Not the Airship, the Notice Chain: What Stratospheric Cell Service Can and Cannot Do for MBG Meal Continuity

MBG Watch · 2026-10-03

The premise

A stratospheric cell tower is an interesting machine. It is not, by itself, a food-safety control.

The useful question for MBG is smaller and more operational: when floods, earthquakes, haze closures, island-route interruptions, portal outages, power failures, or telecom disruption break the usual channels, what record shows that the meal-status decision reached the humans who had to act?

That means kitchens, drivers, schools, posyandu and local cadres, families, health posts, district governments, and the people responsible for pausing, rerouting, replacing, or recalling food. The test is not whether a backup signal exists somewhere above the road. The test is whether the notice chain still worked on the disrupted day.

This piece builds on MBG Watch’s earlier continuity work: “When the Last Mile Goes Quiet: The Communications Readiness Record MBG Kitchens Need,” “If the Portal Is Attacked: The Service-Continuity Record MBG’s Digital Layer Needs,” “When Public Climate Data Becomes Meal Infrastructure: The Data-Continuity Standard MBG Needs,” “Before the Forecast Works: The Rehearsal Record MBG Kitchens Need,” and the island-route work in “When the Sea Route Goes Missing” and the Flores continuity pieces. It also crosses with Rupiah Stability Watch’s 3 October 2026 piece, “Stratospheric Cell Service and the Rupiah: Ordinary Phones, Disaster Continuity, and the Archipelago Operating Ledger,” which treats the same Sceye/SoftBank signal as an operating-ledger issue for public continuity in an archipelago.

What the Sceye and SoftBank evidence actually shows

The core signal is real, but it is still a trial.

Sceye said its Service Test 1 mission flew more than 15,000 km from New Mexico to Japan through the stratosphere, then demonstrated mobile broadband connectivity to unmodified devices through SoftBank’s core network. The company said the platform carried SceyeCELL, a “cell tower in the sky,” and demonstrated text messaging, voice calls, internet access, video streaming, emergency calls, an emergency-alert messaging system, drone communications, and edge computing from the stratosphere. It also reported operation at about 16.5 km altitude, more than seven days in Japanese managed airspace, and station-seeking within a radius as low as 5 km.

SoftBank’s own release is more useful because it states the boundary. It says the companies conducted communications trials with “limited durations, areas and devices” while working toward HAPS commercialization from 2027 onward. The HAPS flew more than 15,000 km, operated off Cape Muroto in Kochi Prefecture, used a 4G LTE-equivalent base station connected to SoftBank’s core network through a ground gateway, and confirmed voice calls, emergency calls, text messaging, apps, video calls, video streaming, earthquake and tsunami warning tests, and a maritime emergency call test.

New Atlas’ later report adds two important cautions. It describes the platform as beaming mobile service to ordinary phones from about 10 miles up, but notes there is no pricing or coverage plan yet. It also quotes Sceye’s CTO cautioning that the “500 terrestrial towers” coverage comparison is not a measure of simultaneous capacity; performance depends on radio frequencies, the backhaul link to the core network, and the number of users online.

So the evidence supports this: a HAPS layer can plausibly preserve or extend an ordinary-phone communications path in selected conditions, especially for warnings and operational messages. It does not support a procurement conclusion for Indonesia, and it does not prove a general public-service continuity guarantee.

Why this matters for MBG

MBG is not only a kitchen program. It is a route, time, temperature, eligibility, incident-response, and public-notice program.

BGN has said its food-safety incident response system includes early detection, tiered reporting, emergency handling at SPPG level, temporary production stoppage, tracing the source of a problem, evaluation, and procedural repair before operations resume. BGN has also described 2027 strengthening around SPPG kitchen standardization, food-handler training, supply-chain strengthening, quality monitoring, food safety, information systems, supervision, and governance for a program aimed at 72,464,886 beneficiaries, including students, pregnant women, breastfeeding mothers, and toddlers.

Those functions depend on communications at the moment the system is least tidy. A kitchen may need to stop production. A driver may need to turn back. A school may need to hold distribution. A cadre may need to tell families that a delivery is delayed or recalled. A health post may need to connect symptoms to a meal route. A local government may need to distinguish a true warning from a rumor.

Indonesia already treats communications restoration as disaster infrastructure. In November 2025, Komdigi said it was deploying ten SATRIA-1 emergency internet points in flood-affected areas of Aceh, North Sumatra, and West Sumatra, with BAKTI Komdigi, BNPB, SAR teams, and the military mobilizing equipment to restore connectivity where communications had been cut. The point for MBG is not that SATRIA-1 and HAPS are the same technology. They are not. The point is that meal continuity in a disrupted archipelago cannot assume that the normal network, portal, road, or public-notice route is available.

