Tuesday, November 17, 2026

Vodafone is putting RAN GPUs to the grill


On 17 November, Iain Morris reported in Light Reading that Vodafone has not yet approved GPUs for its radio access network. Kyriakos Exadaktylos, the group's head of RAN evolution, told an Omdia AI-RAN symposium that Vodafone is building labs with advanced capabilities at its Malaga innovation campus to measure the benefit on network efficiency first, and that "we have not, at the moment, stamped that we have a green light" for GPU deployments. He is, at the same time, convinced that AI will improve the RAN significantly.

This is the third data point in a few weeks pointing the same way. In London, Virgin Media O2 (a Telefonica and Liberty Global joint venture, not a Vodafone company, for those keeping score) ran the AI-RAN trial I covered here and here, which delivered a 20% spectral efficiency gain on Nokia's Marvell based custom silicon, the same figure Nokia had talked about achieving this year on Nvidia GPUs. Vodafone's position, as reported, is that existing algorithms with AI embedded will not require new silicon, and that more advanced models are 2 to 3 years away and still need to be validated.

The technical argument is simple enough. Link adaptation, the function that picks modulation and coding for each user, has to decide in about 40 microseconds according to Gabriel Foglander, Ericsson's head of RAN strategy, and Ericsson already runs a neural network for it on its own ASICs. Models that must answer that fast are small and need deterministic timing, which is what purpose-built silicon does well. CPUs, Arm, GPUs, ASICs, etc... the RAN has seen every hardware pitch in the last decade, and the small model problem favours the incumbent chip.

The economic argument runs the other way, and Ericsson's own executive made it. RAN silicon only sells into the RAN market, which Omdia sized at about $35 billion last year (a small market by the standards of the chip vendors now courting it). If AI software evolves faster than ASIC generations, operators will have to replace baseband more often, and Foglander pointed out that installation, commissioning and truck rolls are a substantial part of that capex. A GPU, oversized for today's models, would last longer in the field.

In my mind, this moves the debate away from the chip. The ASIC wins today on a 40 microsecond task, so the real question is how often an operator will accept a baseband refresh to keep pace with AI models, and at what cost per site. Neither Ericsson nor Nokia has published that cadence or that cost for either path. The 20% figure, measured against an unstated baseline, does not prove one silicon beats the other: it shows the gain came from the software, which is precisely the case for keeping the installed hardware. I think the burden of proof sits with the GPU camp, and a lab like Malaga is where it will be met or not, as mentioned before.

Nokia has the most exposure. It has committed its future RAN products to Nvidia GPUs after Nvidia's $1 billion investment, while its 7,000 site VodafoneThree contract in the UK had only 10 sites activated by April, according to Light Reading. What Vodafone learns in Malaga will decide whether that network is built on Marvell silicon, on GPUs, or on a mix of both.

It will be interesting to see whether Vodafone publishes its Malaga results, baseline included, before VodafoneThree has to make that choice.

Monday, November 16, 2026

Rakuten Mobile's RIC tilts antennas, not economics


Rakuten Mobile reported its third-quarter results on 13 November, and Iain Morris's account in Light Reading the same day is mostly about sites. When Rakuten started out, its executives expected to need 40,000 to 50,000 sites to cover Japan, with simplified radio designs adding 30% to 40% of coverage compared with traditional equipment. It operates 106,000 today, and since KDDI decided in August to end its roaming agreement, Rakuten is building more.

The capex follows. Rakuten's telco business invested about JPY63 billion in 2025 and needs about JPY200 billion this year, of which JPY105.7 billion had been spent by the end of September. The remaining JPY94.3 billion is about 3.3 times what it spent in the final quarter of 2025. Customer additions fell to 297,000 in the quarter from 405,000 a year earlier, which Rakuten blames on concerns about roaming reduction, leaving it with about 11 million customers and a 5% share of the market. Revenue grew 17.1% to JPY111.4 billion and the operating loss narrowed to JPY30.3 billion from JPY37.3 billion, but Light Reading puts the group's cumulative net losses since 2019 at about $7.9 billion.

Rakuten still points to its open RAN. It uses its RAN Intelligent Controller to host software that adjusts antenna tilt without manual intervention, to improve coverage and capacity. Morris notes that the same feature is available from traditional vendors to Japan's other operators, and concludes that virtualization and open RAN have been massively oversold.

