Taiwan’s Climate Crisis and Net-Zero Transition: On the Day the Maanshan Nuclear Plant Referendum Failed, the Choices Set by Physical Limits Had Only Just Begun

30-second overview: On the evening of August 23, 2025, votes were counted in the referendum on extending the service life of the Maanshan Nuclear Power Plant: 4.34 million yes votes, 74% in favor, turnout of 29.53%, and 650,000 votes short of the threshold. The next day, Lai Ching-te announced three principles: “nuclear safety without concern, a solution for nuclear waste, and social consensus.” Seven months later, on March 27, 2026, Taipower submitted its application to the Nuclear Safety Commission to restart Maanshan, with a restart possible as early as 2028¹². The referendum failed, yet Taipower is moving back toward nuclear power. This is the deepest contradiction of a country that imports 98% of its energy and has promised to invest NT$9 trillion to reach net zero by 2050³. The government’s geothermal target is 200 MW by 2030, but only 7.4 MW was actually in commercial operation by the end of 2025, a gap of 27 times; the Lanyu storage site opened in 1982, holds 97,672 barrels of nuclear waste, and has missed its relocation deadline four times⁴⁵. Energy is a question of physical limits.

The Maanshan Nuclear Power Plant in Hengchun, Pingtung, located on the Nanwan coastline. Unit 1 shut down on January 1, 2025; Unit 2 shut down on May 17. Image: M. Weitzel, CC BY-SA 3.0, via Wikimedia Commons

The Day of the Maanshan Referendum

On the evening of August 23, 2025, votes were counted across Taiwan’s 22 cities and counties, and the result of the referendum on extending the service life of the Maanshan Nuclear Power Plant was announced: 4,342,206 yes votes, 1,511,693 no votes, and 74.17% in favor. But turnout was only 29.53%. Under the Referendum Act, yes votes had to reach one-quarter of all eligible voters, or 5,000,523 votes; the yes side fell short by 658,317 votes¹. More than twice as many people voted yes as no, yet the referendum failed.

📝 Curator’s note: The common interpretation is that “74% in favor = clear public support for nuclear power,” but this reverses cause and effect. The Referendum Act is not designed simply to ask which side has more votes. It requires a mobilization threshold, proof that “enough people care.” A turnout of 29.53% means more than 70% of voters chose not to go to the polls. That is a more awkward third signal: many people do not feel strongly enough about energy questions to walk into a polling station.

Two days later, on August 25, President Lai Ching-te held a press conference and gave his response: restarting nuclear power would have to pass through three gates, “nuclear safety without concern, a solution for nuclear waste, and social consensus”². It sounded reasonable, but every one of those conditions is a problem Taiwan has failed to solve for 50 years.

Then came March 27, 2026. Taipower submitted a plan to the Nuclear Safety Commission to restart Maanshan, initiating the post-shutdown safety inspection process for Unit 1. The safety review is expected to take about 18 months, with restart possible as early as 2028¹. This is a return.

Exactly seven months passed between vote-counting and submission for review. In between, nothing changed: nuclear waste was still on Lanyu, no final repository site had been chosen, and social consensus remained divided. But the administrative process moved. That is the question this article seeks to answer: when a democratic vote rejects something while the executive branch simultaneously advances it, who, exactly, is deciding Taiwan’s energy policy?

Lanyu, 1982 to 2057

To understand the Maanshan story, one must first understand the Lanyu story.

In 1982, Taipower opened a low-level radioactive waste storage site on the coast south of Longmen, Lanyu. At the time, Tao residents were told it was a “fish cannery,” a claim that later became one of the most frequently cited cases of deception in Taiwan’s environmental justice history⁶. In 1988, Tao people launched their first large-scale protest, using the traditional ritual of “driving out evil spirits” to express their rejection of nuclear waste. It became the starting point of Taiwan’s Indigenous environmental movement.

Over the next 38 years, relocation promises were broken four times: in 1996, the government promised to move the waste out by 2002, then missed the deadline in 2002; it then missed deadlines again in 2016, 2019, and 2023. As of 2024, the Lanyu storage site held a cumulative 97,672 barrels of low-level radioactive waste. The Atomic Energy Council has required Taipower to complete relocation by 2029, but the destination remains undecided⁴.

If the 2029 deadline is missed again, as the industry widely expects, Lanyu’s nuclear waste will have been stored there from 1982 to 2057, a total of 75 years. An offshore island of 4,000 people will have borne the byproduct time of four national nuclear power plants for longer than most people in Taiwan live.

⚠️ Contested view: Nuclear advocates often say that “nuclear waste is technically solvable; the obstacle is only political resistance.” But the nuclear waste problem has always been about the dimension of time. Lanyu has already spent 44 years with nuclear waste since 1982, and promises have never been fulfilled. In the most optimistic scenario, the waste is moved out in 2029. But what happens after “moving it out of Lanyu”? The siting of a final repository remains blocked, and local resistance in Daren Township, Taitung County remains unresolved. Technically feasible ≠ politically feasible ≠ ethically feasible. Lanyu is where the gap among these three levels becomes concrete.

PanSci reports that spent nuclear fuel rods remain hot and highly radioactive after reactor decommissioning and must be cooled in on-site spent fuel pools for at least five years before they have any chance of being moved out. Meanwhile, land-use issues for dry cask storage facilities at the Jinshan and Kuosheng nuclear plants have been stuck for more than 11 years. The New Taipei City government has refused to approve dry storage facilities, leaving spent fuel rods still stored in on-site fuel pools, already beyond their original designed capacity⁷⁸. “The greatest obstacle to extending nuclear power is the question of where spent nuclear fuel will go,” PanSci’s quotation of the industry consensus, is the most awkward background noise to the Maanshan restart application⁸.

The Physical Limits of Nuclear Waste

Shift the frame to Olkiluoto Island in southern Finland.

Five hundred meters underground, in a granite bedrock layer, a tunnel five kilometers long has been excavated. At the end of the tunnel is Onkalo, the first final repository for high-level radioactive nuclear waste in human history to formally receive a trial operation permit. In August 2024, Finland’s nuclear safety authority STUK issued the permit. The project has been planned since the 1970s; it took nearly half a century⁹.