What the evidence does not support

There are four overclaims to avoid.

First, HAPS is not a food-safety system. It may move a warning, but it does not show whether rice was held too long, whether protein was contaminated, whether a kitchen paused in time, or whether a route remained reachable.

Second, direct-to-phone connectivity is not beneficiary validation. A phone signal does not prove the right child, mother, toddler, or school received the right meal under the right condition.

Third, backup connectivity is not payment correctness. It does not prove that suppliers, kitchens, couriers, or service points were paid correctly, nor that emergency substitutions were lawful and recorded.

Fourth, the public record does not show that MBG is adopting stratospheric connectivity. I found no public BGN source saying MBG has a HAPS plan or a stratospheric direct-to-phone continuity layer. The Sceye/SoftBank signal should therefore be treated as a systems prompt, not a procurement recommendation.

The minimum notice-chain continuity record

If MBG uses, pilots, promises, or relies on any backup communications layer — satellite, HAPS, radio, courier relay, offline-first portal sync, SMS broadcast, local government hotline, or a mixed system — the record should show the same minimum facts.

1. Decision ID. A unique identifier for the stop, delay, reroute, substitution, recall, reopening, or public correction.

2. Disruption source. The signal that triggered action: flood warning, earthquake damage, haze closure, road failure, ferry cancellation, telecom outage, power loss, portal compromise, food-safety report, lab result, school report, health-post signal, or local government instruction.

3. Affected operating unit. The SPPG, kitchen, route, school, posyandu, village, district, delivery batch, or beneficiary group affected.

4. Decision authority. The role that made the decision, not just the office name: SPPG lead, district coordinator, food-safety officer, school principal, health official, local disaster official, or BGN escalation point.

5. Message recipients by role. Kitchen operators, drivers, school contacts, cadres, health posts, local government, families, suppliers, and public-facing complaint channels.

6. Channel used. Portal notice, SMS, WhatsApp group, radio, satellite terminal, HAPS/direct-to-phone path, phone tree, printed notice, courier relay, local government broadcast, or school announcement.

7. Acknowledgement or fallback. Who confirmed receipt, who did not, what fallback channel was used, and when escalation began.

8. Timing. Time disruption was detected; time decision was made; time each recipient group was notified; time the meal was paused, replaced, rerouted, or cleared; time public notice was issued; and time the record was corrected if the first notice was wrong.

9. Public-facing version. A short, plain notice that families and schools can check without exposing private health data or operational security details.

10. Privacy boundary. What personal data was not shared, how beneficiary and illness data were minimized, and who could see the full incident record.

11. After-action repair. What failed in the chain, how the roster or channel list was corrected, and when the next rehearsal will test the repair.

The important feature is not the brand of the communication layer. It is the auditability of the human chain.

What MBG should be able to show before the next disrupted meal day

A useful public standard would be modest.

BGN does not need to publish every incident record in full. Some of it will include private health data, security-sensitive routing details, or personal phone numbers. But it should be able to publish the standard and keep the detailed record inspection-ready.

For a sample of disrupted days, auditors and local oversight bodies should be able to see whether:

• the kitchen knew whether to cook, pause, replace, or discard;
• the driver knew whether to depart, wait, reroute, or return;
• the school or posyandu knew whether to distribute, hold, or warn families;
• the health post knew whether symptoms might be linked to a meal route;
• local government knew which sites were affected;
• families received a public explanation in time to change behavior;
• missed acknowledgements triggered a fallback channel rather than disappearing into a chat group;
• the final record corrected rumors, mistakes, and earlier notices.

This is the same standard whether the message travels through a terrestrial tower, SATRIA-1 emergency internet, radio, a district phone tree, or a future stratospheric layer. Connectivity is a means. The public interest is the chain of notice and action.

What I am uncertain about

I found no public MBG-specific HAPS or direct-to-phone plan. That absence matters, but it is not proof that no internal exploration exists.

I also did not find a public MBG operating record that fully specifies degraded-network notice-chain fallback across kitchens, drivers, schools, cadres, health posts, families, and local governments. BGN’s published materials do describe incident response, reporting, emergency handling, food-safety controls, information systems, and governance, but not a complete public communications-fallback ledger.

The Sceye/SoftBank trial itself is promising only within its test boundary: limited duration, limited area, limited devices, a ground gateway, selected test scenarios, and commercialization still aimed from 2027 onward in Japan. Indonesia’s geography, regulation, spectrum arrangements, economics, and public-service needs would require their own evidence.

The least-harm conclusion is therefore plain: MBG should not be judged by whether it uses airships. It should be judged by whether, when the ordinary route goes quiet, the program can prove that the right warning reached the right person in time to keep an unsafe or impossible meal from becoming a public-service failure.