This year's Han Kuang military exercise included, for the first time, a deliberate degradation of Taiwan's internet service. During two half-hour testing windows, on August 10 in central Taiwan and August 13 in northern Taiwan, mobile data speeds were throttled for 17 million people, roughly 73 percent of the population, while voice calls and text messages remained available. The National Communications Commission (NCC) was candid about the scenario being tested: how Taiwan's communications system would function if the undersea cables carrying civilian and government traffic stopped working.

The drill is a sensible exercise, but it raises a question that has received too little attention: if the cables go down, what is actually in place to carry the traffic?

Taiwan has largely won the argument over the need to defend its undersea infrastructure. Submarine cables are now designated as critical infrastructure by the Ministry of Digital Affairs (MODA) and protected under criminal law. Yet the satellite layer expected to take over when those cables fail has faced far less scrutiny. Taiwan's own budget documents, audit reports, and legislative transcripts point to a system considerably thinner than the term "satellite backup" suggests.

What Redundancy Actually Costs

Matsu, Taiwan's outlying island group located closer to the Chinese coast than to Taipei, has had its undersea cables severed more than 20 times over the past five years. One incident in 2023 left residents without normal communications services for more than 50 days. In June 2025, the Tainan District Court sentenced Yuliang Wang, the Chinese captain of the Togo-registered vessel Hong Tai 58, to three years in prison in Taiwan's first criminal conviction for severing an undersea cable. The sentence was upheld on appeal that August.

The conviction rested on a provision of the Telecommunications Management Act that has since been replaced. Prosecutors charged Wang under Article 72, Paragraph 1 of the Act, which the legislature amended in June 2023 following the Matsu cable disruptions. The amendment expanded the law's protection beyond cable landing stations to include the undersea cables themselves. Before that change, damage caused by dragging an anchor across a cable could generally be prosecuted only as ordinary property damage under the Criminal Code.

Legally, therefore, Taiwan's first successful cable prosecution concerned a vessel anchoring in a prohibited area, rather than gray-zone coercion as such. The 2023 provision has since been superseded by a package of seven submarine-cable laws that entered into force in January 2026. Among other changes, the new framework requires the confiscation of vessels used to damage undersea cables. Taiwan has steadily strengthened the law governing the seabed, but it has devoted far less attention to the communications layer expected to take over when those cables fail.

Against that record, satellites look like the obvious hedge. The relevant systems are low-earth orbit (LEO) constellations: hundreds or thousands of small satellites circling a few hundred kilometers up, each visible overhead for only minutes before handing the connection on. Because no single spacecraft is critical, the design resists attack, though it depends on a dense network of ground stations.

They are a hedge, but a narrow one, and Taiwan's own planning documents show how narrow. The Executive Yuan's national communications resilience plan puts the island's fourteen international cables at a combined 109.62 terabits per second. Eutelsat's OneWeb, the constellation the emergency network runs on, carries a little over one terabit per second of usable capacity for the entire planet. MODA's field testing recorded 90 megabits per second at the receiving sites. The resilience plan sorts the tiers accordingly: fiber in terabits, microwave in gigabits, satellite in megabits. Starlink would change these numbers, though not the conclusion. Its network carries several hundred terabits per second globally against OneWeb's one, but that capacity is shared worldwide, constrained on the uplink, and Taiwan would still draw only a slice of it.

The satellite layer provides a vital triage channel for command and coordination, emergency services, and essential civilian communications. It is not designed to carry the full weight of Taiwan's economy, but to keep the most critical connections functioning until normal service can be restored.

The Network that was cut in Half

The most consequential thing to happen to Taiwan's satellite resilience in two years happened in the Legislative Yuan, not in orbit. By October 2024, MODA had built 773 reception sites across Taiwan for non-geostationary satellites, those that move across the sky rather than sitting fixed above the equator. The network combined Eutelsat's OneWeb LEO service with medium-earth orbit (MEO) capacity from the Luxembourg operator SES, plus receiving points in Japan, Thailand and Guam. It grew out of a two-year, NTD 550 million pilot inspired by Ukraine and Tonga, the Pacific island country whose international communications were severely disrupted after a volcanic eruption damaged its undersea cable in 2022.

The network was sharply reduced after the KMT–TPP majority in the Legislative Yuan cut and froze substantial portions of the 2025 central government budget. On January 21, 2025, the KMT–TPP legislative majority approved cuts and freezes totaling an unprecedented NTD 207.5 billion, or roughly US$6.5 billion. MODA's own NTD 8.659 billion budget was cut by NTD 3.582 billion, about 41 percent, with a further 20 percent frozen. The digital resilience line item supporting non-geostationary satellite communications was hit particularly hard, losing NTD 300 million from an original allocation of NTD 530 million, a reduction of 56 percent.