Entrance to Finland’s Onkalo final repository for high-level radioactive nuclear waste, 500 meters underground in granite bedrock; trial operation permit obtained in 2024. Image: kallerna, CC BY-SA 4.0, via Wikimedia Commons

Onkalo is designed to isolate nuclear waste for more than 100,000 years. How extreme is that timescale? Human civilization is about 10,000 years old; the oldest pyramids are 4,500 years old; 100,000 years ago, our ancestors had not yet left Africa¹⁰.

💡 Did you know: A final repository for nuclear waste must isolate it for so long that it reaches “the end of human memory.” Onkalo’s design team spent years discussing one question: how can people 100,000 years from now be told, “Do not dig here”? By then, no existing language, symbol, government, or religion will remain. The final approach uses nuclear waste warning symbols along with multilingual warnings, but the designers acknowledge that this is only “a message for the next 1,000 years.” No one knows the answer for what comes after those 1,000 years.

What about Taiwan’s final repository? The candidate site for a final repository for low-level nuclear waste is Daren Township, Taitung County, but the siting process has fallen into difficulty under local political resistance¹¹. The siting process for a final repository for high-level nuclear waste has not even begun. Finland spent 50 years getting to trial operation. Taiwan has spent zero years.

PanSci also mentions another “physical-limit solution” that was once seriously discussed: disposing of nuclear waste in space. “The idea of disposing of nuclear waste in space is physically feasible, but it would require extremely stable and insured rockets; otherwise, if a launch failed, the radiation contamination caused to Earth would be difficult to estimate”¹². The failure rate of SpaceX’s Falcon 9 is about 1%, meaning that once in every 100 launches, high-level radioactive nuclear waste would be blasted into the atmosphere. Physically feasible, and also physically infeasible.

This is the physical limit behind the four words “a solution for nuclear waste.” Its timescale is longer than all of human civilization.

The Hydrogen Rainbow: Green, Blue, Gray, White and Gold

If nuclear power is too heavy, can Taiwan route around it?

Over the past five years, hydrogen has been seen as the next wave in the energy transition. The problem is that hydrogen itself is an energy carrier, not an energy source: other energy must first be used to “make” hydrogen, which is then used to generate electricity or as fuel. Where it comes from determines whether it is truly “clean.”

PanSci divides hydrogen into different colors by production method: “Hydrogen color codes correspond to different production methods: gray hydrogen (natural gas SMR, emitting CO₂), blue hydrogen (gray hydrogen + CCS), green hydrogen (water electrolysis powered by renewable electricity), and turquoise hydrogen (methane pyrolysis, solid carbon without CO₂ emissions). From a carbon-emissions perspective, green hydrogen is the ideal, but it has the highest cost”¹³.

Hydrogen color	Production method	Carbon emissions	Cost	Taiwan status
Gray hydrogen	Steam methane reforming of natural gas (SMR)	High (emits CO₂)	Low	Most common in industry
Blue hydrogen	Gray hydrogen + carbon capture and storage (CCS)	Medium (reduced after CCS)	Medium-high	No commercial operation
Green hydrogen	Water electrolysis using renewable electricity	Zero	High	CPC planning stage
Turquoise hydrogen	Methane pyrolysis (decarbonized hydrogen combustion)	Zero (produces solid carbon)	Medium	Hsinta Power Plant trial
White / gold hydrogen	Naturally formed underground	Zero (no manufacturing required)	Pending exploration	None

Taiwan’s hydrogen test site is the Hsinta Power Plant in Kaohsiung. Taipower and Academia Sinica are collaborating to test “decarbonized hydrogen combustion”: natural gas, or methane, is decomposed at high temperature into hydrogen and solid carbon; the process does not produce carbon dioxide, and the solid carbon can be used as an industrial material¹³. The appeal of this technology lies in the fact that it can reuse existing natural gas infrastructure, without rebuilding the entire energy system from scratch.

But hydrogen has its own physical limits. “Although hydrogen is clean energy, its effect as a greenhouse gas is 11.6 times that of carbon dioxide (GWP100). If it leaks during production, transportation, or use, it may instead worsen warming”¹⁴. A hydrogen molecule is the smallest molecule in the universe, and its leakage rate is inherently high. This is a physical limit in materials science, not something engineering effort can fully overcome.

There is also a newer player: white / gold hydrogen. A 2023 study by the U.S. Geological Survey (USGS) estimated that naturally occurring hydrogen formed underground by crustal movement may amount to “tens of billions of tons,” enough to provide human energy needs for the next several hundred years¹⁴¹⁵. France and Mali already have commercial exploration. Taiwan sits on an active plate boundary and theoretically has potential, but currently has no exploration plan at all. This is the energy option furthest from theory.

📝 Curator’s note: The core point readers should remember about hydrogen’s rainbow classification is that behind the phrase “clean energy,” one must always ask: where does the energy come from? Green hydrogen is economical only when renewable electricity is so abundant that there is nowhere left to sell it. Taiwan has not reached that scenario. Until then, hydrogen is in fact another showroom for fossil fuels.

Geothermal Taiwan: 33 GW of Potential vs. 7.4 MW of Reality

If hydrogen is a “carrier dispute,” geothermal is a “depth dispute.”

Taiwan should have been a geothermal power. Located at the junction of the Eurasian Plate and the Philippine Sea Plate, its volcanoes, hot springs, and seismic zones form a natural geothermal resource bank. In 1981, the 3 MW demonstration unit at Qingshui Geothermal in Yilan was built, becoming Taiwan’s first geothermal power plant. But due to technical problems such as downhole scaling and acid corrosion, it was shut down in 1993.

For the next 30 years, geothermal power in Taiwan fell silent. Only in 2020, when the privately invested 4.2 MW Qingshui Geothermal unit restarted commercial operation, did geothermal power reenter public discussion. In 2024, construction began on the 5.4 MW Tuchang Geothermal project in Yilan, expected to start operation in early 2026. By the end of 2025, Taiwan’s total commercially operating geothermal capacity was 7.4 MW¹⁶.

What are the government’s official targets? 200 MW by 2030 and 6 GW (6,000 MW) by 2050. From 7.4 MW to 200 MW is a 27-fold gap; to 6 GW, an 810-fold gap. These are five-year and 25-year timelines.

