Six months ago I said the moment I was really waiting for was an incumbent backing Open RAN for real, instead of nodding along in public and hoping it lost momentum. AT&T and Ericsson just handed me a version of that moment, and it is stranger than I wanted it to be. The deal announced this week is fourteen billion dollars over five years, Ericsson beat Nokia to it, and AT&T is now talking about pushing something like seventy percent of its wireless traffic through open-capable platforms by the end of 2026.

This is brownfield, which is the whole point. Rakuten had a blank island and 1&1 is a fourth operator building from zero. AT&T is one of the three or four largest carriers on earth committing to rebuild a live, national network around open interfaces, and that is the thing nobody had actually done. When AT&T put its name on the alliance in 2018, it cost the company a press release. Five years later it costs fourteen billion, which is a much more honest way to find out whether a founder meant it.

Here is what unsettles me. The deal is a fourteen-billion-dollar award to a single vendor. AT&T used the vocabulary of openness to consolidate onto Ericsson and shove Nokia out of a large slice of its network. An Open RAN contract whose immediate effect is to cut the supplier count from two toward one. I kept circling this same paradox in earlier posts without naming it cleanly, and here it sits in one purchase order: openness working as a procurement lever rather than an outcome. The open sockets are real and will matter later, on the day AT&T can drop a Fujitsu or a Dell radio into an Ericsson-managed network without a forklift. On day one, though, the money bought consolidation.

There is a policy layer under all of this too. Washington has stood up a fund worth well over a billion dollars to seed a Western open ecosystem, and a Tier-1 carrier finally spending gives that money something concrete to point at. It is landing into a rough market, mind you. The RAN business is shrinking, Ericsson and Nokia are both cutting thousands of jobs, and an industry gets meaner about defending its installed base when the pie is getting smaller, not larger.

What I cannot tell yet is whether the open-capable part ever gets exercised. AT&T could build the entire thing on Ericsson and quietly never plug in a second radio vendor, in which case this was an extremely expensive rebrand of a normal upgrade. Or it opens the sockets for real and a genuine multi-vendor network shows up in American cities for the first time. I have been watching this movement since a Chinese white paper about saving electricity in 2010, and I still could not tell you which way it breaks. That is the thing I want to pin down the next time I look at where all of this actually stands.

Sources