I would put it differently. Rakuten proved that a cloud native mobile network, with virtualized RAN, cloud native core, automated operations, etc... can be built and run at national scale, which is what I wrote in 2020, and that part has held. What it did not prove, and was never going to prove, is that software substitutes for spectrum. Rakuten holds 186MHz, about 100MHz of it at 3.5GHz (a poor band for wide area and in-building coverage), against more than 250MHz for SoftBank. No controller recovers the propagation of low-band frequencies an operator does not own.

Automated antenna tilt is not new either. It is a centralized SON function, the ancestor of today's rApps, as mentioned at FYUZ 2 weeks ago. It makes existing sites work better. It does not make a cell reach further than its band allows, and it does not replace a roaming partner who walks away.

In my mind, the 106,000 figure says nothing about whether open RAN works technically. It says the original coverage model rested on an optimistic link budget and on a roaming agreement that would last, and both assumptions were commercial, not architectural. The site count is the cost of spectrum scarcity, and it shows up in the capex line rather than in the architecture diagram. RIC automation delivers real, measurable gains, but they are measured as a percentage of an existing network's performance, not in sites avoided, and the 30% to 40% coverage gain promised 6 years ago was never within a controller's reach.

It will be interesting to see whether Rakuten's next results report the RIC by the coverage it adds, or only by the manual interventions it saves.

AI Gigafactory Northern: Telia brings the fibre


On 16 November, Telia announced it has joined Nokia, British neocloud Nscale, Finnish AI cloud provider Verda and Estonian power systems specialist Skeleton Technologies in a consortium called AI Gigafactory Northern, to bid for the European Commission's AI gigafactory programme. TelecomTV's Ray Le Maistre and Light Reading's Paul Rainford both reported it the same day. The design is federated: facilities in Oulu in Finland, the Tallinn region in Estonia, Valmiera in Latvia and a Swedish site still to be named, with VTT and KTH as research partners.

The Commission aims to mobilise €30 billion across 7 facilities, up to €10 billion from the Commission and national public sources and €20 billion from the private partners in the winning bids. Two thirds of the headline is therefore private money the winners still have to raise, so the €30 billion is a target, not a budget. Bids for large facilities are due on 3 December and for medium ones on 4 February 2027, with winners named in the first half of 2027. Telia joins Telefónica (with the Spanish government and Santander), Orange and Iliad's Scaleway in the AION consortium, Vodafone, linked to a bid with Hochtief, and Deutsche Telekom, still weighing the economics after expanding its own Munich AI factory.

What Telia says it brings is telling: connectivity, data transmission, infrastructure operations and sovereign cloud, through its Telia Cygate integration arm. GPUs are not on the list. This is the role AT&T described for itself last month, fibre and managed connectivity rather than GPU landlord, and I think it is the right one for a mid-sized operator. Telia's balance sheet is not built to carry a large share of €20 billion of private capital, and neoclouds such as Nscale exist to take exactly that risk.

The federated design is where it gets interesting. The consortium says spreading capacity over 4 locations reduces concentration risk, draws on several sources of decarbonised energy and lets it build where power, grid connections, land, cooling, etc... are available. That is sound for energy procurement and resilience. It is less obvious for what a gigafactory is meant to do, which is train very large models. Large training runs want tightly coupled GPUs on a single low latency fabric. Spread them over Oulu, Tallinn and Valmiera and either each site runs its own jobs (in which case the consortium has 4 AI factories, not a gigafactory) or the inter-site network has to behave like a backplane. This is the fabric versus location question I raised in July, and it lands squarely on Telia's part of the bill of materials.

In my mind, that is both the opportunity and the exposure for Telia. If the federated model works, the operator's fibre and operations are what turn 4 buildings into one facility, which is a far more defensible position than renting racks. If it does not, the consortium ends up with a regional inference and sovereign cloud footprint carrying a gigafactory label, which is a reasonable business, just not the one in the Commission's brief. The announcement gives no capacity in MW or GPUs, no split of investment between partners and no design target for inter-site latency, so it is too early to tell which one is being bid.

It will be interesting to see, when the winners are named next year, whether the Commission scores distributed designs like this one as gigafactories or as a collection of AI factories with good fibre.