PanSci cites a National Taiwan University study stating that “Taiwan’s geothermal resources are widely distributed. According to research by National Taiwan University, deep geothermal resources (below five kilometers) have potential generation capacity as high as 33,640 MW, equivalent to about 12 Lungmen Nuclear Power Plants”¹⁷. But that is only a theoretical figure. Developing deep geothermal energy requires EGS, or Enhanced Geothermal System, technology: drilling several kilometers underground and injecting water to create an artificial heat-exchange layer. At present, there are only a few demonstration projects worldwide, and the technology has not yet been commercialized.

Meanwhile, “Taiwan’s shallow geothermal development potential (within three kilometers) is estimated at no more than 1,000 MW, and several experimental projects are now underway in Yilan’s Qingshui area and Taipei’s Datunshan region”¹⁷. Even if shallow geothermal were fully developed, it would provide only about 3% of Taiwan’s total electricity demand.

Geothermal’s advantage is stability. “The advantage of geothermal is that, unlike wind or solar, it is not affected by weather; it is a baseload power source that generates steadily 24 hours a day, which gives it unique value in the energy mix”¹⁸. Not many renewable energy sources can replace nuclear power’s baseload function; geothermal is one of them, provided it can actually be built.

⚠️ Contested view: Slow geothermal development in Taiwan is often blamed on “immature technology.” But PanSci’s interviews with the industry reached a different conclusion: the real bottleneck is subsurface uncertainty + difficulty obtaining loans. Before a geothermal well is drilled, no one can guarantee that water will flow, how hot it will be, or how long it will last. Banks do not lend, and developers hesitate to invest. Japan and New Zealand face similar difficulties, but both countries have government-guided funds that share risk. Taiwan’s geothermal developers can currently only use the financing model for solar photovoltaics: once solar panels are installed, they generate electricity; geothermal wells do not. Copying the financing structure guarantees a bottleneck.

In March 2023, PTS’s Our Island produced a two-part series, “Generating Power with Heat,” following national teams into Datunshan, Yilan, and Taitung communities, giving a full account of both “the diagnostic process of geothermal exploration” and “geothermal power generation in Indigenous communities”:

PTS _Our Island official channel: Episode 1195, “The National Geothermal Exploration Team Sets Out” (2023-03-06). At Datunshan and Jiaoxi, Yilan, the episode follows exploration teams from the Ministry of Economic Affairs’ Geological Survey and Mining Management Agency, Taipower’s Research Institute, and ITRI to see how “geothermal mushrooms” are pieced together from three clues: seismic waves, rock samples, and well temperature gradients. Concrete answers to how much heat lies beneath Taiwan begin here._

PTS _Our Island official channel: Episode 1196, “Geothermal Power in Indigenous Communities” (2023-03-13). The episode presents the two layers of “energy transition vs. local justice” at the same time: in Lizé, Yilan; Hongye, Taitung; and Zhonglun, Chiayi, it asks which line of community negotiation geothermal development enters. Technical problems are one matter; whether a project can enter socially is another matter entirely._

Marine Energy: The Kuroshio’s 9.4 GW at the Trial Stage

After the underground comes the sea.

The Kuroshio current off eastern Taiwan is one of the strongest ocean currents in the world. It flows north year-round at 1.5-2.5 meters per second and is about 100 kilometers wide. In theory, it is an inexhaustible river of energy. In 2021, Academia Sinica completed sea testing of a 100 kW demonstration unit, a milestone in Taiwan’s development of marine energy¹⁹.

PanSci cites an Academia Sinica estimate: “The renewable energy potential of the waters around Taiwan is enormous. The theoretical potential of marine energy, including ocean current energy, wave energy, and ocean thermal energy, is estimated at 9.4 GW. The Kuroshio flows past Taiwan’s east coast and is the most promising source of ocean current energy”¹⁹.

Another direction is OTEC, ocean thermal energy conversion: using the temperature difference between warm surface water (25-28°C) and cold deep water (5°C) to drive generators. The waters off eastern Taiwan have a steep depth gradient and are considered an ideal location for OTEC. “The waters off eastern Taiwan have a large depth gradient and are theoretically an ideal location for developing OTEC, but the technology remains at the experimental stage”²⁰.

But marine energy encounters physical limits even earlier than geothermal: the durability of ocean engineering. Typhoons, salt corrosion, biofouling, and deep-water maintenance are each century-scale engineering challenges. Internationally, no commercial OTEC power plant is operating to date; the leading global case for Kuroshio power generation is a 100 kW demonstration in Okinawa, Japan. Taiwan’s 100 kW test is only the beginning. International experience suggests that going from here to commercialization will take 15-20 years.

Fourth-Generation Nuclear and SMRs: Bill Gates’s Bet

If Taiwan turns back toward nuclear power, will fourth-generation nuclear be the answer?

PanSci reports: “The biggest difference between the Natrium reactor and a traditional nuclear power plant lies in its coolant. Traditional nuclear reactors use water as a coolant, while Natrium uses liquid metal sodium. Sodium has a high boiling point, can operate at higher temperatures and improve reaction efficiency; sodium’s thermal conductivity is 100 times that of water”²¹.

This is the sodium-cooled fast reactor promoted by TerraPower, the company founded by Bill Gates. In April 2026, TerraPower’s Natrium project officially broke ground in Kemmerer, Wyoming, with completion expected in 2030²², a year behind the original schedule but still a key milestone in the commercialization of fourth-generation nuclear power.

The selling point of fourth-generation nuclear is the “small modular reactor” (SMR): reducing generation capacity from the traditional 1,000 MW scale to 100-300 MW, with factory prefabrication and on-site assembly that, in theory, can lower costs and shorten construction time.

But the physical limits remain. PanSci identifies two key risks:

“Fast neutron reactors need to use highly enriched uranium fuel, and breeding reactions generate plutonium-239, an important material for manufacturing nuclear weapons. Therefore, how to manage nuclear materials and prevent nuclear proliferation becomes a difficult problem fast neutron reactors must confront”²³.

“The construction of the Natrium reactor marks a major advance in fourth-generation nuclear power plant technology, but its development also comes with major challenges”²⁴. Liquid sodium reacts violently with water and is flammable; reactor operation and maintenance impose extremely high demands on materials science, and there is still no safety data from large-scale commercial operation.

