Following Taiwan's Drone Defense Money: Three Budgets and a Bill Stuck for Two Months

🌏 中文版

Up front: this is budget-process and industry-structure analysis, not investment advice. No tickers, no timing.

The four-criteria piece noted that ~80% of Taiwan's drone output comes from public procurement, which makes budget continuity the sector's dominant variable. This article unpacks that sentence: which budgets exist, what they buy, where they're stuck, and when money might move.

Following the money starts with separating three verbs: appropriated, disbursed, recognized. Years can separate them, and most industry coverage only reports the first.

Three budgets, three different natures

These three flow to entirely different recipients, and conflating them is the most common analytical error:

• The NT$44.2 billion is grants, not orders. Per the Industrial Development Administration, the Ministry of Economic Affairs approved 31 cases of R&D grants through its industrial innovation platform in 2024–2025, supporting work on airframes, flight/navigation control, communications, payloads, and AI autonomous recognition. Receiving a grant means you're doing R&D; it does not mean you have revenue.
• The NT$210 billion is procurement. But it doesn't exist yet.
• Annual budgets are the most fragmented and the most certain. Small amounts, scattered tenders — but they happen every year and require no special statute.

The full timeline of how NT$210 billion got stuck

Worth recording in full, because it demonstrates exactly what the interruption mechanism for "policy demand" looks like.

Note the date in that last row. The draft's period was written to start on 1 August 2026. That date has passed and the statute is still in negotiation.

This is what the gap between "policy demand" and "structural demand" from the four-criteria piece looks like in practice — demand didn't disappear, the disbursement mechanism jammed, and the reasons have nothing to do with drones (the disputes are special budget vs. annual budget, and which ministry should administer it).

The Executive Yuan version buys only three items

Badly underappreciated. Per Koo's committee testimony, the Executive Yuan's statute restricts procurement to specific items:

1. Littoral surveillance drones
2. Littoral attack drones
3. Small suicide unmanned boats

Koo's framing: the Ministry of National Defense procures to operational requirements and "of course should be the competent authority," while "industrial development belongs to the Ministry of Economic Affairs."

What that restriction means: if the NT$210 billion eventually passes, it will not spread evenly across "drone-adjacent" companies. It concentrates on airframe and system suppliers for those three items. Companies doing agricultural spraying, survey, or inspection are not on this path.

Other vehicle types are handled separately: unmanned surface vessels, unmanned underwater vehicles co-developed by NCSIST with foreign partners, and counter-drone systems get introduced via small-batch procurement, while UGVs are already in the 2027 defense budget — that is, via annual appropriations, not the special statute.

How much the public sector is buying

Assembling the scattered procurement plans:

• Over 100,000 airframes planned across the next three years: 50,898 commercial-government units and 48,750 defense units (the defense portion appropriated separately)
• The Armaments Bureau announced in November 2025 procurement of nearly 50,000 new drones worth NT$50 billion
• NT$6.887 billion of militarized-commercial drone procurement was completed in 2024
• The National Police Agency's critical infrastructure counter-drone tender: 50 systems, NT$3.67 billion, qualification round on 24 July 2026

That last one's award design deserves separate note: per UP Media, systems are graded A, B, and C by verified capability, with the top three in each grade delivering proportionally and up to 9 vendors winning; NCSIST manages the project but cannot bid. Deliberate distribution, not concentration — the four-criteria piece covered how that compresses any single vendor's upside. (in Chinese)

Three paths, very different certainty

When assessing any company, first ask which path it stands on. Path A has the largest imagination and the least certainty; Path C has the smallest amounts and happens every year. Market narrative usually discusses A; what actually books is often C.

Three risks specific to budget process

The four-criteria piece already covered acceptance risk (the Army's NT$988M counter-drone contract failing three times and being terminated in full). Three more belong to the budget process itself:

One: legislative risk isn't "will it pass" but "in what form." The Executive Yuan version runs on a special budget; the KMT version on annual appropriations at NT$40 billion a year; the TPP version on regular appropriations. Three forms, three capacity-planning rhythms for suppliers. Koo said it himself — if the government version fails, "we'll have no choice but to put it in the annual budget," while objecting to boxing all unmanned vehicles inside NT$40 billion a year. The form change changes the annual executable ceiling.

