Defence · Company direction
An archipelagoguards itself.
Dewata AI is moving into the defence industry on a forty-year plan. Software first, then robotics, then weapon systems. Built for Indonesia, through Indonesia's official channels.
A forty-year direction. No product, licence, or contract yet.
- Status
- Direction, not product
- Buyer
- The state, through official channels
- Control
- A human decides
- Legal basis
- Law 16/2012
Software first.Then the steel.
01 · Why
The sea is vast.The guards are few.
Indonesia has 17,380 islands and 6.4 million km² of water. Crewed patrol ships and aircraft will never be enough to cover all of it.
What can cover it: cheap sensors, data fused into one picture, and uncrewed machines in large numbers. That is a software problem before it is a steel problem, and software is what we do every day.
-
17,380
Named, charted islands
Badan Informasi Geospasial: Indonesia's island count, 2024 -
6.4M km²
Sea area
KKP, Data Rujukan Nasional Kelautan: Maritime Affairs Ministry, Navy Hydrography, and BIG, 2018 -
108,000 km
Coastline
Mongabay Indonesia: Launch of the official marine territory data, 2018
02 · The stages
Forty years.Each stage has a gate.
Realistically, a software company from Denpasar cannot jump straight to missiles. So we start with what is closest, and only move up a stage once its gate is met.
- Running or in research
- Planned
- Direction
+Show the numbers
| Stage | Status | Earliest start |
|---|---|---|
| 00 Funded by ventures already running | Running | 2026 |
| 01 Defence software | Research | 2026 |
| 02 Robotics and uncrewed systems | Planned | 2028–2029 |
| 03 Stakes in defence companies | Planned | 2029–2031 |
| 04 Guided weapon systems | Planned | 2033–2036 |
| 05 Our own factory and supply chain | Direction | 2036–2041 |
| 06 A defence venture fund | Direction | 2038–2044 |
| 07 Full system integrator | Direction | 2041–2046 |
| 08 Exports to partner states | Direction | 2046–2051 |
| 09 Satellites and space | Direction | 2051–2056 |
| 10 Next-generation technology | Direction | 2056–2061 |
| 11 Handed to the next generation | Direction | 2061–2066 |
2026–2036 · Build the base
- 00Running
2026 Earliest: Now
Funded by ventures already running
What gets built
- Dewata Tech and Dagangku AI keep running and growing
- A team trained on AI, data, and systems that cannot go down
Starts when
Already met. Every later stage is funded from here, not from debt.
- 01Research
2026 Earliest: Now
Defence software
What gets built
- Maritime awareness: radar, AIS, and satellite imagery in one picture
- Command and control: who sees what, who decides what
- AI-driven simulation and training
Starts when
Buildable from Denpasar today. To move on to stage 02: one real trial with a defence or maritime security user.
- 02Planned
2028–2029 Earliest: 2–3 years out
Robotics and uncrewed systems
What gets built
- Surveillance drones for sea and border patrol
- Uncrewed surface vessels
- Ground robots for logistics and explosive ordnance disposal
- Counter-drone systems for critical sites
Starts when
Defence Ministry approval to enter the defence industry, an in-house hardware team, and a lawful test range.
- 03Planned
2029–2031 Earliest: 3–5 years out
Stakes in defence companies
What it looks like
- Listed defence companies such as Palantir. Passive holdings, no board seat
- Private companies such as Anduril only if a route opens. Their shares are restricted, and a foreign investor asking for more than a passive stake is reviewed by CFIUS in the US
- Defence startups in Indonesia and Southeast Asia, as a minority shareholder and a partner
- Every investment follows the same limits as this page
Starts when
Twelve months of reserve secured, the core product not short of capital, and the split of surplus between this and Bitcoin published on the thesis page before it starts. No debt.
Before we can build it, we can own a piece of it, and learn from their reports every quarter. Someone else's shares will never replace a product of our own.
- 04Planned
2033–2036 Earliest: 7–10 years out
Guided weapon systems
What gets built
- Guided rockets and missiles
- Precision bombs and munitions
- Short-range air defence
Starts when
Approval from the Minister of Defence, a state-owned defence partner, and Defence Ministry oversight from production through sale. This one we will not do alone.
Precision is also an ethical choice. The more exact the weapon, the fewer people it reaches who should not be reached.