Wednesday, November 4, 2026

Here we go again: The Killer App syndrome, featuring ISAC in 6G


I am in Seattle this week for FYUZ, the TIP event and if there is one acronym on every 6G slide from the vendors and the US operators, it is ISAC. AT&T called integrated sensing and communication "the big shiny object in 6G" from the stage, as Mobile World Live reported on November 4, and Samsung, Ericsson and the rest of the usual suspects have a keynote, a demo or a birds of a feather session on it before the show closes on Thursday. Having spent a career watching each generation of cellular get designed around one headline use case, I have seen this movie before.

ISAC, in plain terms, is the use of the signals a base station already radiates to detect objects that are not connected to the network, the way a radar does. The echoes of the downlink reference signals, received by the same base station (monostatic) or by a neighbouring one (bistatic), are processed to extract the range, speed and direction of a drone, a vehicle, a crowd, etc... 3GPP added a sensing channel model in Release 19 and published in June 2026 the Release 20 study (TR 38.765) on base station sensing of drones in FR1, reusing the CP-OFDM waveform, while SA2 drafted the architecture in TR 23.700-14. Full ISAC is a 6G affair, Release 21 and later, which is why AT&T wants a limited version on 5G Advanced rather than wait, and why Ericsson told Fierce Network in July it expects small scale deployments within two years, the Los Angeles Olympics of 2028 being the stated target.

This is where my skepticism comes from. 2G was designed for digital voice, and texting, an afterthought squeezed into the signalling channel, became the killer app and the most profitable bits the industry has ever carried. 3G was sold on video calling and push to talk, and what actually paid, modestly, was MMS (I ran the product team that shipped the first commercial MMSC at Logica, so I remember how modestly). 4G was architected around IMS and VoLTE, and the application that filled the pipes was video from YouTube and Netflix, which no operator designed for and none monetized. 5G was built around network slicing and the enterprise connectivity products it was supposed to enable, and seven years in, the only 5G product with a revenue line of its own is fixed wireless access, a service that was in nobody's keynote in 2018, as I wrote here and here. In each generation, the standard was optimized for the use case the industry wanted, and the market picked another one that the standard merely tolerated.

What is different this time is who is doing the pushing. The US Department of Defense's FutureG office calls ISAC a "mission-first priority" against drone swarms (Tom Rondeau, July 2026) and awarded Cohere Technologies 28 million dollars to turn commercial cell sites into drone detectors, Ericsson Federal demonstrated drone detection in Plano in February with the same office on the press release, and Verizon's trial with Samsung at a soccer fan event in Dallas and its FCC experimental licence in 4.8 to 4.9 GHz, reported by Light Reading in September, both point at NTIA's Mission 6G 28 showcase at the Los Angeles Olympics, an initiative NTIA describes as industry led, industry planned and industry funded. Ericsson Federal's CEO put it plainly to Mobile World Live in October: sensing gives the US military a reason to invest in commercial cellular infrastructure beyond connectivity.

In my mind, read backwards, that sentence describes a national air defence sensor layer built on operator capex, with the requirements set in Washington and the bill presented to the operators and, eventually, their subscribers. The industry is not naive about it; AT&T's own speaker said on stage that some assume defence money will fly in through the window, and asked the right question, which is, outside defence and public safety, who pays the bill? Nobody at FYUZ had an answer beyond "new revenue streams", which is the same answer slicing got in 2018.

Because ISAC reuses the radio, the pitch is that sensing comes almost for free. It does not. Monostatic sensing means listening for an echo while the base station is transmitting, which today's TDD massive MIMO radios do not do; it requires a separate, isolated receive panel or self interference cancellation, in other words new hardware at the cell site, and the Release 20 work on base station monostatic sensing exists precisely because that architecture is not in the field. Second, every resource given to sensing is taken from communication: beam time swept across empty sky instead of pointed at subscribers, denser reference signals, wider bandwidth and longer integration windows to resolve speed, and transmissions in idle periods where energy saving would have switched the carrier off. The ISAC panel at FYUZ said as much; the test vendor on stage described sensing as blanking out resource blocks stolen from the operator, and the operator's own 3GPP delegate explained that the exposure layer needs checks and balances on every request precisely because sensing is resource intensive and cannot be an always-on API. Third, processing: the sensed data runs to terabytes, the result is worthless unless the processing sits close enough to the radio to make a decision in time, and one panelist expects purpose-built analog silicon at the cell site because the sensing runs continuously across a lot of spectrum, which he conceded will require a lot of additional money. ABI Research makes the same point, sensing is a continuous processing workload, unlike connectivity, with no revenue model to cover it. Fourth, everything above the radio is new: an exposure layer with its northbound API, about 75% specified and due in March 2027 according to the delegate, a new sensing function in the packet core to translate an external request (area, time window, detection or tracking) into the right set of base stations, the RAN procedures to go with it, roaming between networks when the requested area crosses them, and consent, since the delegate conceded that people who are not the operator's subscribers get sensed. Fifth, the one that gets no slide time, liability: a network that tracks objects that never attached to it, "operationally invisible" in Cohere's words, will be treated by regulators, courts and customers as a surveillance system, in Europe before anywhere else.