Does Taiwan have an SMR plan? At present, there is no official plan. Even if assessment began now, international experience suggests that site selection, environmental impact assessment, safety review, and commercial operation would take 15-20 years. In other words, fourth-generation nuclear is not the answer to net zero by 2050; even in the most optimistic scenario, it would not come online until 2045-2050.

📝 Curator’s note: In international public discourse, fourth-generation nuclear is often treated as “the nuclear power of the future,” and therefore turned into “a good reason to delay the energy transition now”: if better technology will exist 15 years from now, why hurry today? This is the most dangerous confusion in a question of physical limits. The engineering bottleneck for renewable energy is “not enough has been built yet.” The bottleneck for fourth-generation nuclear is “not enough safety and nonproliferation data from commercial operation has accumulated.” The two timelines cannot substitute for each other. If Taiwan misses the renewable energy construction window before 2030, SMRs in 2045 will not rescue the climate ledger.

Offshore Wind: The Asian-Leading Piece of the Puzzle

Return the frame to what has already happened.

The Formosa 1 offshore wind farm off Miaoli entered commercial operation in 2019 and was Taiwan’s first large-scale offshore wind farm. Image: Ministry of Economic Affairs, Republic of China, Attribution, via Wikimedia Commons

The Taiwan Strait is one of the best wind fields in the world. “Because of topography, the Taiwan Strait forms a ‘channeling effect,’ making wind speeds in the strait far higher than in surrounding waters and making Taiwan one of the world’s most promising locations for offshore wind development”²⁵. In winter, the northeast monsoon is squeezed into the strait by the Central Mountain Range and the hills of Fujian, producing average wind speeds of 10-12 meters per second. This geographic fact has made offshore wind the central wager of the energy transition.

From almost zero in 2016 to about 4.5 GW of cumulative installed capacity by March 2026³, Taiwan’s offshore wind expansion has ranked among the fastest in Asia. Denmark’s Ørsted completed 920 MW of capacity in the Greater Changhua 2b and 4 offshore wind farms off Changhua²⁶. The third-phase zonal development launched in 2026 allocates 3.6 GW in this round, with completion and grid connection targeted for 2030 to 2031³.

The government’s blueprint is even larger: 13 GW of offshore wind by 2030, and a challenge target of 55 GW by 2050.

But turbines at sea bring not only electricity, but also conflict. In February 2022, more than 100 Changhua fishers traveled north to protest outside the Executive Yuan, accusing the government of “destroying fishers” for the sake of wind power²⁷. Navigation exclusion zones around offshore wind farms blocked waters where they had worked for generations. In May 2025, a court ruled that the navigation restrictions were illegal, the first time a Taiwanese court challenged the spatial governance of offshore wind²⁸.

Solar photovoltaics have taken another path. In 2024, solar PV installed capacity reached 14,281 MW, accounting for 68% of total renewable energy capacity, with generation of 14.9 billion kWh²⁹. Rooftop, ground-mounted, floating, and agrivoltaic installation models have made solar power the main force among renewables. But agrivoltaic policy has prompted doubts about “fake farming, real power generation,” forcing the Ministry of Agriculture to tighten inspections. In a country with only 790,000 hectares of arable land, every plot’s use is a political question.

Rooftop solar panels at Xihu Service Area. Taiwan’s solar PV installed capacity reached 14,281 MW in 2024, accounting for 68% of renewable energy. Image: lienyuan lee, CC BY 3.0, via Wikimedia Commons

Wind and solar are the fastest-moving pieces in Taiwan’s energy transition, but they are inherently intermittent: when the sun sets, power stops; when the wind stops, power stops. This is also the argument nuclear advocates use most often in discussions of extending Maanshan: “Renewable energy is unstable and requires baseload.” The question returns to the geothermal section: baseload renewable energy is not being built fast enough. The 27-fold gap to the target is the political support behind the 2028 restart timeline for Maanshan.

That Afternoon on May 13

At 2:37 p.m. on May 13, 2021, at the Lubei extra-high-voltage substation of the Hsinta Power Plant in Kaohsiung, an operator opened switch No. 3541. He was supposed to open No. 3542³⁰.

This human error triggered a busbar ground fault. Four generating units tripped, instantly removing 2.2 GW of generation capacity. Starting at 3 p.m., Taiwan implemented six rounds of rolling regional power rationing, 50 minutes per round, affecting about four million households. Making matters worse, solar generation declined as the sun set, while drought had reduced hydropower output. Coal-fired units came back online at 7 p.m., and power was fully restored only at 8 p.m.

Four days later, on May 17, Hsinta Unit 1 failed again, bringing a second round of outages. The two incidents together affected more than 5.62 million households³⁰.

The May 13 and May 17 outages exposed how fragile a power system in transition can be, and the problem went far beyond the level of human error. The government’s solution is energy storage: 1.5 GW of battery storage planned for 2025, expanding to 8.6 GW by 2030. But storage costs remain high, and the technology is still maturing.

This is the most honest side of the energy transition: the old system is no longer enough, and the new system is not yet ready. Whether the Maanshan life-extension referendum passed or failed could not change that reality. It could only delay or accelerate the moment of confronting it.

Giving Carbon a Price

On August 7, 2023, the Taiwan Carbon Solution Exchange was inaugurated in Kaohsiung’s Asia New Bay Area, with initial paid-in capital of NT$1 billion and planned capital of NT$1.5 billion; the Taiwan Stock Exchange invested NT$600 million, and the National Development Fund invested NT$400 million³¹. On December 22 of the same year, the first batch of international carbon credit transactions was completed: 45 companies purchased about 88,500 metric tons of CO₂ equivalent in international carbon credits for more than US$800,000³¹.

In 2025, Taiwan’s domestic carbon fee system formally came into force, marking the country’s “first year of carbon pricing”³². Formosa Plastics’ energy-efficiency improvement project was listed at NT$3,000 per metric ton, while Hanbao Farm’s biogas power project was priced at NT$3,000 to NT$4,000. But the market is still searching for its footing: trading volume is low, and companies generally see domestic carbon credit prices as too high.

At the same time, technology giants such as TSMC and Foxconn have already moved onto another battlefield. Under the RE100 initiative, these companies have pledged to use 100% renewable energy. TSMC plans to reach net-zero emissions by 2050. When international customers make green electricity a supply-chain threshold, green power supply becomes an issue of industrial survival, not simply environmental protection.

