Relativity Networks raises $22 million to bring a faster kind of fiber to data centers
As the global race to build out AI infrastructure intensifies, developers are projected to pour as much as $4 trillion into data centers by the end of the decade. However, these massive projects are increasingly hamstrung by rigid constraints, including power grid limitations and complex political hurdles. While the industry has largely accepted the speed of traditional fiber as a static variable, one startup is betting that a breakthrough in transmission technology could fundamentally rewrite the geographical rules of data center construction.
Relativity Networks announced on Tuesday that it has secured $22 million in SAFE note funding. The round was led by Rhapsody Venture Partners, with participation from Bell Ventures Inc. and Faster Than Glass LLC, among others. In addition to the capital injection, the company confirmed a significant $40 million follow-on order from a major, unnamed hyperscaler.
Redefining Speed with Hollow-Core Fiber
At the heart of Relativity Networks’ value proposition is hollow-core fiber, a specialized technology that remains rarely deployed in modern networking. Unlike conventional fiber-optic cables, which transmit light through solid glass, hollow-core fiber guides light through a vacuum chamber located in the center of the cable. By removing the glass medium, the signal travels much closer to the theoretical speed of light.
The performance gains are measured in microseconds, but in the world of high-frequency compute, these fractions of a second are transformative. CEO Jason Eisenholz notes that while a signal typically takes five microseconds to travel one kilometer in standard fiber, hollow-core technology slashes that latency to just 3.5 microseconds.
Why Latency Matters at Scale
In the early days of AI, when compute was confined to a single rack of GPUs, fiber latency was a negligible concern. Today, however, the physical scale of data centers has ballooned, with campuses sprawling across hundreds of acres and dozens of distinct buildings.
“The largest systems are distributing the compute across multiple campuses to reach the power that exists,” Eisenholz explained. “They’re moving to where the warm shell is, but they still need to operate as one synchronized machine.”
By reducing latency by 30%, Relativity Networks provides developers with a critical advantage: the ability to span significantly larger distances between facilities without sacrificing performance.
The Third Era of AI Infrastructure
According to Eisenholz, the industry is currently transitioning through three distinct phases of optimization:
- Era 1: Optimization focused purely on raw compute power (the "GPU, GPU, GPU" phase).
- Era 2: Optimization of internal data center networking to maximize that compute.
- Era 3: Optimization of geography.
By alleviating the spatial constraints that have traditionally limited where data centers can be placed, Relativity Networks aims to help hyperscalers build more efficiently. As compute projects continue to scale to unprecedented levels, the ability to knit together disparate campuses into a single, synchronized unit could prove to be a major turning point for the future of digital infrastructure.