La plateforme stratosphérique de Sceye revient aux États-Unis après un vol stratosphérique aller-retour sans précédent entre les États-Unis et le Japon

Sceye et SoftBank ont achevé avec succès la mission stratosphérique ST1, qui a parcouru près de 30 000 km entre le Nouveau-Mexique et le Japon sur un mois. Le HAPS a opéré plus de sept jours dans l’espace aérien japonais à 16,5-17 km d’altitude, assurant SMS, appels, internet et streaming, avec une précision de maintien de zone de 5 km. Le traitement embarqué a atteint un aller-retour moyen de 68 ms, réduisant la latence de plus de 40% par rapport à un traitement cloud, renforçant la perspective de déploiement commercial d’infrastructures de connectivité HAPS.

Analysis

SoftBank (9434 JP) gains a low-capex option on rural coverage and disaster-resilience services, but the near-term earnings relevance is negligible: the commercial question is cost per covered subscriber-hour, not whether a demonstration can carry traffic. A HAPS layer could reduce tower-build requirements in low-density Japanese regions and improve network differentiation versus KDDI (9433 JP), NTT (9432 JP) and Rakuten (4755 JP); however, it also introduces aviation permissions, spectrum coordination, fleet replacement and weather-related availability risks that terrestrial-network economics do not bear.

The more material second-order read is competitive pressure on satellite direct-to-device narratives. AST SpaceMobile (ASTS), Globalstar (GSAT) and Iridium (IRDM) trade partly on the scarcity value of non-terrestrial coverage; a credible stratospheric alternative offers materially lower latency and potentially cheaper capacity in markets where regulators permit persistent operations. That said, HAPS is not a substitute for global satellite coverage over oceans, polar routes or politically inaccessible airspace, so the threat is concentrated in disaster response, fixed wireless access and rural/mobile coverage rather than core global connectivity.

Over the next 1-3 months, treat this as a technology-validation catalyst rather than a revenue catalyst. The thesis becomes investable only if SoftBank discloses paid trials, usable capacity per platform, operating availability across seasons and a deployment cost below avoided terrestrial capex; absent these metrics, optimistic 6G framing should not drive a sustained rerating. A regulatory delay in Japan, inability to maintain station-keeping through adverse conditions, or service economics above satellite/terrestrial alternatives would falsify the HAPS disruption case.

Key Decisions for Investors

• Maintain a watch-list long in SoftBank (9434 JP), not a new core position, ahead of its next results and any paid-service disclosure; initiate only if management quantifies HAPS capex/opex and identifies a commercial launch window. Upside is incremental network-quality differentiation, while downside is limited near term because the project is unlikely to move consolidated EBITDA.
• Avoid chasing ASTS or GSAT on this development alone; instead, set an alert for disclosed Japanese HAPS commercial contracts or spectrum approvals. If either occurs, consider a 3-6 month relative-value hedge: short a basket of ASTS/GSAT against long 9434 JP, sized small, because HAPS primarily pressures regional direct-to-device valuation premiums.
• Prefer IRDM over ASTS/GSAT within non-terrestrial connectivity until HAPS unit economics are independently demonstrated: IRDM's global, safety-critical and maritime exposure is less substitutable. Reassess if HAPS platforms show multi-season availability and international airspace approvals.
• For Japanese telecom exposure, monitor KDDI (9433 JP) and NTT (9432 JP) for defensive responses through satellite partnerships or rural-network capex changes. A material reduction in SoftBank rural capex guidance without offsetting HAPS operating costs would be the first financially actionable confirmation.