A Stanford professor is now running a telecom standards forum called xRAN, and that detail alone tells me the ground is shifting. Dr. Sachin Katti is directing the xRAN Forum, launched this month by AT&T, Deutsche Telekom, and SK Telecom, and the pitch is blunt: the radio access network should be software running on ordinary servers, with open interfaces between the pieces, instead of a sealed box you buy whole from one vendor.

I have been waiting for something like this since I read China Mobile’s C-RAN paper a few years ago. That one was about pulling the baseband out of the tower and pooling it centrally, mostly to save on electricity. It solved a power problem and left a bigger one untouched: the pooled baseband was still proprietary hardware from a single supplier. Centralizing your dependence is not the same as removing it. xRAN goes after the dependence itself.

The core move is to split the base station along clean seams and publish those seams as interfaces anyone can build to. Decouple the control plane from the user plane. Define an open interface so a radio from one company can talk to baseband software from another, and turn the monolithic eNodeB into modules that different vendors can supply. If it works, an operator stops buying a network and starts assembling one.

What makes me take it seriously is who signed up. AT&T and Deutsche Telekom are not startups looking for attention, they are two of the largest operators on the planet, and they are tired of a market where a handful of vendors set the terms. Put a Stanford lab in the middle to referee the interfaces and you have something that looks less like a vendor consortium and more like an attempt to rewrite the supply chain.

There is a second thing happening in parallel that I cannot ignore. Back in February, Facebook launched the Telecom Infra Project, and its OpenRAN work is chasing a similar goal from the opposite end, focused on getting cheap gear into the ground in places the big vendors underserve rather than on clean academic interfaces. SK Telecom is in both camps, which tells me the operators are hedging. Nobody knows yet whether the open network arrives as a specification or as a deployment.

I will say the obvious risk out loud. Open interfaces are easy to announce and brutal to make real, because the fronthaul between radio and baseband is where latency goes to die, and a spec on paper is not a working link. Everyone remembers pledges of open telecom interfaces that quietly became one vendor’s flavor. And 5G is coming down the pipe fast enough that xRAN has a narrow window to matter before the next generation locks its own architecture in.

Still, I did not expect the disaggregation push to arrive with a university name attached and two tier-one operators behind it. That combination is new. I am going to be watching whether anyone actually ships an interface, or whether this is one more forum that produces slides.

Sources