Relativity Networks Raises $22M to Make AI Fiber Faster

Relativity Networks is building for a problem created by AI’s appetite for power. As data-center campuses grow larger and compute is pushed toward locations where electricity, land and infrastructure are available, the physical distance between machines becomes harder to ignore. The Florida-based company believes one of the answers lies inside the fiber itself: moving light primarily through air instead of solid glass so data can travel between computing sites with less delay.

The company has raised $22 million through SAFE financing, with backing from Rhapsody Venture Partners, Bell Ventures Inc., Faster Than Glass LLC and other investors. The new capital arrives alongside reported commercial momentum, including a $40 million follow-on order from an unnamed hyperscaler, giving Relativity Networks an important test of whether a technology developed through years of optical research can become infrastructure deployed at data-center scale.

Relativity Networks was founded by Dr. Jason Eichenholz and Dr. Rodrigo Amezcua Correa, combining commercialization experience with deep photonics research. Jason Eichenholz previously co-founded lidar company Luminar Technologies, while Rodrigo Amezcua Correa pioneered anti-resonant hollow-core fiber designs at CREOL, the University of Central Florida’s College of Optics and Photonics. Their work builds on more than a decade of research into changing how light moves through optical fiber.

Conventional fiber guides light through solid glass, where propagation is slower than through air. Relativity’s hollow-core approach instead guides light through an air-filled center. The company says this can reduce propagation time from roughly 5 microseconds per kilometer to about 3.4 microseconds, translating into approximately 32% lower latency over the same distance.

Those microseconds become more valuable as thousands of accelerators have to stay synchronized across buildings or separate campuses. Faster transmission can give operators more freedom to place computing capacity near available power while preserving tighter communication between sites.

Relativity Networks is ultimately testing a bigger infrastructure idea: as AI scales beyond individual buildings, the material carrying the data may become just as strategically important as the switches, chips and software sitting around it.