Nvidia and Intel Capital Invested in CScale
The strategic investment targets AI infrastructure reliability as gigawatt-scale data centers emerge.
Updated on Sept. 30, 2026 in Data Centers
Live Poll
Do you expect the cost of new AI-driven technology to make personal electronics more expensive?
Nvidia and Intel Capital have become strategic investors in CScale, a company focused on AI infrastructure technology. CScale aims to address the challenges of interconnect bandwidth and potential optical failures in large-scale computing environments.
Why it matters
As AI clusters grow, optical failures can threaten compute stability at a fleet-wide level. Securing reliable interconnect systems is essential for the future deployment of massive, gigawatt-class data centers.
CScale uses a combination of photonics, electronics, and software to architect interconnects that contain optical failures. This technology is designed to maintain compute operations even when individual system components fail.
The players
Nvidia
Nvidia is a leading technology company that designs graphics processing units and hardware for artificial intelligence applications.
Intel Capital
Intel Capital is the venture capital arm of Intel Corporation, which invests in startups and technology companies across the global tech industry.
CScale
CScale is an infrastructure startup founded in 2023 that specializes in AI interconnect technologies and photonics.
Sanjai Kohli
Sanjai Kohli is the founder and CTO of CScale who previously co-founded SiRF, a company known for GPS technology.
Facebook is a global social media and technology company now operating as Meta that frequently acquires smaller engineering firms to bolster its infrastructure.
The details
Founded in 2023, CScale focuses on building robust architectures capable of supporting the high-bandwidth demands of modern AI hardware. By isolating failures, the company seeks to prevent single-point disruptions from cascading across hundreds of thousands of AI accelerators.
Timeline
Facebook acquired Inovi in 2014.
CScale was founded in 2023.
Nvidia and Intel Capital invested in CScale in September 2026.
Gigawatt-class AI datacenters are expected to appear by the end of the decade.
The Tech Race
The transition toward gigawatt-class AI datacenters represents a significant shift from localized server racks to massive, interconnected computational fleets. By securing infrastructure reliability, firms like CScale are positioning themselves to support the next generation of global AI workloads.
As data centers become more reliable through these architectural improvements, users can expect more consistent performance from large-scale AI applications. Enhanced stability reduces the likelihood of service interruptions for major cloud-based artificial intelligence tools.
The takeaway
Reliable hardware architecture is becoming the limiting factor in the rapid expansion of artificial intelligence capabilities. Investors are increasingly focused on the underlying photonic and interconnect technologies that keep these massive computing systems running without failure.
Further reading
For more context on the evolving requirements for large-scale AI facilities, explore the Data Centers section.
Live Poll
Do you expect the cost of new AI-driven technology to make personal electronics more expensive?