After the EU Carbon Border Adjustment Mechanism (CBAM) enters its formal system in 2026, it will raise carbon costs and reporting pressure for exports to the EU of high-carbon products such as steel, cement, aluminum, fertilizers, electricity, and hydrogen³³. Taiwan’s manufacturing sector is dominated by energy-intensive industries; steel, petrochemicals, cement, and paper together account for 60% of industrial emissions. This is another physical limit, a timetable imposed on Taiwan by the structure of international trade.

In his 2024 National Day address, Lai Ching-te announced the launch of a “second energy transition,” covering three directions: diversified green energy, deep energy conservation, and advanced energy storage³⁴. But in 2025, the share of renewable energy remained clearly below the original 20% target; depending on statistical definitions, it stood at roughly 12.7% to 13.1%³⁵. The Ministry of Economic Affairs has shifted to saying that the share is expected to reach 20% starting in November 2026, and about 30% by 2030.

Algal Reefs, the Tao People, and Meinong: Fault Lines of Environmental Justice

Every energy pathway has its opponents, and every opponent has a history.

Taoyuan’s algal reefs. The 2021 “Cherish the Algal Reefs” referendum, Referendum No. 20, opposed Taipower’s construction of the third liquefied natural gas receiving terminal on the Datan coast, with the goal of protecting the world’s largest columnar algal reef formation. The referendum failed, and the compromise plan for the terminal went ahead: the port area was pushed farther offshore to avoid high-density algal reef zones. Algal reef scholars still considered the environmental impact assessment insufficient, but the Environmental Impact Assessment Committee approved the review in 2023. The controversy has not subsided. This is the physical / ecological intersection where “to reduce carbon, natural gas must be built; to build natural gas, the algal reefs must be disturbed.”

The Tao people of Lanyu. From 1982 to 2026, 44 years of nuclear waste storage history forms the longest-standing wound in Taiwan’s environmental justice history. Tao people were still protesting missed relocation deadlines in 2024; in May of the same year, the Atomic Energy Council announced that it would require Taipower to complete relocation by 2029. But where it will be moved remains unanswered⁴.

Meinong’s anti-dam movement. In the 1990s, the movement against the Meinong Dam mobilized resistance through the “Meinong Yellow Butterfly Festival” and “Hakka spirit,” ultimately forcing the dam project back. It is a classic case of community-based environmental movements in Taiwan. Reading Meinong today, one finds that its spirit still influences other energy battlefields: whenever a wind turbine, solar panel, or transmission line enters a local community, it will encounter the response, “We oppose bearing the costs of transition ourselves; we do not oppose the energy transition itself.”

📝 Curator’s note: Common environmental justice discussions often stop at “balancing development and environmental protection,” but that framing flattens the problem. The real commonality among Lanyu, the algal reefs, and Meinong is that they are all aftereffects of decisions made after the 1980s, paid for by social movements from the 1990s to the 2020s. Before 2050, the energy transition will generate many new “Lanyus” and “algal reefs”: Changhua fishers affected by offshore wind, Yilan Indigenous communities affected by geothermal development, and Tainan salt flats affected by solar power. The real question is whether Taiwan can avoid repeating the decision-making model of 1982.

For more detailed historical context on environmental justice, see History of Taiwan’s Environmental Movement and Taiwan’s Marine Pollution Governance and Conservation Challenges.

NT$9 Trillion and Physical Limits

Only when all energy sources are placed in the same table do the differences among physical limits emerge.

Energy source	Taiwan’s theoretical potential	2025 status	Government target / timeline	Main physical limit
Offshore wind	Among the world’s best	4.5 GW	13 GW by 2030; 55 GW by 2050	Marine engineering / fishery conflicts
Solar PV	Rooftop + agrivoltaics	14.3 GW	31 GW by 2030	Land acquisition / intermittency
Geothermal (shallow)	≤ 1,000 MW	7.4 MW	200 MW by 2030; 6 GW by 2050	Subsurface uncertainty / financing
Geothermal (deep EGS)	33,640 MW (theoretical)	Laboratory stage	2040+	EGS technology not commercialized
Marine energy	9.4 GW (theoretical)	100 kW trial	2030+	Durability of ocean engineering
Hydrogen (green hydrogen)	Requires large amounts of renewable electricity	Hsinta Power Plant trial	2030+	Electrolysis cost / leakage GWP
Maanshan life extension	1,902 MW	Shut down in 2025	Restart as early as 2028	Nuclear waste / nuclear safety review
Fourth-generation nuclear SMRs	No local plan	U.S. trial operation in 2030	2045+	Sodium-cooling safety / nuclear proliferation

This table answers one question: Can Taiwan reach net zero by 2050 without relying on nuclear power?

Technically, yes. The National Development Council’s roadmap lists 12 key strategies and estimates NT$9 trillion in investment³⁶. But this requires offshore wind, solar PV, geothermal, marine energy, hydrogen, and energy storage to all meet their respective targets at the same time. Right now, geothermal is 27 times short, marine energy is still at the kW level, hydrogen is still in trials, and storage remains expensive.

Every physical limit is time.

In Hsu Huang-hsiung’s models, Taiwan after 2060 has no winter³⁷. In coastal risk assessments, western low-lying areas face higher pressure from sea-level rise and storm surges³⁸. From 1911 to 2020, Taiwan’s annual average temperature has already risen 1.6°C, almost one and a half times the global average of the same period, 1.1°C³⁷.

A Country Warming One and a Half Times Faster

In the summer of 2017, Hsu Huang-hsiung of Academia Sinica’s Research Center for Environmental Changes looked at the data on his screen and produced a forecast that even his colleagues were reluctant to say publicly: if emissions trends remained unchanged, winter in Taiwan might disappear completely after 2060³⁷. Winter days would fall to zero, and summer would stretch to seven months.

This is not science fiction. The number of days above 35°C in Taipei rose from three days a year in the 1960s to 15 days in the past decade³⁹. Southern Taiwan is worse: Tainan and Kaohsiung already have more than 30 high-temperature days per year.

