Point2 Technologies Extends Series B to $76M to Build RF Interconnects for AI Data Centers
Inside AI data centers, the real tension is not in the models, it is in the milliseconds between them. Signal trying to outrun heat, bandwidth pushing against physics, cost creeping in every time data hesitates. Point2 Technology just leaned into that pressure, extending its Series B to $76M, led by Maverick Silicon with NVentures and UMC Capital stepping in like they know exactly where this story is headed
Sean Park, CEO, and Thomas Byunghak Cho, CTO, are not chasing headlines, they are chasing constraints. Alongside them, HM Bae, Founder and Chairman, plus a bench that includes Reza Norouzian, Chief Strategy Officer, Jake Eu, SVP Engineering, and Gus Lignos, SVP Worldwide Sales, are building around a simple reality. In AI infrastructure, the constraint is no longer compute alone. It is the connective tissue. The wires, the latency, the power draw that quietly taxes every breakthrough. Point2 Technology is leaning into that pressure with RF based interconnects that treat distance and efficiency less like limits and more like variables to bend.
E Tube sounds almost polite until you realize what it is doing. RF over plastic dielectric waveguide, built for multi terabit throughput inside environments where every nanosecond compounds into real dollars. This is not a cosmetic upgrade. This is about rethinking how racks talk, how clusters scale, how hyperscale operators stop burning power just to keep signals alive long enough to matter.
The investor mix tells its own story. Maverick Silicon leading, NVentures bringing NVIDIA’s proximity to the problem, UMC Capital tying into manufacturing gravity. That is not passive capital. That is alignment around a very specific choke point in the AI stack. When the people building chips and fabs start leaning into your interconnect layer, you are no longer a side character.
What stands out is how this round came together. Not noise, not theatrics. Just a company that kept building through the hard part, proved enough signal in a space full of skepticism, and earned the right kind of attention. Prior extensions, strategic partners like Molex, and now a cap table that reads like a map of the semiconductor supply chain. That is how you stack credibility without announcing it every five minutes.
For operators and builders, the takeaway is simple and uncomfortable. You can scale compute all day, but if your interconnect cannot keep up, you are just creating expensive traffic jams. Point2 Technology is betting that the next leap in AI performance will not come from more chips alone, but from how intelligently those chips are connected.
And if they are right, this is not just a funding round. It is a signal that the real race is happening in the spaces between the hardware, where speed, power, and latency decide who actually wins when the models start running hot.