Two: item risk. The statute names three items, which makes the overlap between "drone-adjacent company" and "beneficiary of this budget" quite narrow. Reasoning from industry heat to individual companies goes wrong here.

Three: line-item intervention. The legislature intervenes at a granularity commercial orders never see — alongside the counter-drone termination above, it also cut NT$10 million and froze another NT$10 million of the 2026 budget pending a written report.

What to track

If you track only three things:

1. The outcome of cross-party negotiation on the statute — whether and in what form it passes determines whether the NT$210 billion path exists at all
2. When the special budget bill itself is submitted — Koo said the special budget would follow statute passage. Passing the statute is only the first gate; the budget bill is what authorizes disbursement
3. Unmanned vehicle line items in the annual defense budget — the portion unaffected by the statute that happens every year; UGVs entering the 2027 budget is the sample of this path

What these three share: they all sit in the legislature, not in the industry. Which is what makes this sector unusual right now — most of the variables determining revenue are outside any company's control.

This is budget-process and industry-structure analysis, not investment advice. No securities are recommended. Evaluate and bear your own risk.

Review points (added 2026-08-10)

This is the most time-sensitive post in the series and it shipped without a review mechanism. Here it is. Content is current as of 2026-08-06; below is what expires and where to re-check it.

Do not confuse this with the other special act. This post is about the Special Act on Procurement of Indigenous Defence Unmanned Vehicles (NT$210bn, 2026-08-01 to 2031-12-31, stuck in negotiation). A separate Special Act on Procurement for the Plan to Defend National Security and Strengthen Asymmetric Capabilities was promulgated and in force on 2026-05-11, and its Article 5 lists "combined soft-kill/hard-kill counter-drone systems" — that one is handled in Who May Bring a Drone Down. They are different statutes with different amounts, statuses and item lists.

One comparison this post did not carry: per CNA and other reporting, opposition caucuses tabled their own versions — the KMT caucus proposed NT$240bn over six years appropriated through the ordinary annual budget (NT$40bn a year), with a further TPP version. The dispute is mainly about special budget versus annual appropriation, not about whether to buy. This paragraph rests on secondary reporting; I did not read the draft texts.

References

Legislative timeline (all in Chinese)

• CNA — Opposition blocks the Executive Yuan drone special statute; item deferred (2026-06-26)
• CNA — Executive Yuan and opposition drone bills all referred to committee (2026-07-03)
• CNA — Three committees to review the drone bills clause by clause on the 27th (2026-07-25)
• CNA — Committee clears the unmanned vehicle statute with all clauses reserved for cross-party negotiation (2026-07-27)
• CNA — Koo hopes the Executive Yuan drone statute passes; UGVs in the 2027 budget (2026-08-03)
• CNA — Executive Yuan: a special budget balances operational urgency with industry development (2026-07-16)

Budgets and procurement (in Chinese)

• Executive Yuan — Premier Cho: NT$44.2 billion from 2025–2030 to strengthen drone industry capacity
• Executive Yuan — Drone budgets are critical to democratic supply chain strategy and military capability (NT$7.2B+ in annual agency budgets)
• CNA — Legislature reviews drone statute; MOEA urges support for the Executive Yuan version
• CNA — Taiwan builds a democratic drone supply chain (100,000+ airframe procurement plan)
• UP Media — National Police Agency opens tender for 50 critical infrastructure counter-drone systems
• CIO Taiwan — From AI recognition to edge computing (31 approved platform grant cases)

On this site

• The Drone Supply Chain Against a Four-Criteria Framework
• Taiwan's Drone Supply Chain: Where the 267 Companies Are, and Which Layer They're Stuck On
• Four Drone Business Models, and Why Selling Airframes Is the Worst One