The operational effects were immediate. MODA confirmed that only around 195 priority sites, roughly one quarter of the network, could be maintained in 2025, and even those remained in service only because police and fire agencies absorbed the maintenance costs. Funding for MODA's high-altitude balloon platform was eliminated altogether.

The same pattern soon reappeared on a larger scale. The 2026 general budget, submitted in August 2025, remained blocked from committee review for more than 200 days. The DPP administration's NTD 1.25 trillion special defense budget, announced in November 2025, had likewise been prevented from reaching committee ten times by early 2026. The proposal includes NTD 12.5 billion for satellites and related ground equipment.

At the same time, the Ministry of National Defense committed more than NTD 4.6 billion to lease satellite capacity for drone data links. Leasing can add capacity quickly, but it also depends on sustained annual funding, making it particularly vulnerable during a prolonged budget standoff. That dependence has strengthened the case for developing a sovereign satellite capability of Taiwan's own.

Sovereign Capability and Its Honest Limits

Taiwan has begun to pursue that goal, though progress so far has been uneven. In October 2025, the Executive Yuan approved an expanded thirteen-year national space program with a budget of NTD 71 billion.

The Beyond-5G (B5G) satellite program it funds now shows two experimental spacecraft, 1A and 1B, scheduled on the Taiwan Space Agency (TASA) website for 2028 and 2030, slipping repeatedly from earlier 2026 and 2027 dates. The 1A payload went to the American firm CesiumAstro at NTD 880 million after the Industrial Technology Research Institute (ITRI), Taiwan's state-backed applied research body, had its 1B payload work terminated over technical bottlenecks. TASA's director-general, Jong-Shinn Wu, has told legislators that continuous independent coverage of Taiwan would need 100 to 200 satellites, against a near-term plan of two experimental spacecraft and roughly four communications platforms.

Commercial providers have therefore become the practical backbone of Taiwan's satellite communications system. Chunghwa Telecom, Taiwan's partly state-owned telecommunications incumbent, holds exclusive local distribution agreements for Eutelsat OneWeb's low-earth-orbit service and SES's medium-earth-orbit service. In 2025, it received Taiwan's first commercial licenses for both LEO and MEO satellite communications. In January 2026, it also signed a memorandum of understanding with SES to develop what would be North Asia's first dedicated SES gateway. Far EasTone, one of Chunghwa Telecom's main competitors, secured the Taiwan distribution agreement for Amazon Leo in May 2026.

Taiwan has also moved to make its market more accessible to foreign satellite operators. On July 21, 2026, the Legislative Yuan amended Article 36 of the Telecommunications Management Act, with support from all political parties, to allow the government to waive existing foreign-ownership limits on a case-by-case basis for satellite companies. The amendment was widely understood as an effort to accommodate Starlink. A supplementary resolution also called for local ground stations and data localization.

SpaceX, however, has not applied to enter the Taiwanese market. Talks broke down in May over the company's demand for full ownership, while MODA has also noted that Taiwan's 99.9 percent mobile coverage makes the market less commercially attractive than places with large connectivity gaps. The legislature has therefore removed much of the legal obstacle, but the commercial incentive remains limited.

The Chokepoint Has Moved

While Taiwan argued over line items, the PRC has been launching. Guowang, filed for nearly 13,000 satellites, had roughly 160 to 190 in orbit by mid-2026; SpaceSail's Qianfan crossed 238 by early July and targets 648 by year-end. Roughly half of China's 2025 satellite payloads served these two constellations, and PLA Daily, the People's Liberation Army's official newspaper, cites Starlink's role in Ukraine as the motivation.

How Beijing thinks about suppressing such a network is partly on the public record. In November 2025, four researchers at Zhejiang University and the Beijing Institute of Technology's Yangtze Delta academy published a modeling study on jamming a Starlink-class constellation from the air. The paper gives its purpose as supporting the "regulation and management" of mega-constellations. The reference point for its simulated jamming grid is a set of coordinates at the mouth of the Tamsui River. It found that one jammer, flying at 20 kilometers and drawing 400 watts through a narrow-beam antenna, can deny reception across roughly 38.5 square kilometers of ground. 