In the same building, Wang Chung-ho of the Institute of Earth Sciences was calculating another set of numbers. His conclusion was equally disturbing: the rate of sea-level rise around Taiwan is twice the global average³⁸. Multiple simulations indicate that sea-level rise and storm surges will expose Taiwan’s western low-lying coasts to higher flood risk; among the six special municipalities, New Taipei, Tainan, and Kaohsiung are of particular concern in terms of exposed population and land area.

The temperament of rain has also changed. Taiwan’s total rainfall has not significantly declined, but rain no longer falls when it should, and when it falls, it pours. Spring rainfall has dropped sharply, and the dry season has become drier. In 2021, Taiwan suffered its worst drought in 56 years. Reservoir levels hit historic lows, and TSMC at one point sent water trucks to supply its factories⁴⁰. In May of the same year, two major blackouts struck Taiwan in succession.

The number of days with more than 200 millimeters of rain in a single day rose from an annual average of five days in the 1960s to eight days in recent years. In 2009, Typhoon Morakot set a cumulative rainfall record of 2,884 millimeters at Alishan⁴¹, with three days of rain equivalent to a full year of rainfall in Taipei. During that disaster, Xiaolin Village in Jiaxian, Kaohsiung, was buried before dawn by debris from the collapse of Xianfu Mountain, killing 491 people⁴².

“Each chair represents one family member.” Survivor Wang Min-liang later said this to visitors at Xiaolin Memorial Park. He founded the Sunlight Xiaolin Community and led the tribe’s Taivoan song and dance troupe on performances across Taiwan. (From PTS’s Our Island)

The 2024 National Climate Change Science Report, led by Hsu Huang-hsiung, states that extreme rainfall events that currently occur once every 50 years may occur once every 10 years in the future⁴³. Yunlin, Tainan, and Keelung are the areas with the highest coastal flood risk.

For a country of 23 million people, Taiwan’s carbon emissions are disproportionately large: in terms of fossil-fuel CO₂ emissions, annual emissions are about 280 million metric tons, about 11.7 metric tons per person, placing Taiwan among the world’s higher emitters; depending on database and statistical definitions, its ranking is around the global top 20-plus⁴⁴. Emissions are highly concentrated in energy use and power supply, with the energy sector accounting for the largest share; the power generation mix remains the core of decarbonization pressure. The root of the problem lies in electricity generation: in Taiwan’s 2024 generation mix, gas accounted for about 42.4%, coal about 39.3%, gas surpassed coal for the first time, renewables about 11.6%, and nuclear about 4.2%³⁵. This is still a highly fossil-fuel-dependent energy system, and Taiwan imports 98% of its energy. Energy security and the climate crisis are the same question.

Democracy and Physics in Parallel

On the evening of August 23, 2025, the Maanshan referendum pushed every contradiction in this question onto the vote-counting screen.

74% in favor, 29.53% turnout, threshold not met, Taipower’s March 2026 submission, restart as early as 2028. At the same time: 97,672 barrels on Lanyu, 50 years for Finland’s Onkalo, geothermal power 27 times short, marine energy still at 100 kW, fourth-generation nuclear not until 2045. Every number asks: Can the speed of democracy keep up with the speed of physics?

Democratic timeline	Physical timeline
2025/08/23 referendum vote count	Lanyu opened in 1982 and may still be storing waste in 2057
2025/08/25 three-principles press conference	Nuclear waste isolation for 100,000 years
2026/03/27 Taipower submission	Finland spent 50 years on a final repository
2028 earliest possible restart	Geothermal power 27 times short
2050 net-zero target	Marine energy still at the 100 kW trial stage

No one knows whether NT$9 trillion can buy a different future. But we are already beginning to see the consequences of not spending that money: Hsu Huang-hsiung’s 2060 without winter, Morakot’s 2,884 millimeters, the rolling rationing of May 13, the rupture of the algal reef referendum, and Lanyu’s 44-year wait.

PanSci reports, citing industry consensus, that “the world’s fastest-moving final repository is Finland’s Onkalo project, which received a trial operation permit in August 2024. The project has been planned since the 1970s and took nearly half a century to reach the trial operation stage”⁹. Taiwan has not even chosen a site for its final repository. Even if Maanshan restarts in 2028, every new fuel rod produced during that restart period will still need somewhere to go.

Lanyu’s 97,672 barrels will not disappear because a referendum passes or fails. They are there now, will very likely still be there in 2029, and will still be there in 2057 if relocation misses its deadline again.

✦ On August 23, 2025, the referendum failed. On March 27, 2026, Taipower submitted the application anyway. Between these two dates, the physical limits did not change once. What changed was whether we are willing to acknowledge that Taiwan, with 98% of its energy imported, is lining up to face every physical limit that no one wants to face.

---

Further Reading:

• Taiwan and the Nuclear Energy Debate — This article addresses energy and physical limits; that article addresses the nuclear debate itself: forty years of anti-nuclear and pro-nuclear arguments, three referendums, and the social struggle over nuclear waste on Lanyu
• History of Taiwan’s Environmental Movement — From anti-nuclear activism to anti-air-pollution movements, how the Tao people of Lanyu, Meinong’s anti-dam movement, and the algal reef referendum shaped today’s energy politics
• Taiwan’s Marine Pollution Governance and Conservation Challenges — 80% coral bleaching at the Maanshan Nuclear Power Plant outlet, marine debris, and the ecological intersection of offshore wind
• Taiwan’s Hot Springs and Geothermal Energy — From the failure of Qingshui Geothermal in 1981 to its 2024 restart, how 30 years of geothermal silence formed
• Taiwan’s Environmental Justice and NIMBY Controversies — Lanyu, the algal reefs, and Meinong: the distributive politics of energy transition costs
• Taiwan’s Industrial Transformation and Upgrading — From energy-intensive manufacturing to green energy industries, TSMC’s RE100, CBAM, and the energy ledger of Taiwan’s “sacred mountain protecting the nation”
• Taiwan’s Agricultural Modernization — The agricultural transformation pressure and land-use conflicts behind agrivoltaics
• Plum Rain — Local observations of climate change in “spring rain not arriving, plum rain concentrating”