The comparison the ISAC slides never show is with what already exists. Drone detection is a market of 2 billion dollars in 2026 growing to 6 billion by 2034 according to Fortune Business Insights, radar being the largest segment, RF detection the cheapest and optical the fastest growing, and products like Dedrone's fuse all three with recognition models that tell a drone from a bird or a balloon, which an echo at 3.5 GHz cannot. Cameras with facial, scene and object recognition are deployed at scale in stadiums, airports, ports and borders, and the EU has a 250 million euro border management call for exactly this. These sensors are purpose built, mature, specialized, and what they lack, connectivity, backhaul and compute for inference, is what operators are good at providing. The operator opportunity in my mind is to connect, backhaul and host these sensors at the edge, not to rebuild a mediocre radar out of a radio designed for throughput.

In my opinion, the technology works, the AT&T and Verizon trials show it, and for a stadium, a port or a military base with a federal customer paying, a few hundred sites will carry sensing. That is a managed service for a handful of government buyers, which ABI Research itself now frames as ISAC sold inside a private network or edge platform rather than as a mainstream 6G feature, with commercialization after 2030. It is not a reason to design the 6G air interface, the radios and the baseband around sensing, and it is certainly not a reason for operators to fund national security infrastructure on their balance sheets while the government keeps the spectrum and sets the requirements. If the killer app of 6G follows the pattern of the last four generations, it will be something the standard tolerates rather than something it was designed for, and the sensing hardware will sit at the cell site the way the IMS cores sat in the data centres, waiting for a use case. Will the Pentagon pay the full price of the sensor network it is asking for, or will it, as is now customary, leave the bill with the subscribers?

Tuesday, November 3, 2026

Open RAN at 10: Who Killed the Near-RT RIC?

 I am in Seattle this week for FYUZ, opening the SMO and RIC panel. Ten years ago, at Telefonica, my team was trying to connect the unconnected in Latin America, and the traditional RAN was simply not affordable there, so we disaggregated it. We launched the TIP OpenRAN working group in 2017, co-authored the OpenRAN RFI with Vodafone in 2018 and presented the results at the TIP Summit, which has since become FYUZ (I still have the T-shirt). This conference is, in a way, where it started for me, and a good place to take stock of what ten years have delivered.

Open RAN set out to open the radio, with RU, DU and CU from different vendors on general purpose hardware. On the way, it created something else, a software layer for intelligence: the SMO, the Non-RT and Near-RT RIC (brothers from a different mother, as I called them in 2023), rApps and xApps, and more recently dApps inside the CU / DU. None of this came from nowhere, of course. Centralised SON is the ancestor of the rApp, distributed SON the ancestor of the xApp, and one vendor told me that 55% of its SON deployments are already multi-vendor. The promise was any vendor's app, on any vendor's controller, over any vendor's RAN.

For this session, I interviewed and surveyed the SMO, RIC and app vendors on where things really stand. rApps have won the application model and they are in production. The largest deployment I found runs about 12 rApps on roughly 150,000 cells, but vendor and operator sit in the same group. Third party rApps are live at US tier ones, including on AT&T's network, and one RAN vendor's energy rApp runs on a third party RIC at another tier one. What I could not find anywhere is the original promise, a third party app on another vendor's SMO over a different vendor's RAN. The leading directory lists 117 rApps, 94 of them from independent vendors, and only a handful are in public production.

