Our Take
HAPS is real hardware with a 12-day flight proof, but claiming parity with satellites on cost and complexity needs independent data—Sceye's numbers are company-reported only.
Why it matters
Stratospheric platforms could serve disaster zones and dense urban areas faster than satellites and cheaper than ground infrastructure. The first production test with Softbank signals whether this category moves from engineering demos to commercial viability.
Do this week
Infrastructure leads: monitor Sceye's August Japan results for latency and uptime metrics before planning any HAPS integration into backup connectivity stacks.
Sceye launches stratospheric platform trial with Softbank in Japan
Sceye, a New Mexico-based company, will dispatch a helium-filled airship roughly 200 feet long to hover 18 kilometers above Japan's coast in August 2026. The craft will carry a custom antenna designed to supplement Softbank's 5G network with direct-to-device data transmission. This test follows a spring 2025 proof-of-concept flight in which Sceye's platform stayed aloft for 12 days, including more than 88 hours of stationary positioning off Brazil's coast.
Sceye's platform belongs to a category called HAPS (high-altitude platform stations or systems). Competitors include Airbus's Aalto subsidiary. These systems operate in the stratosphere, a layer thin enough to require lightweight construction but high enough to cover large ground areas without the cost and orbital complexity of satellites.
The airship must solve a core engineering problem: staying aloft while generating enough solar energy during daylight to power an electric fan that corrects its position when wind pushes it off station. The craft proved it could do this in a 2024 test flight (company-reported).
Latency and cost, not altitude, are the real advantage
Stratospheric platforms sit roughly 35 times closer to ground than geostationary satellites. That distance reduction cuts signal transmission energy and latency, making HAPS more efficient for dense coverage zones where a single platform can serve an area a satellite would require multiple passes to cover.
The business case is clearer for disaster response and remote regions than for general broadband. A single airship can supplement existing 5G during network outages, and the electrical overhead to maintain station-keeping is lower than the fuel cost of keeping manned aircraft aloft. But no independent benchmarks yet exist comparing HAPS power consumption, uptime, or coverage cost per gigabyte against satellite or terrestrial alternatives.
Sceye's CEO Mikkel Vestergaard Frandsen claims the platform offers "space-like conditions, without the cost of going to space and without the complexity of being in orbit." That framing is company-reported. Whether it holds up under Softbank's commercial test remains the open question.
What to track in the Japan test results
Three metrics will determine whether HAPS moves from prototype to infrastructure candidate. First: sustained uptime. The 12-day Brazil flight was a test; a production deployment must sustain months-long station-keeping with automatic repositioning during seasonal wind shifts. Second: latency to device. Direct-to-device 5G transmission avoids backhaul delays, but only if the antenna can track users moving across the platform's footprint. Third: failure modes under real weather. A stratospheric platform in a typhoon season faces stresses the Arizona and Brazil test flights did not encounter.
Softbank's involvement signals they believe the trial is worth the operational risk. But "supplementing" an existing network is different from replacing terrestrial infrastructure. Expect the August announcement to focus on connection counts and coverage maps rather than cost-per-bit or customer churn impact. Those comparisons are what determine actual deployment velocity beyond Japan.