Summary
- Rakuten Symphony was selected for a Japanese METI grant supporting the next phase of its Open RAN work with Kyivstar.
- The first phase completed 4G laboratory validation of features and interoperability.
- The second phase will put 4G LTE into field tests around Kyiv, carrying commercial customers in a live brownfield network.
- The project will evaluate distributed and centralised RAN architectures; 5G will still be assessed in a dedicated laboratory.
- Expected project cost is about ¥710 million, while the potential METI subsidy is capped at ¥350 million.
A network trial changes character when ordinary customer traffic enters it. Laboratory interoperability asks whether components can work together under controlled conditions. A brownfield field test asks whether they can coexist with legacy equipment, operational procedures and the service expectations of people already paying for connectivity.
Rakuten Symphony and Kyivstar are crossing that boundary around Kyiv. Their first phase validated Open RAN and cloud functions in a 4G laboratory. The next phase is intended to carry commercial customers over 4G LTE in a live network while comparing distributed and centralised radio architectures.
That is a material step. It is not a commercial 5G launch.
Brownfield is where integration costs become visible
Open RAN separates functions that a traditional supplier may deliver as a more tightly integrated system. The promise is greater choice and a more software-led operating model. The operational burden is that interfaces, timing, performance and fault ownership must work across components.
A brownfield network adds another layer. Kyivstar is not building an isolated demonstration with no installed base. The project is described as part of replacing legacy infrastructure. New radio and cloud functions therefore have to enter an operating environment without treating existing customers as test equipment.
The comparison between D-RAN and C-RAN matters for that reason. Distributed architecture keeps more processing near radio sites. Centralised architecture can pool functions but asks more of transport, timing and resilience between sites and the processing layer. The announcement does not declare a winning design. It says both will be evaluated.
Useful evidence will include stable handovers, busy-hour performance, fault recovery, power behaviour and the amount of operational work required to isolate a problem. The current release does not publish those measures, a site count or a field-test completion date.
Commercial users do not make the trial a commercial product
“Carrying commercial customers” means the test will encounter real traffic and real service consequences. It does not say that Kyivstar is selling an Open RAN tariff or that the architecture is deployed across its network.
The distinction protects two different claims. The customer can be commercial while the infrastructure beneath the session remains under validation. Likewise, a successful group of sites would prove performance in that tested configuration, not automatic suitability for every region, spectrum band or traffic pattern in Ukraine.
War makes the test more consequential and less easy to generalise. A network serving Ukraine has to contend with damaged infrastructure, unstable power and unusual continuity demands. Those conditions can reveal resilience weaknesses quickly. They can also make costs and operating practices different from a peacetime nationwide rollout.
The 5G evidence remains behind the laboratory door
The programme expands to “4G/5G configurations”, but the location of each generation is explicit. The 4G LTE work goes into the field around Kyiv. The 5G capabilities remain in a dedicated laboratory as a foundation for a possible future commercial deployment.
It would therefore be wrong to describe this phase as a Kyiv 5G Open RAN launch. It also should not be merged with Ukraine’s separate, government-coordinated multi-operator 5G city pilot. The two programmes test different systems and have different participants and evidence.
The next meaningful 5G milestone would be a defined field configuration, authorised spectrum conditions, named sites and operating results. None is supplied here.
The subsidy buys evidence, not scale
Rakuten Symphony expects the project to cost about ¥710 million. METI may cover up to ¥350 million. The ceiling is slightly less than half of the expected cost, and “up to” does not establish that the maximum has already been paid.
The announcement does not allocate the remaining funding among Rakuten Symphony, Rakuten Mobile, Kyivstar or other sources. It also does not translate the budget into radios, sites or nationwide coverage. Cost per site cannot be inferred.
What the grant can purchase is a more credible body of evidence: whether open interfaces survive live traffic, whether centralisation is practical in this transport environment, and how much operating effort a legacy replacement requires.
If that evidence is favourable, a later commercial decision would still need scope, procurement, financing and a migration calendar. For now, Kyivstar’s Open RAN programme has moved from a controlled room into a working network. That is a stronger test—and still a test.