Why? One independent vendor summed it up: the algorithm ports, the plumbing does not. Bringing a first app onto a new SMO costs around three FTE-months, about half for each one after, and R1 delivers standards-based portability, not plug-and-play. Every vendor I spoke to, large or small, named the same three gaps: data semantics (interfaces connect, but counters and models differ by vendor), conflicts between automation loops (in an EU testbed, two vendors' energy-saving xApps switched the same cells on and off against each other) and onboarding, certification and accountability. The industry's answer, unfortunately, is shifting from portability to certification per platform, curated catalogues rather than open marketplaces. I have seen this movie before, with MANO and then with the SMO itself: the barriers were never technical, they were commercial.


The biggest change since last year is that, of the three intelligence layers, the middle one did not survive. Regular readers know my feelings about the Near-RT RIC, and at MWC last year I called it a flawed proposition as a host for third party xApps. A tier-one SMO vendor has now discontinued investment in its Near-RT RIC and xApps, given limited industry demand, a second keeps time-critical AI inside its own RAN software, and independents bypass the RIC altogether, one of them delivering the same E2 code straight to the base station. A separate xApp server, and its power bill, broke operators' cost models, and the VMware and Juniper exits stranded the integrations app vendors had built. The use cases moved up to rApps for anything slower than a second, and down into dApps or the RAN software itself for real-time control.

Two surprises. The first commercial dApps are in space: beam weights and satellite handover are not native in the standards, so one vendor fills the gap with its own AI models running in the DU, in production on a GEO network with LEO targeted for 2027. And the US government is no longer just a buyer: the Pentagon's FutureG office funded OCUDU, an open source CU / DU now backed by a Linux Foundation ecosystem with AT&T, Verizon, Ericsson, Nokia and NVIDIA. Defence wants a programmable scheduler, and RAN vendors guard it as their crown jewel. In my mind, the three RIC tiers are collapsing into one AI plane, training central and inference placed wherever the loop latency requires, which I think starts at central offices rather than cell sites, whatever the GPU vendors and their partners say.

So where does this go by 2028? In a vendor's words, not simply more RIC. Tier-one vendors now describe SMO and RIC as the base for intent-driven, agentic operations, autonomy sits at level 2.5 to 3 today, and a reality check on energy: measured savings average 15 to 16%, while headlines quote 35 to 40% peaks. The test for 2027 is the first production rApp combining a third party app, another vendor's SMO and a different RAN. I left the panel with three questions: who pays for porting, who arbitrates conflicts between apps and who is liable, and will certification replace portability? It will be interesting to see whether Open RAN's tenth year is the one where the original promise is at last delivered, or quietly replaced.

Monday, November 2, 2026

VMO2 AI-RAN Trial Used Custom Silicon, Not GPUs


Light Reading reported on 2 November that Virgin Media O2's AI-RAN trial at Paddington, central London, did not use an Nvidia GPU. Nokia told the publication that the site ran its AI-RAN software on existing Marvell custom silicon, a Layer 1 chip Nokia calls ABIP, and Light Reading withdrew its original story. My post VMO2 Finds AI-RAN Does Not Need GPUs said the trial ran on a GPU, following that original report, and that detail was wrong, it has been corrected since. The result VMO2 described is unchanged: AI-based channel quality estimation raised spectral efficiency by 20%, and he said a GPU is not required to do it.

Nokia relies exclusively on Marvell silicon for Layer 1 in its RAN products today, and ABIP is the last custom chip it plans to develop. Its successor, ABIG, is built on Nvidia technology. The 20% at Paddington equals the figure Nokia gave in the summer for AI-RAN on GPUs, when it projected 50% in 2027 and a doubling by 2028. A white paper from Chetan Sharma Consulting that Nokia endorsed was more cautious, noting that a site already close to its theoretical spectral efficiency has less headroom for AI gains. Neither the Paddington result nor the GPU figures come with a baseline.

This is the position I took in Ericsson and Nokia at odds on AI-RAN silicon and in Is AI-RAN the future of telco?: using AI in the RAN and buying GPUs are separate decisions. The evidence now covers both vendors. Ericsson reports about 10% from an AI successor to link adaptation on about 60,000 deployed base stations, and Nokia's trial reached 20% on installed silicon. Nokia's argument for GPUs is that they receive years of software upgrades while custom silicon is replaced more often. That is the case to be tested, and it depends on how much AI headroom each generation of ASIC has, which neither vendor has published. Operators weighing ABIG against ABIP-class hardware need that figure, together with a baseline and a cost per site, before they commit to the upgrade.