Taiwan's counter-drone shield cannot see the drones the PLA has flown since 2023

Taiwan's Audit Office has reported that the military's fixed-site counter-drone programme, built to protect airbases and other key military facilities, cannot detect or jam the 5.1GHz frequency used by newer People's Liberation Army drone models ^[1][3]. The system was delivered by the island's principal weapons developer and is already behind schedule, which makes this less a story about one bad procurement than about the mismatch between a static radio-frequency architecture and a threat that changes frequency faster than a contract can be written ^[2][7].

First, the scale. The programme cost NT$3.55 billion, or about US$111 million ^[1]. That is a modest sum for national base defence, and worth stating plainly because the figure has been mangled: the Korean outlet en.sedaily.com rendered the cost as "$3.55 billion" in its text while its own headline said $110M ^[13]. The audit's implied exchange rate is roughly NT$32 to the dollar ^[1]. A hundred million dollars of fixed infrastructure is a cheap layer to replace and an expensive one to be wrong about.

Now the timing problem. The defence ministry, according to the sedaily report, said most of the technologies flagged as problems emerged after the project began in 2021, and spokesperson Sun Li-fang said NCSIST had already verified new frequency-jamming capabilities and would add functions through future upgrades ^[9]. Take that at face value and the arithmetic is still unflattering: the PLA has fielded 5.1GHz drones since 2023, two years into the programme ^[4][4]. The threat moved during the build, and the build finished anyway.

The 5.1GHz gap is not the only one. The audit also found the system failed to counter some frequency bands used by BeiDou-3, China's satellite navigation system, and warned that effectiveness against fibre-optic-guided and AI-guided drones had not been verified ^[5][6]. Delivery slipped six to 12 months on construction delays and component shortages, and the audit concluded that "the system's detection and countermeasure functions remain incomplete and implementation has fallen behind schedule" ^[7][8]. Lo Ching-sheng of the Taiwan International Strategic Study Society told the same reports that no concrete upgrade timing has been given: "The project has only just been completed, and there is no concrete answer on the timing of upgrades" ^[10].

The second data point is the one operators should weigh. A separate army effort to buy 26 fixed anti-drone systems worth NT$980 million was cancelled last month after the equipment failed three consecutive performance tests ^[11]. That works out to about NT$37.7 million per unit, roughly US$1.2 million each ^[2], and the cancelled buy was about 28 per cent the size of the audited programme ^[3]. Kuomintang lawmaker Ma Wen-chun described the pattern as "a repeated practice of pushing ahead with purchases before verification of equipment performance is even complete" ^[12].

Two things to watch. One is whether the ministry publishes a dated upgrade path for the 5.1GHz and BeiDou-3 bands rather than a promise of future supplementation, since a jammer is only as current as its last frequency list ^[9][3][5]. The other is what replaces the cancelled 26-unit army order, and whether the retest gate comes before payment this time ^[11][12]. Buyers of any sensing-and-jamming stack, defence or commercial, should note the structural lesson: a fixed installation validated against a 2021 threat model was obsolete on delivery, and the audit, not the adversary, found out first ^[1][9].