Switching off the constellation is a hard problem. Reaching its gateways, the large ground stations that connect it to the internet, is not. Taiwan's traffic runs through gateways in Japan, Thailand and Guam, and the nearest ground station for SES's medium-orbit fleet sits at Stanley Fort in Hong Kong until the planned Taiwanese facility opens. The chokepoint has migrated from the seabed to a handful of buildings and the jurisdictions that host them.

Beijing has made that vulnerability explicit. In response to the July amendment, the Global Times quoted a Chinese military commentator outlining two possible approaches: 1) attacking or jamming ground terminals, and 2) disrupting the satellites themselves in orbit. He also pointed to Elon Musk's substantial business interests in China and suggested that these should factor into SpaceX's decisions. The implication is clear, a satellite constellation that Taiwan does not own can be pressured not only through its physical infrastructure, but also through the commercial interests of its operator.

Where Taiwan Has Leverage

Taiwan's contribution was never likely to be measured by the number of satellites it could place in orbit. First, Taiwan's greater advantage lies in the ground segment and the broader supply chain. The Ministry of Economic Affairs has assembled a national team for satellite ground equipment that includes Wistron NeWeb and Quanta, while Taiwanese firms have already developed prototype Ku-band antennas for commercial satellite broadband.

The remaining gaps are more specialized, in radio-frequency and signal-processing chips, as well as the phased-array circuits that allow flat-panel antennas to track moving satellites electronically rather than mechanically. Closing those gaps would strengthen both Taiwan's communications resilience and its industrial position. Ground terminals, after all, are the part of the system that Taiwan and its partners can manufacture, deploy, and stockpile in advance.

Second, Taiwan can contribute through spectrum diplomacy. Although it is not a member of the International Telecommunication Union (ITU), the UN agency responsible for coordinating radio frequencies and orbital slots, Taiwan has direct experience navigating a system in which it cannot submit coordination requests on its own and must instead rely on filings made through foreign operators. It is also excluded from the UN Committee on the Peaceful Uses of Outer Space, while its launches are recorded in the UN register under "Taiwan, Province of China."

That experience gives Taiwan a practical role in helping democratic partners understand how political exclusion affects access to spectrum, orbital resources, and satellite services. The 2027 World Radiocommunication Conference (WRC-27), where ITU members will renegotiate international spectrum rules, makes that contribution especially relevant. Because the conference will be held in Shanghai, the rules affecting Taiwan's access will be discussed on the territory of the government responsible for excluding it from the process. The European Parliament's January 2026 support for EU–Taiwan cooperation on undersea cables and satellite constellations offers a starting point for partners to bring Taiwan's technical experience and policy concerns into those discussions.

Third, Taiwan can turn exercise design into a form of international cooperation. Its recent drill will be the country's first nationwide test of a degraded internet environment. Japan is developing its own sovereign satellite capability through a ¥150 billion constellation program and a ¥283.1 billion defense satellite contract, while Europe's planned IRIS² constellation is expected to begin launching around 2029. Both will need to understand how communications systems perform when terrestrial networks are disrupted.

Taiwan will be among the first democracies to generate that operational experience. In January, a workshop held under the Global Cooperation and Training Framework, the capacity-building platform Taiwan operates with the United States, Japan, Australia, Canada, and the United Kingdom, brought 197 participants from 26 countries to the Taiwan Space Agency and the Toucheng cable landing station. The lessons from Taiwan's exercises can therefore become shared practice among its democratic partners.

The policy priorities are straightforward. Taiwan should fund the satellite ground segment and reception network as permanent operational infrastructure, rather than as temporary pilot projects. It should require domestic gateways and appropriate data localization as conditions for foreign satellite operators seeking access to the Taiwanese market, building on the approach set out in the July legislative resolution. It should instead define satellites as a complementary layer to undersea cables, with each system designed around the role it performs best.The two systems perform different functions: cables sustain the full flow of Taiwan's economy, while satellites preserve critical communications when that flow is disrupted. Resilience depends on designing them as complementary layers, not interchangeable ones.

Taiwan's comparative advantage lies in the experience it has gained from confronting a communications threat that most democracies have only begun to consider. By strengthening its ground infrastructure, developing trusted supply chains, and sharing the lessons of its exercises with international partners, Taiwan can turn that experience into a wider contribution to democratic resilience.

Taiwan has already shown that vulnerability can become expertise. Its experience protecting undersea cables has helped move it from being seen as a potential weak point in the network to becoming a source of stronger standards and shared practice. The same transition is now possible in space. The task is not simply to build a backup for when the cables fail, but to create a resilient communications architecture that Taiwan and its partners can rely on when it matters most.