Relativity Networks raises $22M for faster data center fiber

Data center developers are expected to spend as much as $4 trillion by the end of the decade, and they are already constrained by political and power-grid considerations in where they can build. Relativity Networks is betting that faster fiber can loosen some of that geographic math.

On Tuesday, Relativity Networks announced $22 million in SAFE note funding, drawn from Rhapsody Venture Partners, Bell Ventures Inc., and Faster Than Glass LLC, among others. A SAFE note converts into equity once a company closes its first priced funding round — a standard mechanism for pre-seed and seed-stage startups. The company also secured a $40 million follow-on order from a leading hyperscaler that declined to be named for the story.

Relativity Networks builds hollow-core fiber, a rarely deployed technology that transmits data 50% faster than conventional fiber. Standard fiber sends light through fiber-optic glass; hollow-core fiber instead sends the same light through a vacuum chamber running through the center of the line, pushing transmission speed closer to the theoretical limit of light.

The gap shows up in microseconds. CEO Jason Eichenholz estimates a signal takes roughly five microseconds to travel one kilometer over conventional fiber, a figure hollow-core fiber can cut to only three and a half microseconds.

That difference barely mattered when AI compute ran on a single rack of GPUs. But as AI infrastructure has scaled, so has the physical distance between GPUs — data center campuses now commonly sprawl across hundreds of acres and dozens of buildings. Eichenholz sees a particular opportunity in multi-campus deployments, where operators knit together existing data centers to function as a single system.

“The largest systems are distributing the compute across multiple campuses to reach the power that exists,” he told TechCrunch. “They’re moving to where the warm shell is, but they still need to operate as one synchronized machine.”

Cutting latency by 50% also means operators can space data centers 50% farther apart before latency becomes a limiting factor, loosening some of the spatial constraints that have shaped recent buildouts. Eichenholz argues this shift in geography could become the industry’s next major front.

“The first era of AI optimized for compute,” he said. “It was GPU, GPU, GPU. The second era optimized the networking inside the data center to take advantage of that compute. The third era that we see coming is optimizing the geography.”

Why it matters

Faster fiber doesn’t add compute capacity, but it changes where AI infrastructure can physically be built — potentially easing the power and land constraints that have slowed data center expansion.