(Source: Kirill Ivanov / Alamy Stock Photo)

Fear that a geopolitical clash between China and the US might rip apart the 3GPP, the international body presiding over mobile standardization, arguably reached its zenith several years ago. At a plenary in Madrid last month, delegates from Huawei, China's biggest vendor, trampled the same halls as those from Ericsson and Nokia, its European rivals. The 3GPP has survived the first rounds of 6G standardization. But a fight that mirrors the geopolitical contest has broken out over the shape of the technology. It threatens to produce a Chinese 6G that competes against one backed by Europe and the US.

The fight is over what the industry refers to as the "core." If the radio access network (RAN) is the surface skin of the infrastructure, the core is its brain or central nervous system. Software-based – and traditionally hosted at a few data centers as opposed to thousands of mobile sites – it generates a smaller amount of revenue for the large vendors than the RAN. But it is also responsible for some of the critical features and big network decisions.

At stake is how much the core should change with the transition from 5G to 6G. Responding to the wishes of some of the big telcos they serve, Ericsson and Nokia, along with various other European and US stakeholders, are pushing for a more evolutionary approach, a tinkering with the fundamentals that would park AI in a separate domain. Other non-Chinese players in an adjacent camp also want AI's role limited to use in certain 6G network functions.

"The 6G core is a very lightly tweaked version of the 5G core," said Luke Ibbetson, the head of research and development for Vodafone, expressing the views of people in these camps. "It's not a new thing." Ahead of the plenary, 17 operators outside China, including all the European and US incumbents, submitted a joint position paper calling for "evolution as the starting point" in 6G system architecture. The 5G core should be "an important architectural foundation for 6G," said the companies.

Evolution is the same word that Ericsson has used since last year to frame its thinking about 6G. "What we're pushing for is a move to that standalone 5G core," said Marie Hogan, Ericsson's head of 6G portfolio strategy, back in June. "The 6G core will reuse a large part of that." As her comments reflect, many telcos have not even upgraded from the old 4G core, supporting what is called the "non-standalone" version of 5G.

Revolution is in the air

But Huawei and ZTE, China's two main 5G vendors, together with China Mobile, the country's biggest operator, are calling for something more radical – a revolution instead of an evolution. Rather than keeping AI in a separate domain, or inserting it into 6G network functions, they want to see AI agents deeply integrated into procedures and even what Ruth Brown, a senior principal analyst at Light Reading sister company Omdia, called "agent-handled connectivity" in her recent research.

Just weeks ahead of the next 3GPP meeting, there appears to be very little appetite for compromise, according to one European telco executive who was in Madrid and who asked not to be named. "I am surprised that people don't want to compromise," he said. "Everyone wants simplicity as long as they get their own solution." Worsening relations between China and the West seem to be manifesting themselves within global standardization activities.

This would not, of course, be the first time companies have disagreed about the make-up of a forthcoming generation. During 5G standardization, Huawei was charged by some observers with forging alliances to champion its proposals over those of competitors. But Brown warns that 6G's bust-up could result in possible fragmentation.

"This could leave operators managing two divergent cores: a revolutionary single-vendor solution in Chinese domestic markets and an evolutionary multi-vendor architecture in the rest of the world," she said in a blog published on Light Reading. "Post-Madrid discussions must finalize an architecture by November and deliver a system solid enough to keep a divided industry unified under a single global standard."

Recent market developments have made this 6G duality look more possible. Seven years ago, Börje Ekholm, then Ericsson's CEO, put his company's share of the Chinese market at 10%. By late 2025, the combined share of Ericsson and Nokia had dwindled to around 3%, according to Nokia executives. The Nordic vendors have been replaced by Chinese alternatives, including rivals to Huawei and ZTE, said Gabriel Brown (no relation to Ruth), another Omdia senior principal analyst.

Big European operators, in turn, claim to have been ditching Huawei's core network products, even if they have been less proactive in removing its RAN equipment. None of the main American telcos are customers of Huawei or ZTE.

Fudging it

Outside China, much of the opposition to a 6G overhaul stems from the disappointment telcos feel about 5G. "They're not saying let's not be innovative, but I think they're saying let's not be too radical either, because they've spent a lot on 5G," said Ruth Brown. Despite the investment, operators have struggled to monetize 5G standalone features like network slicing, which reserves portions of capacity for specific user groups. "It was a big wow feature and that's still in progress," she said.

Even so, a criticism of the conservative 6G camps is that timidity could eventually prove costlier if an AI-native core quickly becomes essential. Gabriel Brown draws an analogy with car makers that tried evolving combustion engines into electric vehicles (EVs) rather than starting from a clean slate. "The chassis design really wasn't ideal, but it kind of worked," he said. "What they've now found is that EV-native is the better way to do it. Even Volkswagen and Toyota are now doing EV-native designs, but they went through a five-year period of trying to fudge it."

The revolutionary approach could also produce a more elegant design that scraps unnecessary tech and is ultimately more efficient. Traditional signaling would have multiple hops or stopping-off points between the user equipment and the core, involving use of a tool called the access and mobility management function (AMF). The China camp is proposing a new signal routing function – independent of the AMF – that could be more streamlined. "I guess their rationale is trying to remove some of this kind of procedural, multi-hop bloat," said Ruth Brown.

The ideas of this more revolutionary camp are referred to in the standards arena as Direction 3. As implied by the numbering, other stakeholders are not in neat alignment. Direction 1, whose supporters include Ericsson and Nokia, wants AI in a separate domain. Direction 2 – backed by NTT Docomo, Vodafone and Asian vendors including LG Electronics, NEC and Samsung – supports the injection of AI into dedicated network functions.

One possible compromise scenario, as described in Ruth Brown's blog, is a "partial merge" of the first two camps. AI would remain a separate domain, but there would be some tweaks to the routing functions to satisfy Direction 2 demands. That would still, however, leave a big gap with Direction 3, and the contrasting visions do not look easily reconcilable.

A desire to conceive a 6G core at the outset reflects the industry's determination to avoid a repeat of what is now regarded as something of a 5G fiasco. It has saddled the world with non-standalone and standalone variants for years.

The approach of 6G might finally rectify it, prompting a shift to standalone by the non-standalone laggards. "Most operators are viewing 5G standalone as the baseline from which they will deploy 6G," said Ibbetson. But compared with the split that happened between the two 5G variants, an irreconcilable bust-up over the 6G core standard could have far more serious implications.