Japan, South Korea and China are pursuing three distinct paths toward sixth-generation mobile networks, and the gap between their strategies has widened rather than narrowed as 2026 has progressed. NTT is pushing an optical-first architecture through its IOWN initiative, Korean carriers are running pre-standardization field trials tied to the country's semiconductor and equipment industry, and Chinese operators are folding low-earth-orbit satellite links into the terrestrial network rather than treating space and ground infrastructure as separate systems. None of the three has committed to a single technical blueprint, and that divergence is starting to worry equipment vendors who would rather sell one radio design to the whole region than three.
Japan bets on light instead of radio
NTT's Innovative Optical and Wireless Network programme, known inside the industry simply as IOWN, is the most structurally different of the three approaches. Instead of treating optical fibre as a backhaul afterthought behind the radio network, IOWN pushes photonics deeper into the network core, aiming to replace electronic switching with optical processing at exchange points and eventually inside base stations themselves. NTT has framed the payoff in terms of power draw and latency rather than raw throughput, arguing that an optical core can cut end-to-end delay to a fraction of what today's fibre-electronic hybrid networks manage.
The commercial case is not abstract. NTT has been extending IOWN-linked optical transport between selected exchange buildings in the Kanto region through 2026, and it has said repeatedly that a nationwide optical-core rollout is intended to arrive well before the international 6G standard is finalized, rather than waiting for 3GPP to hand down a single global specification. That sequencing matters: if IOWN's optical layer becomes the default transport for Japanese carriers before 6G radios exist, equipment makers selling into Japan will need to support a network core that looks meaningfully different from what ships in Seoul or Shenzhen.
NEC, Fujitsu and the domestic supply chain
NTT is not building IOWN alone. NEC and Fujitsu have both committed engineering resources to optical transceiver and switching hardware tied to the initiative, and the Ministry of Internal Affairs and Communications has treated IOWN as a strategic asset worth defending on export terms, not just a domestic upgrade. Sony and NTT Data have separately explored optical interconnects for data centre applications that overlap with IOWN's core technology, which gives Japan a broader industrial base for the approach than a single-carrier project would normally command.
Korea trials quietly, standardizes loudly
South Korea's approach looks almost the opposite of Japan's in tone, even though the underlying goal — being first to a workable 6G network — is identical. SK Telecom and KT have both run pre-standardization 6G trials through the first half of 2026 involving terahertz-adjacent spectrum bands and AI-native radio access network software, but neither carrier has publicized detailed throughput or latency figures the way they did during the 5G rollout years. The trials appear designed to feed Korean proposals into 3GPP standardization meetings rather than to generate consumer-facing marketing claims, a shift that reflects how burned the industry got by 5G's gap between marketed speeds and delivered performance.
Samsung's network equipment division and the state-backed Electronics and Telecommunications Research Institute have both been named as technical contributors to Korea's 6G proposals, and the government's Ministry of Science and ICT has tied its 6G research budget to the broader physical-AI and robotics push it announced earlier this year. That linkage is deliberate — Seoul wants 6G framed as infrastructure for autonomous factories and robotics rather than as a faster phone network, since the domestic market for another consumer speed upgrade is already saturated. Whether that framing survives contact with actual 3GPP Release 21 negotiations, where terahertz spectrum allocation and network-AI integration are still contested even among close allies, is a separate question the trials alone cannot answer.
China integrates satellites into the core network
China's telecom operators and its satellite programmes have converged in a way neither Japan nor Korea has attempted at comparable scale. State-linked satellite constellations, including the Guowang and Qianfan (Thousand Sails) low-earth-orbit projects, are being designed from the outset to interoperate with terrestrial 5G-Advanced and future 6G networks rather than to operate as a standalone backup service for remote areas. China Mobile, China Telecom and China Unicom have all discussed direct-to-device satellite connectivity trials in 2026, aiming at a network architecture where a handset hands off between a terrestrial tower and a satellite link without the user noticing the switch.
This satellite-terrestrial integration strategy solves a problem that is more acute for China than for Japan or Korea: covering a country with vast rural and maritime territory without building terrestrial towers everywhere. But it also gives Chinese standards bodies a strong incentive to push satellite integration into the core 6G specification at 3GPP, rather than leaving it as an optional regional feature — a point of friction that has already surfaced in preliminary standardization discussions this year. Huawei and ZTE, both still under partial restrictions in Western markets, have used satellite integration as a technical area where they can claim leadership without needing access to markets that have excluded their radio equipment on security grounds.
Why the divergence matters for equipment makers
A genuinely unified global 6G standard would let Ericsson, Nokia, Samsung, Huawei and ZTE build largely interchangeable radio and core equipment, the way 4G and, imperfectly, 5G allowed. Three national roadmaps pulling toward optical-core architecture, AI-native radio access networks and satellite-terrestrial integration respectively make that outcome considerably less likely. Equipment vendors now face a familiar but unwelcome choice: build modular systems flexible enough to support all three approaches, at higher engineering cost, or pick a lane and risk losing sales in the markets that chose differently.
- NEC and Fujitsu are best positioned for Japan's optical-core path, since photonics and transceiver design is already their core business rather than an adjacent bet.
- Samsung and Nokia have the software-defined radio depth that Korea's AI-native RAN trials reward.
- Huawei and ZTE hold the edge on China's satellite-terrestrial integration, an area where Western vendors have limited standing regardless of technical merit — and where the field is unlikely to stop at these three players as more national programmes weigh similar bets.
The semiconductor side of this fragmentation is arguably sharper than the equipment side. Baseband and RF front-end chips tuned for satellite handoff protocols are not the same silicon as chips optimized for optical-core signalling, and fabless designers serving the region will need to decide fairly soon which combination of standards their next generation of chips actually needs to support, since design cycles for advanced RF silicon run two to three years ahead of commercial deployment.
The standardization fight still ahead
3GPP's own 6G timeline points to a first full specification release around the end of this decade, with commercial networks following in the early 2030s — a schedule industry participants have described as aggressive but not yet officially revised. Japan, Korea and China are all active members of the standardization process rather than outsiders trying to route around it, which means none of the three divergent paths described above is guaranteed to survive into the final global specification unchanged. What is more likely, based on how 5G standardization actually played out, is a core specification broad enough to accommodate regional variation, with optical transport, AI-native RAN and satellite integration all surviving as optional profiles rather than universal requirements.
That outcome would suit incumbent equipment vendors more than it suits any single government's industrial strategy, since it lets Ericsson and Nokia keep selling into all three markets without picking a side. It would be a less satisfying result for Tokyo, Seoul and Beijing, each of which would prefer the eventual global standard to look like the version they already built.