Summary

  • NTT reports average capacity of 1.94× SISO in 2×2 MIMO and 2.67× in 3×3, plus successful separation of two simultaneous 920 MHz IoT signals.
  • The tests establish orbital measurements, not a commercial offer: price, paying customers, coverage and deployed sensor count remain unpublished.

NTT’s satellite experiment has acquired two useful numbers and one conspicuous blank. A 2×2 MIMO test averaged 1.94 times the capacity of a single-antenna link; a 3×3 test averaged 2.67 times. In a separate trial, a satellite separated two 920 MHz IoT signals sent at the same time and frequency from different places on Earth. The blank is the one a buyer would ask for next: the price of using it.

The two results belong to different links. MIMO tested the feeder connection from satellite to ground stations. The satellite sent multiple signals on the same frequency; separated ground antennas received them, and processing at the ground station disentangled them. That is a capacity result for the backhaul path. It is not the data rate promised to a sensor or a service customer.

The IoT test addressed the other end of the story. Two low-power terminals, located at different sites, transmitted together on the same 920 MHz channel. The satellite recovered both signals. Two is a real engineering milestone: the receiver had to tell overlapping transmissions apart. Two is also still two. It does not tell an asset operator how many meters, pumps or remote instruments a service could support, how often each would report, or what a terminal would cost.

NTT and the Japan Aerospace Exploration Agency mounted LEOMI on RAISE-4, a small demonstration satellite launched on 14 December 2025. Steady-state experiments began on 13 March 2026. NTT says the tests will continue for about a year, with different antenna configurations and satellite passes, plus varied IoT transmissions and radio-environment checks. The October release reports roughly five months of MIMO trials, not a completed service rollout.

That distinction is where the market sits. The tests show that more traffic can fit through a satellite feeder link and that two terrestrial sensors can share a narrowband satellite path. A product still needs an operating schedule, field terminals, data handling, service commitments and a bill someone can approve. NTT’s release names no paying customer, price, coverage promise or deployed sensor fleet.

There is a practical planning result too. NTT says measured capacity changes tracked estimates from an NTT–JAXA channel model. Predictable capacity swings could help engineers choose satellite antennas and place ground receivers around an orbit. That is valuable design evidence: it can reduce guesswork before a system is built. It does not yet tell a customer what reliability or monthly cost to expect.

NTT has moved the story from concept to orbital measurements. The next year of tests must show whether the capacity gains repeat across passes and configurations, and whether the two-terminal trick grows into a useful field service. Until an offer and a customer appear, the clever part is in orbit; the price tag is still firmly on Earth.

Analysis

The next review should break the results into three ledgers: repeated capacity across satellite passes and antenna layouts; the number and type of IoT terminals that can share a channel; and the full ground-side cost of service. NTT plans more configurations and radio-environment tests through the experiment year. A repeatable service-level result, named terminal class and customer trial would change the commercial reading. A display at NTT R&D Forum from 18 to 25 November is a near-term evidence check, not a substitute for operating data.

Sources