2036–2046 · Make it at home
- 05Direction
2036–2041 Earliest: 10–15 years out
Our own factory and supply chain
What gets built
- A plant for drones, robots, and weapon-system components in Indonesia
- A local supply chain for electronics, batteries, and materials
- Engineers and technicians we train ourselves, from Bali and beyond
Starts when
Repeat orders from stage 02 or 04 large enough to justify a plant, a defence industry licence, and Defence Ministry oversight on the factory floor.
- 06Direction
2038–2044 Earliest: 12–18 years out
A defence venture fund
What it looks like
- Our own fund for Indonesian defence technology startups
- Seeded from operating profit and stage 03 returns
- Every startup it backs follows the same limits
Starts when
Stage 03 has run long enough to show checkable results. If outside capital ever joins, the structure follows OJK rules.
Anduril and Palantir were startups once. This stage is about finding the Indonesian ones early.
- 07Direction
2041–2046 Earliest: 15–20 years out
Full system integrator
What gets built
- One command network across land, sea, and air
- Layered air defence for cities and critical sites
- Systems from stages 02 and 04 joined into one
Starts when
Designated by the government as a lead integrator, systems from stages 02 and 04 in use by the armed forces, and the stage 05 plant under Defence Ministry oversight.
2046–2056 · Beyond Indonesia
- 08Direction
2046–2051 Earliest: 20–25 years out
Exports to partner states
What gets built
- Systems proven in Indonesia, for other countries
- Training and after-sales support in the buyer's country
Starts when
An Indonesian government export licence for every contract, and only to states the government approves. Never to non-state actors.
- 09Direction
2051–2056 Earliest: 25–30 years out
Satellites and space
What gets built
- Our own maritime observation satellite constellation
- Military communications that depend on no other country
- Ground stations in Indonesia
Starts when
Partnership with the national space authority, and frequency and orbit licences through official government channels.
2056–2066 · The next generation
- 10Direction
2056–2061 Earliest: 30–35 years out
Next-generation technology
What gets built
- Sensors, autonomy, and guidance designed entirely in Indonesia
- Research with Indonesian universities, with patents filed here
- Defences against threats that do not have a name yet
Starts when
Honestly, this far out cannot be planned yet. What we can promise: the limits on this page stay the same, however advanced the technology gets.
- 11Direction
2061–2066 Earliest: 35–40 years out
Handed to the next generation
What gets built
- The company led by the next generation, ideally Balinese who grew up inside it
- The limits on this page written into the company's articles of association
- Ownership stays Indonesian
Starts when
The next leaders have been proven inside the company, and the limits are fixed in the articles before leadership changes hands.
The timings are our estimate, not a promise, and they get rougher the further out they go. Stage 05 onwards is a direction, not yet a plan. If a gate is not met, the stage slips rather than being forced.
03 · Limits
What we willnot build.
A weapons company without written limits is waiting to be used by the wrong people. These are ours, and they cover what we invest in too.
- No sales to non-state actors, militias, or individuals
- No exports without Indonesian government approval
- No anti-personnel landmines or cluster munitions
- No chemical, biological, or nuclear weapons
- No weapon that selects and engages a target without a human decision
- No use in watching or policing our own citizens
04 · Where it stands today
What exists,and what does not.
A row here changes only when its status actually changes. Not before.
| Item | Status |
|---|---|
Ventures funding it Dewata Tech and Dagangku AI | Running |
AI software and data capability Used every day across our two ventures | In place |
Defence product None yet, prototypes included | Not yet |
Defence Ministry approval Not applied for until stage 01 has results | Not yet |
Contracts or customers Government or private, none yet | Not yet |
Hardware team and facilities Built in stage 02 | Not yet |
Stakes in defence companies Not a single share yet, stage 03 at the earliest | Not yet |
State-owned defence partner Sought in stage 04 | Not yet |
Note
- This page describes the company's direction. It is not an offer of products, services, or securities.
- Private ownership of a main defence equipment business requires the approval of the Minister of Defence and falls under the Defence Ministry's oversight system, under Law 16/2012 on the Defence Industry as amended by the Job Creation Law.
- Indonesia is a party to the Ottawa Convention on anti-personnel mines. The cluster munition limit on this page is the company's own choice, not a legal obligation in force today.
- Palantir and Anduril are named as examples. We have no relationship with either company, and naming them is not an investment recommendation.
Engineer, veteran, or in the defence industry?
We are looking for people who want to build this from stage 01, under the same limits set out on this page.