References

Image Sources

• Exterior of the Maanshan Nuclear Power Plant in Hengchun, Pingtung (hero): Maanshan Nuclear Power Plant, Nan Wan (photograph: M. Weitzel, Wikimedia Commons, CC BY-SA 3.0)
• Onkalo underground repository in Olkiluoto, Finland: Onkalo spent nuclear fuel repository entrance (photograph: Posiva Oy, Wikimedia Commons, CC BY-SA 4.0)
• Formosa 1 offshore wind farm off Miaoli: Hai Long offshore wind farm (Wikimedia Commons, CC BY-SA 4.0)
• Rooftop solar panels at a freeway service area: Xihu Service Area solar panels (Wikimedia Commons, CC BY-SA 3.0)

1. Central Election Commission: Announcement of Results for the August 23, 2025 National Referendum — ; Central News Agency: Maanshan Life-Extension Referendum Gets 4.34 Million Yes Votes but Fails to Meet One-Quarter Threshold; Taipower: Explanation of Submission of Maanshan Restart Plan to Nuclear Safety Commission (2026/03/27) — August 23, 2025 Maanshan life-extension referendum: 4,342,206 yes votes (74.17%), 1,511,693 no votes, turnout of 29.53%, failing to meet the Referendum Act threshold of one-quarter of all eligible voters (5,000,523 votes), so the referendum did not pass. Taipower submitted an application for the Maanshan restart plan to the Nuclear Safety Commission on March 27, 2026; the safety inspection schedule is expected to take about 18 months, with restart possible as early as 2028.↩
2. Central News Agency: Lai Ching-te’s Remarks After Maanshan Referendum Propose Three Principles: Nuclear Safety, Nuclear Waste, Social Consensus — On August 25, 2025, President Lai Ching-te issued a formal response to the result of the Maanshan life-extension referendum, proposing three principles for any future restart of nuclear power: nuclear safety without concern, a solution for nuclear waste, and social consensus, and instructing the Ministry of Economic Affairs and the Nuclear Safety Commission to initiate safety inspection procedures and assessment.↩
3. Bureau of Energy, Ministry of Economic Affairs: Announcement Launching the Phase 3 Offshore Wind Zonal Development Selection Mechanism — The Ministry of Economic Affairs announced on March 27, 2026, that as of March 26, 2026, Taiwan’s cumulative offshore wind installed capacity was about 4.5 GW; the third-phase allocation capacity is 3.6 GW, with completion and grid connection targeted for 2030 to 2031.↩
4. Atomic Energy Council: Announcement of Stored Volume at Lanyu Storage Site (2024) — Official Atomic Energy Council announcement: as of 2024, the Lanyu low-level radioactive waste storage site held a cumulative 97,672 barrels. Since opening in 1982, it has undergone repeated missed relocation promises in 1996, 2002, 2016, 2019, and 2023; the Atomic Energy Council requires Taipower to complete relocation by 2029. Content Curation Partner per MOU 2026-05-05.↩
5. Bureau of Energy, Ministry of Economic Affairs: Geothermal Power Targets and Commercial Operating Capacity (2025) — Government geothermal power policy targets: 200 MW by 2030 and 6 GW (6,000 MW) by 2050; as of the end of 2025, Taiwan’s commercially operating geothermal capacity was about 7.4 MW, mainly Qingshui Geothermal’s 4.2 MW in Yilan and some small units, about 27 times short of the 2030 target.↩
6. Wikipedia: Lanyu Storage Site — Before the Lanyu storage site opened in 1982, Taipower told Tao residents that it was building a “fish cannery,” without fully informing them of the nuclear waste storage nature of the site; in 1988, Tao people launched the first “driving out evil spirits” protest, becoming the starting point of Taiwan’s Indigenous environmental movement.↩
7. PanSci: After Kuosheng Exits, Nuclear Waste Still Has to Stay for 20 Years — Content Curation Partner per MOU 2026-05-05. The Kuosheng Nuclear Power Plant was formally decommissioned at the end of 2023, but spent nuclear fuel rods remain hot and highly radioactive after reactor decommissioning and must be cooled in on-site spent fuel pools for at least five years before they can possibly be moved out; Daren Township, Taitung County is a candidate site for a final repository for low-level nuclear waste, but the siting process is stuck under local political resistance.↩
8. PanSci: What Is the Real Problem with Extending Nuclear Power? — Content Curation Partner per MOU 2026-05-05. Land-use issues for dry storage facilities at nuclear power plants have been stuck for more than 11 years. The New Taipei City government has refused to approve dry storage facilities, leaving spent fuel rods from the Jinshan and Kuosheng plants still stored in on-site fuel pools, beyond their original designed capacity; the greatest obstacle to extending nuclear power is where spent nuclear fuel will go.↩
9. PanSci: Since Nuclear Waste Has Nowhere to Go, Are There Other Methods? — Content Curation Partner per MOU 2026-05-05. The world’s fastest-moving final repository is Finland’s Onkalo project, which received a trial operation permit in August 2024 and has taken nearly half a century from planning in the 1970s to the present; final repositories must isolate waste for more than 100,000 years, a timescale far exceeding the existence of human civilization.↩
10. Posiva Oy: Onkalo Final Repository Design Overview — Official explanation from Posiva, the Finnish company operating the Onkalo repository. Its design goal is to isolate high-level radioactive nuclear waste for at least 100,000 years, with a multi-barrier system (copper canister + bentonite + granite bedrock) and a long-term memory warning system design.↩
11. Daren Township Office, Taitung County: Low-Level Radioactive Waste Final Repository Issue — Daren Township in Taitung County is one of two candidate sites for a final repository for low-level nuclear waste, the other being Wuqiu, Kinmen County. Local opinion is divided, Indigenous communities strongly oppose it, and a siting referendum has not yet been successfully held.↩
12. PanSci: Feasibility Analysis of Space Disposal for Nuclear Waste — Content Curation Partner per MOU 2026-05-05. Space disposal of nuclear waste is physically feasible but would require extremely stable and insured rockets. If a launch failed, radiation contamination of Earth would be difficult to estimate; with existing rocket failure rates, about one in every 100 launches carries risk, which does not meet practical engineering needs.↩
