Relativity Networks' hollow-core fiber bet: $22M round, $40M hyperscaler order

Relativity Networks is an Orlando startup selling ChronoCore hollow-core fiber — cable that guides light through an air core instead of glass, cutting propagation latency by roughly half — so AI data center campuses can stay synchronized across greater distances.

That as AI compute sprawls across campuses, latency becomes the wall — hollow-core fiber cutting travel time ~50% lets data centers span farther before sync breaks.

$4.6M pre-seed announced Feb 2025, with total funding approaching $11M by July 2025 including Prysmian support; $22M SAFE notes announced Aug 2026, alongside a $40M follow-on order from an unnamed leading hyperscaler (TechCrunch, 2026-08-19; optics.org).

Jason Eichenholz, former CTO of lidar maker Luminar and a CREOL graduate, founded Relativity Networks in 2023 with UCF optics professor Rodrigo Amezcua Correa, whose anti-resonant hollow-core fiber designs came from a decade of CREOL research. The thesis: the first AI era optimized compute, the second optimized in-building networking, and the third — the AI Geography Era — optimizes distance, with faster fiber as the lever.

The company raised a $4.6M pre-seed by Feb 2025, deployed field installations in the US with multi-million-dollar contracts, and reached nearly $11M total funding by July 2025 with support from Prysmian, the world's largest cable maker. In August 2026 it announced $22M in SAFE notes from Rhapsody Venture Partners, Bell Ventures and Faster Than Glass, plus a $40M follow-on order from an unnamed leading hyperscaler, after testing a 24-fiber ChronoCore cable installed in standard microducts with Dura-Line.

Scaling from pilot to product: a hyperscaler order in hand, manufacturing at scale with Prysmian, and positioning against roughly $4T of projected data center spending this decade — but the technology must move from test routes to wide carrier-grade deployment.

The bet targets an overlooked fixed layer: hyperscalers spend on GPUs and networking, while Relativity changes the medium itself — a wedge competitors cannot copy with software.

The math is concrete and testable: 5 to 3.4 microseconds per kilometer means roughly 50% more reach before latency breaks synchronization, giving the pitch a number.

A $40M order from an unnamed hyperscaler plus manufacture at scale with Prysmian moved the story from lab tech to supply chain, the usual graveyard of fiber startups.

The timing rides the AI buildout's real constraint — power scarcity — rather than a hype cycle, because campus sprawl is already happening.

When everyone optimizes the loud layers — GPUs, switches, software — the overlooked fixed layer is the opportunity: faster fiber changes where AI factories can be built.

As of 2026-09-02 Relativity Networks is live and scaling from pilots: an unnamed leading hyperscaler placed a $40M follow-on order, and the company is manufacturing fiber and cable at scale with Prysmian while qualifying installation partners for carrier-grade routes. The company frames its platform as the 'Scale Across' layer for the AI Geography Era, with a white paper on the synchronization budget. Total disclosed funding is about $33M ($11M through July 2025, $22M in August 2026 SAFE notes), against a data center market projected to spend up to $4T by the end of the decade.