臺灣海峽動態 — 2026年8月23日
Taiwan's cabinet approved the first defence budget above NT$1 trillion on 20 August — NT$1.1225tn, about 3.01 per cent of projected GDP and roughly 18 per cent above 2026 — and sent it to a legislature the governing party does not control, which has already cut one special budget by a third and has not passed the drone procurement bill. Three days earlier the armed forces ran their first live-fire FPV drone exercises, in the same month an official audit reported technical shortcomings in counter-drone development. Beijing spent the week working the other lever, announcing twelve measures to integrate Kinmen with Xiamen ahead of November's local elections.
各項動態
Taiwan's Executive Yuan approved a central government budget proposal for 2027 on 20 August 2026 totalling NT$3.9266 trillion, of which NT$1.1225 trillion is for national defence — the first time the figure has exceeded NT$1 trillion, equivalent to about US$35.1 billion and to 3.01 per cent of projected GDP. The figure combines the regular defence budget with the special budgets used to fund major procurement, and stands roughly 18 per cent above the NT$949.5 billion budgeted for 2026. President Lai Ching-te described defence investment as an investment in peace and stability across the strait, and has set a target of 5 per cent of GDP by 2030. The proposal was to be sent to the Legislative Yuan for review by the end of the month.
Why it matters. A cabinet budget proposal is a statement of intent by an executive that does not command a legislative majority, and the module's standing discipline on legislation applies exactly as it does to sanctions bills in other modules: the observable is enactment, and then disbursement, not approval. The relevant history is specific rather than general — the same chamber cut the May 2026 special budget request from NT$1.25tn to NT$780bn, and has not passed the proposed NT$210bn drone procurement bill. What the proposal does establish is a measurable public commitment: 3.01 per cent now against 5 per cent by 2030 is a claim that will either be met or not, and the gap is now on the record in the government's own numbers. Reading the announcement as delivered capability would be the error; reading it as meaningless because it is contested would be the opposite one, since the 2026 special budget did eventually pass above the opposition's own counter-proposal.
Taiwan's armed forces conducted their first live-fire first-person-view drone exercises on 17 August 2026, described in reporting as integrating lessons from the war in Ukraine. A proposed NT$210 billion drone procurement bill remains unresolved in the Legislative Yuan. Separately, an official audit has reported technical shortcomings in Taiwan's counter-drone development, and a procurement programme in that area collapsed after repeated testing failures.
Why it matters. Drones are the clearest case in this module of the gap between a doctrine Taiwan has accepted and a capability it has bought. The asymmetric argument — many cheap survivable systems rather than a few large ones — has been settled in Taiwanese public debate for two years, and a first live-fire exercise is evidence that it has reached training. It has not reached procurement: the bill is unpassed and the counter-drone side, which is the defensive half of the same problem, has an audit finding against it and a failed programme behind it. That combination is the specific thing to watch rather than the headline budget total, because drones are where a legislative delay converts directly into a capability gap on a timescale measured in months rather than years — the systems are cheap, fast to field, and useless if not ordered.
The People's Republic of China announced twelve measures to integrate Taiwan's outlying Kinmen county with the mainland city of Xiamen, centred on tourism and trade. Kinmen holds local elections in November 2026, and its magistracy is held by the opposition Kuomintang. The islands sit a few kilometres from the Chinese coast and roughly 190 kilometres from Taiwan proper, and have long had a distinct economic relationship with Xiamen including a ferry link and a water pipeline.
Why it matters. This module records Beijing's pressure as a system with several instruments, and the outlying islands are where the economic instrument is strongest because the geography makes it credible: Kinmen's practical dependencies run west, not east, and always have. An integration package is the same argument as a law-enforcement patrol made in a register nobody can call coercion. The timing against a local election is the part that needs care in both directions — announcing measures before a vote is an attempt to shape a political environment, and treating a Kinmen electoral result as a referendum on unification would badly misread a local contest in a county with its own politics. Vigil records the announcement and the calendar, and draws no inference about either the intent or the effect.
本頁的定位
這是單一衝突本週的註明日期回顧。同一週的跨衝突版本收錄了Vigil所涵蓋各場衝突中最重要的發現。本回顧所更新的常設圖像,是臺灣海峽涵蓋內容。Vigil如何為其主張評級、註明日期並標示來源,載於方法論;任何事後發現的錯誤,均記錄於更正紀錄。
透過電子郵件 透過電子郵件接收Vigil的每週簡報與新的分析札記。Vigil Dispatch寄送哪些內容
← 上一份臺灣海峽簡報 — 2026-08-16下一份臺灣海峽簡報 — 2026-08-30 所有臺灣海峽回顧所有每週簡報