The wireless industry has been buzzing with speculation and wild theories about AT&T’s plan to push Nokia out of their network. My team has been watching this, and while other analysts have posed speculations about the reasons why, we have actually talked to the players and have been able to determine the details about what’s really happening.
1. First of all, we have been quiet about this deal because we heard multiple rumors about Ericsson financing the buildout for AT&T, or taking on a major financial burden to incentivize AT&T to sign this deal. Quarterly earnings for Ericsson and AT&T came out this week, and I see no evidence of any material financial incentive. In fact, AT&T stated that they will be reducing their exposure to vendor financing. Our conclusion is that AT&T is taking this step for operational reasons, not financial reasons.
2. Secondly, this is NOT a cloud RAN deal per se. AT&T plans to migrate to a virtualized mobile network, but this deal with Ericsson doesn’t result in any immediate progress down that path. AT&T will be deploying dedicated hardware from Ericsson in 2024/2025, and will start deploying cloud RAN in 2026. There are operational reasons as well as supply chain reasons for a two-year delay here.
3. This is an open RAN deal. Despite their decision to homogenize on a single RAN software approach, AT&T will be free to buy from many hardware suppliers, in the short term for radios and in the long term for servers. Ericsson, Fujitsu and probably others will compete to provide radio hardware, presumably using the Ericsson ULPI (uplink performance improvement) partitioning scheme, even though there’s no O-RAN Alliance standard yet. The O-RAN Alliance launched this activity as a new work item in July. Similarly, virtualization will eventually allow AT&T to buy computing hardware from Dell, HPE or any other server manufacturer — or to simply outsource RAN computing to hyperscaler partners.
4. This deal will not follow the conventional logic that open RAN saves CAPEX for the operators. Replacing their new Nokia radios won’t save money on CAPEX. But, crucially, the OPEX savings may be significant. This is an opportunity to collapse five or six frequency bands into a smaller total footprint on the tower, avoiding rent increases of $1,000 per month per sector or more in some cases. Energy savings will also come into play. We don’t believe that OPEX savings is enough to justify the $14 billion deal, but it helps.
5. A more important rationale comes down to a desire for a single software platform. AT&T wants “one throat to choke," not just for RAN integration, but for network slicing and edge computing integration with the RAN. AT&T is just starting a journey where they will offer premium services to nationwide customers. When Domino’s wants a network slice to deliver pizzas anywhere in the USA, AT&T doesn’t want to develop the RAN slices with two teams, on two different software platforms. In this way, AT&T is expecting significant savings in R&D costs associated with future development — and lower operational costs in maintaining their network.
6. There has been some silly talk about a negative impact on CommScope, as the Nokia configurations often use CommScope antennas. On the contrary, I see this deal as a new opportunity for CommScope, as they’re the only antenna vendor left on the independent market that can integrate active and passive antennas. Achieving OPEX savings in a complex network will depend on FDD/TDD integration, so strong antenna design will be more important than ever.
The big Q
Here’s the billion-dollar question: Will other operators follow AT&T?
I believe that the answer is “yes, but not right away.” AT&T is gambling here, hoping to simplify the development of premium 5G services by cutting the required RAN development in half. But they have no experience yet with vRAN, let alone cloud-native RAN. They have several hurdles to jump over before they can achieve the Open Cloud RAN vision. Their direction is right, but their timing requires them to start with purpose-built hardware and then awkwardly change over to cloud RAN later. This is all reflected in our forecasts on Virtual RAN and Open RAN.
Most other operators will wait until they have completed a series of field trials on cloud RAN and open RAN, then migrate their network to a single RAN software vendor. Few operators can afford to duplicate AT&T’s expensive approach of ripping out hardware to brute-force the network into a homogeneous platform.
If AT&T’s gamble is successful, they could find themselves ahead of T-Mobile and Verizon in 2027-2028, with an ability to develop end-to-end slices quickly, and to offer nationwide premium services without long development delays.
Joe Madden is principal analyst at Mobile Experts, a network of market and technology experts that analyze wireless markets.
"Industry Voices" are opinion columns written by outside contributors—often industry experts or analysts—who are invited to the conversation by Fierce staff. They do not necessarily represent the opinions of Fierce Wireless.