Indonesia's first beyond-visual-line-of-sight (BVLOS) drone food-delivery trial is now flying in Bali. It also shows how a regulator can let an industry test real operations before the rulebook is finished.
What is flying, and under what authority
DroneDJ reported on 7 October 2026 that Grab, flight operator SKYY Network and A2Z Drone Delivery are running deliveries between Café del Mar Bali and Nuanu Creative City, a 44-hectare development. A2Z supplies the Longtail Dual aircraft and automated AirDocks for landing, recovery and charging. The flights are "being conducted under a certificate issued by Indonesia's Directorate General of Civil Aviation (DGCA), following a Specific Operations Risk Assessment (SORA) process." DroneDJ also says the route sits in controlled airspace associated with Ngurah Rai International Airport, and calls the trial "reportedly the first" BVLOS food delivery operation in Indonesia (DroneDJ).
Grab Indonesia's Bali Nusra operational head, Prisca Yosevine, said the technology "is not intended to replace our Driver-Partners, but to complement their role" (AJOT).
The strongest case for caution
The sceptic's argument deserves a fair statement. BVLOS flight removes the pilot's eyes from the aircraft, and this route is in controlled airspace near an international airport, where one conflict with a crewed aircraft could be catastrophic. The corridor also crosses populated areas, so a failure puts people on the ground at risk, and a certificate granted to one operator on one route is not a public rule others can read, rely on or contest. Those are legitimate concerns, and any proportionate framework has to answer them.
Why the pilot is nonetheless the right way to start
SORA is not a loophole. It asks the operator to classify the ground and air risk of a specific operation, then prove that mitigations meet matching safety objectives. DroneDJ's description of the Longtail Dual includes redundancy: it "uses two batteries, allowing the aircraft to maintain flight control if one power source fails." A regulator that demands evidence for a defined route, aircraft and operator is applying proportionality, because the paperwork scales with the risk rather than banning a category outright.
The alternative is the one that has stalled drone logistics in many jurisdictions: waiting until a complete rulebook exists, which means no real-world data on which to write it. The Bali trial produces that data. It shows how an operator deconflicts with airport traffic, how local communities respond, and what failure modes actually appear. This is the kind of evidence-based regulation that serves both safety and innovation.
The rulebook is lagging the aircraft
Indonesia's wider legal framework is still moving. Law No. 21 of 2025 on Airspace Management (Pengelolaan Ruang Udara) was enacted and promulgated in Jakarta on 24 December 2025, published as LN 2025 (192), and revokes several provisions of the 2009 Aviation Law, including Articles 401 and 402 (JDIH Kemenko Infra). The statute describes airspace management as covering planning, use, control and oversight, so it sets the architecture but leaves operational detail to subordinate rules.
On 31 August 2026, Director General of Civil Aviation Lukman F. Laisa told a Jakarta press conference: "Drones are no longer just for recreational use. They are now being used for cargo and even passengers. We need to align this by creating new regulations." He said Indonesia already has regulations on pilots, aircraft registration, navigation and organisations, and that the ministry is guiding two domestic drone companies and overseeing two foreign ones seeking to operate in Indonesia (Antara). The news hook for this article describes the Ministry of Transportation as still drafting cargo and passenger drone rules, revising PM 37/2020. Antara's report, as retrieved, names no regulation number and gives no timeline, so the schedule is unknown.
Where the gap becomes a risk
A certificate-by-certificate approach works while the number of operators is small. Four companies in the ministry's pipeline is manageable. It becomes a bottleneck, and a source of discretion, once dozens of operators want similar routes. Three problems follow.
- Predictability. Investors in drone logistics need to know what a BVLOS approval will require before they ship aircraft. If the standard lives only inside individual certificates, the first applicants write the de facto rules.
- Transparency. Citizens near flight corridors cannot assess risk if the SORA reasoning is not published in some form. Publishing templates and decision criteria costs little and builds the trust that community coordination in Bali depends on.
- Airport interface. Operating in airspace tied to Ngurah Rai is the hardest case. The eventual rules should state who coordinates with air traffic control, how approvals are revoked, and what data operators must report.
What good rules would look like
Indonesia does not need to start from zero. The DGCA has already shown it will apply a recognised risk-based method. The codified successor to PM 37/2020 should do three things:
- Adopt a published, tiered pathway in which low-risk routes get light-touch approval and controlled-airspace routes get stricter conditions, rather than one blanket standard.
- Require operators to publish summaries of their risk assessments and incident data, so the evidence from pilots like Bali's reaches the public and the next applicant.
- Set decision deadlines for the regulator, so that a safe operation is not held up by indefinite review.
The Director General's own remark that every ICAO Annex change "must be followed by an update to our regulations" points the right way: align with international practice rather than invent a distinctive local regime that exporters of drone technology must re-certify against. Indonesia is also seeking a seat on the ICAO Council in elections on 19-20 November 2026, per Antara, which gives it a reputational reason to be seen applying global norms.
Bottom line
The Bali pilot is a good example of regulation doing what it should: permitting a novel service under evidence-based conditions without waiting for perfection. But a pilot is not a framework. The test for Jakarta is how quickly it converts what the DGCA learns in Bali into public, predictable, proportionate rules before the first approvals harden into an unwritten regime.