software. The startup describes its product as an "integrated light engine," emphasizing the importance of reliability in the rapidly evolving AI infrastructure landscape.
Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.