NTT Docomo has signed a memorandum of understanding with US equipment maker Abside Networks to develop AI-driven networks built on high-altitude platform stations (HAPS) — the latest step in the operator's push to carry traffic above terrestrial masts.

The agreement covers four work areas: HAPS system architecture, network control, data utilisation, and the application of AI to forecast demand and manage radio beams, capacity, spectrum and interference in a stratospheric RAN.

The partners also plan to study the collection and use of data on network quality, operations and stratospheric sensing, along with potential integration with external applications and service platforms.

What is Docomo actually buying?

HAPS platforms operate at roughly 20km altitude — about 60 times the height of a typical urban macro tower — and use unmanned aircraft or balloons to deliver wide-area connectivity. Docomo frames them as a route to places where conventional mobile coverage is difficult or uneconomical: mountainous regions and maritime areas.

The MoU supports what Docomo calls a "three-dimensional communications infrastructure that integrates AI with non-terrestrial network (NTN) initiatives, including HAPS". The operator says it wants to combine terrestrial networks with systems in the stratosphere and space to deliver wider-area, higher-capacity services as AI usage drives up demand for connectivity.

That demand argument deserves scrutiny. Docomo has not disclosed projected traffic volumes, target coverage areas, or any commercial deployment date for HAPS services. The MoU is a research agreement, not a procurement contract, and neither party has quantified the investment involved.

Why Abside?

Abside brings two relevant assets. First, it is developing radio systems based on large antenna arrays for precise beam control, paired with an AI-driven software platform designed to provide "integrated control of networks across the ground, stratosphere and space", according to the company.

Second, it has direct stratospheric experience: Abside worked on Google's Project Loon balloon initiative, which Alphabet scaled up before ultimately winding it down. That history cuts both ways — it demonstrates real expertise in stratospheric communications, but Loon's commercial closure remains the industry's clearest warning that HAPS business models are hard to close.

Neither Docomo nor Abside addressed that precedent in announcing the deal.

What remains unsaid

Docomo acknowledges the technical obstacle plainly: the stratospheric environment requires specialised RAN technologies and equipment, which is why it needs a partner rather than adapting its existing stack. The operator has been conducting research and demonstrations aimed at delivering HAPS-based communications services, but no trial results, throughput figures or coverage targets accompany this announcement.

The missing specifics matter:

• No deployment timeline or target market
• No financial terms attached to the MoU
• No spectrum plan for stratospheric operation
• No stated performance benchmarks for the AI-driven beam control

A crowded stratosphere

Docomo is not starting from zero. The operator previously partnered on a $100 million investment in a HAPS supplier, according to Mobile World Live reporting, and the broader industry push into NTN — spanning satellites as well as stratospheric platforms — is reshaping how operators plan coverage in hard-to-reach geographies.

For now, the deal signals intent rather than capability. Docomo and Abside will next have to show that AI-managed beams from 20km up can move from architecture documents to measurable capacity — and, more importantly, to a service someone will pay for.