Taiwan's Drone Cluster Is Forming. Can Its Own Party's Bill Undo It?

Taichung Mayor Lu Shiow-yen spent last week asking Taipei for more central government money to build out a drone industry corridor anchored in her city's existing semiconductor, optics, and aerospace base. Her own party is drafting the bill that would work against that geography: a site-dispersal mandate requiring drone production and storage to spread across at least three counties, with no single site holding more than 25 percent of capacity. Lu is KMT. So is the caucus writing the mandate.

Two caveats belong at the top. The KMT's provisions discussed below are still draft language moving through committee, not enacted law, and the outcome of the current legislative standoff over Taiwan's domestic drone budget remains genuinely unresolved. The military commentator quoted most extensively here, Zhuge Fengyun, was giving his own reading of the bill's likely effects, not speaking as one of its drafters.

A mayor's contradiction points at the real geography

Taiwan's drone base already has a shape, and it isn't evenly distributed: production and R&D cluster around Taoyuan, Taichung, and Chiayi, built on precision manufacturing, optics, and aerospace supply chains that predate the drone boom by decades. We've covered the scale of that cluster before, more than 260 companies and counting, so the point worth making here is narrower: the geography is now a live political fault line, not just an industrial fact.

At a forum on Taichung's unmanned vehicle industry, Lu argued her city and its neighboring counties already have the semiconductor, optics, aerospace, materials, and precision-manufacturing base to turn drone research into exportable product, and asked Taipei for the resources to build on it. DPP Legislator Ho Hsin-chun, at the same event, went further, arguing that stable government funding is what lets companies commit to R&D in the first place. Neither mentioned the provision working through the legislature that would cap how much of that funding could concentrate in Taichung at all.

A bill written against the geography it's meant to fund

The KMT caucus frames the dispersal mandate as a resilience measure: spreading production and storage across three administrative regions, it says, reduces the risk that a single strike or disruption could take out Taiwan's drone manufacturing capacity in one go. The localization targets sit alongside a proposed ban on high-risk supply chains for five core components, including flight control systems, navigation chips, and communications modules, plus a supply chain security certification regime the caucus says would apply across the industry.

The dispersal rule is where the draft collides hardest with how the cluster actually works. Zhuge told reporters the three-county, 25-percent-cap requirement is close to meaningless as written, since a manufacturer can satisfy it by changing a registered address without moving any actual production. What it would do, in his assessment, is write siting rules into statute that constrain military deployment flexibility better left to tactical judgment, while chipping away at the clustering effect that lets Taoyuan, Taichung, and Chiayi suppliers share subcontractors, testing facilities, and skilled labor. DPP Legislator Wang Ting-yu raised a separate objection: that the KMT draft's attempt to compel Executive Yuan budgeting through statute conflicts with the constitutional separation of budgetary authority, a dispute that could stall the bill in committee independent of its industrial-policy merits.

The provision isn't happening in a vacuum. The underlying budget it would attach to funds a specific, large program: 1,446 coastal reconnaissance drones, 208,200 coastal attack drones, and 1,320 unmanned surface vessels through 2031, per the Cabinet's original special-act text. That bill was blocked from the agenda twice in June, after the KMT and TPP had already stripped drone procurement funding out of the NT$1.25 trillion special defense budget back in May. On July 3, the Legislative Yuan referred all four rival versions, the Cabinet's NT$210 billion special-budget bill, the KMT's version, the TPP's version, and DPP Legislator Lin Chu-yin's version, identical to the Cabinet's, to joint review by the Finance, Economics, and Foreign Affairs and National Defense committees. The fight now is over which version survives committee markup, not over whether the bills reach committee at all. The KMT caucus has put a number on its own version at its own press conference: NT$240 billion over six years, funded through the annual central government budget at roughly NT$40 billion a year rather than as a special budget.

The localization mandate meets a chip supply that doesn't exist yet

The KMT draft's other headline provision requires commercial drones to reach 50 percent domestic content within two years and 80 percent within four. Zhuge called this the bill's "fatal flaw": Taiwan's current control systems, communications modules, and AI components are, by his account, largely sourced from the US and Europe, and no domestic supplier can absorb that share of the bill of materials on the timeline proposed.

