Data center developers are expected to invest as much as $4 trillion by the end of the decade, but finding enough power and suitable locations is becoming increasingly difficult. Relativity Networks is betting that faster fiber could help data centers overcome some of those geographic constraints.
The company announced $22 million in SAFE note funding from Rhapsody Venture Partners, Bell Ventures and Faster Than Glass, among others. It also secured a $40 million follow-on order from an unnamed hyperscaler.
Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram
Hollow-Core Fiber Cuts Latency
Relativity Networks develops hollow-core fiber, which can transmit data 50% faster than conventional fiber. Instead of sending light through glass, the technology uses a hollow center containing air or a near-vacuum, allowing signals to travel closer to the speed of light.
For a signal traveling one kilometer, CEO Jason Eichenholz estimates that conventional fiber takes about five microseconds. Hollow-core fiber can reduce that to roughly 3.5 microseconds.
Connecting Data Centers Across Greater Distances
As AI systems grow, computing resources are increasingly spread across large campuses and multiple facilities. That makes network latency more important, particularly when separate data centers need to function as a single synchronized system.
Eichenholz sees an opportunity to connect existing facilities across multiple campuses, allowing companies to locate computing infrastructure closer to available power without sacrificing as much performance.
A 50% reduction in latency could effectively allow compute facilities to be separated by greater distances while maintaining similar networking performance.
From Compute To Geography
Eichenholz describes the evolution of AI infrastructure in three stages: first optimizing computing power, then improving networking within data centers, and now optimizing where computing infrastructure is physically located.
For Relativity Networks, faster fiber could therefore become an important piece of the next phase of AI infrastructure expansion.







