Summary

  • NTT DOCOMO and Abside Networks agreed to study HAPS system architecture, AI-assisted radio control and network-data use. Their 14 October announcement does not name a supplier award, aircraft, trial date, spending commitment or launch plan.
  • DOCOMO has an earlier HAPS relationship and a successful AALTO Zephyr link demonstration on record. Those steps establish context; neither answers how Abside’s proposed control work would enter an operating service.

High-altitude connectivity is easy to describe as a coverage map. The harder question is who controls the radio system when the map becomes a service. NTT DOCOMO and Abside Networks’ new memorandum of understanding brings that question into view, but it does not answer it.

The companies signed the MOU on 13 October and announced it the next day. They say they will jointly study future technical cooperation using high-altitude platform stations, or HAPS, and next-generation networks using artificial intelligence. The listed work is specific: system architecture for wide-area, high-capacity and many-user communications; AI for demand forecasts and control of beams, capacity, spectrum and interference in a stratospheric radio access network; collection and use of quality, operations and sensing data; and possible integration with outside applications and service platforms. The announcement describes a study. It does not identify a purchase, delivery duty, aircraft, flight, test calendar, customer or commercial service date. NTT DOCOMO and Abside Networks, 14 October 2026

The control list matters because HAPS is not simply a taller base station. The platforms operate around 20 kilometres above ground, where the radio footprint and interference conditions differ from terrestrial sites. A controller that moves beams or reallocates capacity would have to work with ground networks, other non-terrestrial links and the operator’s service rules. The MOU names some of the variables. It does not say whether Abside would supply equipment, software, both, or a research component; nor does it state who would approve automated changes or carry responsibility if a decision degraded service.

That gap should be read beside DOCOMO’s earlier HAPS work. In 2024, a Japanese consortium led by DOCOMO and Space Compass announced a commitment to invest up to US$100 million in AALTO through HAPS JAPAN, subject to closing conditions and regulatory approvals. The release targeted a Japanese launch and global entry into service in 2026. It was a dated target, not proof that service began or failed to begin. In January and February 2025, DOCOMO and partners used AALTO’s Zephyr over Kenya to relay a ground LTE signal to smartphones. The companies reported that the link was established, using a 2 GHz service link and a 38–39.5 GHz feeder link. That is a successful demonstration of a defined communication path; it is not a measurement of commercial coverage, cost, capacity at scale or customer demand. DOCOMO’s 2024 AALTO announcement · DOCOMO’s 2025 Kenya demonstration

The new MOU may address another layer of the system. AALTO’s Zephyr is an aircraft platform; RAN equipment and network orchestration decide how the communication service behaves. DOCOMO’s announcement says the parties’ visions and technologies may complement one another and mentions a future network spanning the ground, stratosphere and space. But it does not explain how the study relates to the AALTO/Space Compass work, whether one platform would be used, or whether the partners will test interoperability. A new research partner is not, by itself, evidence that the existing operating path has changed.

Abside’s public product page offers useful maturity context, with an important attribution limit. The company describes its HAPS equipment as drawing on Project Loon work, reports more than 700 units produced and fielded and more than 500,000 users served, and lists 4G LTE as proven. The same page marks 5G NR as in development and says range testing is pending. Those are Abside’s own product claims, not an independent audit and not evidence that the DOCOMO study will use the same hardware. They do make it harder to treat “AI-driven HAPS” as a finished commercial product. Abside HAPS product page

DOCOMO also has a Noto HAPS partner programme aimed at identifying use cases and resilience benefits; its public page gives a local flight horizon from fiscal 2028 onward. That programme has a different geography and purpose from a future nationwide service. It reinforces the need to track each milestone by what it tests: link feasibility, local use cases, integrated radio control, commercial acceptance or customer service. DOCOMO’s Noto programme

The commercial value of the memorandum therefore remains unpriced. It gives DOCOMO a route to examine a control layer and gives Abside a route to discuss its RAN and AI work with a national operator. The public record cannot show the cost of the study, whether either company has committed equipment or staff, or how a successful experiment would earn a return. DOCOMO’s subscriber base is distribution context, not a forecast of HAPS customers.

For now, the useful result is narrower: the study agenda recognizes that stratospheric coverage depends on more than an aircraft and a radio link. It also depends on how operators predict demand, allocate spectrum and capacity, and intervene when automation meets a service constraint. The next evidence should be a named test platform and date, measured acceptance criteria, an explicit division of RAN and controller responsibilities, and a path from a successful test to a paid service. Until then, the MOU is a control-system research option—not a rollout.

Sources