That assessment lines up with a structural problem Taiwan's own semiconductor sector hasn't solved on its own. Writing in UP Media, National Taipei University of Business professor Chang Jui-hsiung laid out the "three chips" gap: flight-control chips, positioning modules, and communications chips, none of which Taiwan currently mass-produces domestically, despite TSMC and UMC operating some of the most advanced fabs in the world. The constraint isn't fabrication capability, it's order volume. No chip designer has filed for military-grade certification and no fab has opened a dedicated line, because the domestic drone market hasn't been large enough to justify either investment. Demand, not capacity, is the bottleneck, which means a localization quota on the output side does nothing to fix a shortage sitting on the input side. Chang's argument is that only a large, sustained government order book, the same NT$210 billion the two legislative factions are fighting over how to spend, can break that cycle.

Where foreign entrants fit, and what the mandate could cost them

The clearest sign the cluster works is who's already using it. Redwire's Coast Guard contract for its Penguin Mk2.5, awarded through Taiwan Color Optics rather than as a direct foreign-vendor deal, is the example we flagged in our own reporting as a channel worth watching. We won't re-run that analysis here. The open question the dispersal and localization provisions raise, and one the bill text doesn't yet answer, is whether a subsidiary-routed structure like Redwire's would count as domestic content at all under a strict origin test, or whether it falls into the same ambiguity that's already dogging "non-red" claims across the sector. A mandate built to keep Chinese components out could end up complicating exactly the kind of foreign-OEM entry point that's currently working.

The regional contrast is with Japan, not with anything happening domestically. Portuguese firm Tekever is negotiating to build the first foreign-operated military drone production line on Japanese soil, with trading house Marubeni acting as its Japan sales agent, to manufacture its Ukraine-combat-tested AR3 and AR5 reconnaissance drones for the Japanese market and Asian export, with the Self-Defense Forces a likely eventual customer. That's a different model from the certification-and-standards gap our June 29-July 5 roundup covered on the Taiwan-Japan side. Tokyo is courting a dedicated foreign production facility; Taiwan is drawing foreign OEMs in through subsidiary and joint-development structures layered onto an existing cluster. The KMT's draft, if enacted as written, would push Taiwan's version further from the model that's currently working and closer to the one Tokyo is still trying to build from scratch.

What the industry is already building instead

Writing separately in UP Media, Taiwan think tank advisor Chou Yu-ping argued that unmanned systems and counter-UAS are the areas best suited to a technology-transfer-plus-co-production model with the US, precisely because they draw on Taiwan's existing semiconductor, ICT, and AI strengths rather than requiring an aerospace industrial base built from scratch. That's a description of what Redwire is already doing through its Taiwanese subsidiary structure: not full domestic origination, and not a simple import either, but shared production sitting inside the existing cluster. A hard localization percentage measures the wrong thing if the industry's actual competitive position comes from co-production relationships a strict domestic-content test would penalize rather than reward.

What we're tracking

All four bills are now in joint committee markup after the July 3 referral, and Kung's signal that the Cabinet would accept either budget structure suggests more room for compromise than the two prior June blockages implied. We assess the probability that the dispersal and hard localization provisions survive committee markup in anything close to their current form at roughly 20 to 30 percent, weighted down by the trivial-workaround critique undercutting the dispersal rule's practical rationale, the semiconductor supply constraint undercutting the localization timeline, and a KMT mayor publicly asking for the kind of concentrated funding her own party's bill would cap. The more likely outcome is a compromise that keeps the funding scale under dispute while dropping or loosening the geographic and content mandates. Watch whether Lu's corridor ask gets any traction in Taichung independent of the national bill, and whether more foreign OEMs use the subsidiary-routing model before the legislature settles anything.

→ Subscribe to Taiwan Drone Weekly for ongoing coverage: www.taiwandrones.com/subscribe
→ Suppliers: contact us at info@taiwandrones.com

This analysis draws on reporting from the Liberty Times, UP Media commentary, and TaiwanDrones' own prior coverage, and reflects the state of public information as of July 10, 2026. Passage probability estimates represent analytical judgment, not legal advice or an official policy forecast. The draft bills discussed remain subject to committee revision and should not be treated as final law.