13. PanSci: Can Improved Natural Gas Power Technology Avoid Producing Carbon Dioxide? Gray Hydrogen, Blue Hydrogen, Green Hydrogen — Content Curation Partner per MOU 2026-05-05. Hydrogen color codes correspond to different production methods: gray hydrogen (natural gas SMR emitting CO₂), blue hydrogen (gray hydrogen + CCS), green hydrogen (renewable-electricity water electrolysis), and turquoise hydrogen (methane pyrolysis producing solid carbon without CO₂ emissions); Taipower and Academia Sinica have collaborated on decarbonized hydrogen combustion tests at Hsinta Power Plant.↩
14. PanSci: Elon Musk Dismisses It; Bill Gates Treasures It! Hydrogen Energy — Content Curation Partner per MOU 2026-05-05. In addition to green hydrogen, the emerging white / gold hydrogen refers to naturally occurring underground hydrogen; the USGS estimates reserves may reach tens of billions of tons. But hydrogen’s GWP100 is 11.6 times that of carbon dioxide, and leakage would worsen warming; the academic community still debates a GWP range of 7-37.↩
15. USGS: Geological Hydrogen — A New Energy Frontier (2023) — The U.S. Geological Survey’s 2023 geological hydrogen research report estimated that global underground natural hydrogen reserves may reach tens of billions of tons, enough to provide humanity with hundreds of years of energy needs; France and Mali already have commercial exploration cases, while Taiwan has an active plate boundary but currently no exploration plan.↩
16. Central News Agency: Yilan Tuchang Geothermal Power Plant Breaks Ground, Expected to Start in 2026 — Construction began in 2024 on the 5.4 MW Tuchang geothermal power plant in Yilan County, expected to start in early 2026. It is Taiwan’s second commercially operating geothermal power plant at MW scale; as of the end of 2025, Taiwan’s commercially operating geothermal capacity was about 7.4 MW, including Qingshui Geothermal’s 4.2 MW and other small units.↩
17. PanSci: Is Geothermal Power Development Feasible in Taiwan? (Part 1) — Content Curation Partner per MOU 2026-05-05. According to National Taiwan University research, deep geothermal resources below five kilometers have potential generation capacity of 33,640 MW, equivalent to about 12 Lungmen Nuclear Power Plants, but development requires EGS, Enhanced Geothermal System technology, which remains in the R&D stage; shallow geothermal potential within three kilometers is estimated at no more than 1,000 MW.↩
18. PanSci: Geothermal Advantages and Applications in Taiwan — Content Curation Partner per MOU 2026-05-05. Geothermal power is not affected by weather and is a baseload power source that generates steadily 24 hours a day, giving it unique value in the energy mix; but subsurface uncertainty causes financing difficulty, the fundamental bottleneck delaying Taiwan’s geothermal development.↩
19. PanSci: The Higher the “Sacred Mountain Protecting the Nation,” the Greater the Power Pressure: Is Taiwan’s Marine Energy the Solution? — Content Curation Partner per MOU 2026-05-05. Theoretical potential for marine energy around Taiwan, including ocean currents, waves, and thermal gradients, reaches 9.4 GW; the Kuroshio flowing past Taiwan’s east coast is the most promising source of ocean current energy, and Academia Sinica completed testing of a 100 kW demonstration unit in 2021.↩
20. PanSci: The Taiwan Possibility for Ocean Thermal Energy Conversion (OTEC) — Content Curation Partner per MOU 2026-05-05. OTEC uses the temperature difference between warm surface water (25-28°C) and cold deep water (5°C) to generate electricity. The waters off eastern Taiwan have a large depth gradient and are theoretically ideal, but the technology remains experimental, with no commercial power plants operating globally.↩
21. PanSci: Bill Gates’s Fourth-Generation Nuclear Power Plant Finally Begins Construction — Content Curation Partner per MOU 2026-05-05. The biggest difference between the Natrium reactor and a traditional nuclear power plant is its coolant: traditional reactors use water, while Natrium uses liquid metal sodium; sodium has a high boiling point, can operate at higher temperatures to improve reaction efficiency, and has 100 times the thermal conductivity of water.↩
22. TechOrange: TerraPower Natrium Project Breaks Ground in Wyoming in 2026 — TerraPower’s Natrium fourth-generation nuclear power plant project officially broke ground in Kemmerer, Wyoming, in April 2026, slightly delayed from the original plan, with completion expected in 2030. It is a key global milestone for commercialization of fourth-generation sodium-cooled fast reactors.↩
23. PanSci: Nuclear Proliferation Risks of Fourth-Generation Nuclear Power — Content Curation Partner per MOU 2026-05-05. Fast neutron reactors need highly enriched uranium fuel, while breeding reactions produce plutonium-239, an important material for manufacturing nuclear weapons; how to manage nuclear materials and prevent nuclear proliferation is a difficult problem fast neutron reactors must face.↩
24. PanSci: Safety Challenges of the Natrium Reactor — Content Curation Partner per MOU 2026-05-05. Construction of the Natrium reactor marks progress in fourth-generation nuclear power plant technology, but development comes with major challenges; liquid sodium reacts violently with water and is flammable, reactor operation and maintenance impose extremely high demands on materials science, and large-scale commercial operation safety data remains lacking.↩
25. PanSci: Offshore Wind Turbines Are Expensive and Troublesome to Build, So Why Is Taiwan Still Developing Them Aggressively? — Content Curation Partner per MOU 2026-05-05. Due to topography, the Taiwan Strait forms a “channeling effect,” making wind speeds in the strait far higher than in surrounding waters and making Taiwan one of the world’s most promising locations for offshore wind development.↩
26. PV Magazine: Taiwan solar and offshore wind targets — Report on Ørsted’s completion of 920 MW in the Greater Changhua 2b and 4 offshore wind farms, and Taiwan’s plan to add 8.2 GW of solar PV and offshore wind by the end of 2026.↩
27. Environmental Information Center: Changhua Fishers Protest Offshore Wind (2022) — Report on more than 100 fishers protesting at the Executive Yuan against offshore wind farm navigation exclusion zones blocking waters where generations had worked, with the slogan “destroying fishers.”↩
28. Environmental Information Center: Court Rules Offshore Wind Navigation Restrictions Illegal (2025) — Taiwan’s first court ruling challenging the spatial governance of offshore wind, finding that navigation restrictions infringed on fishers’ rights and causing shock in energy circles.↩
29. Taipower: Renewable Energy Generation Statistics — Official Taiwan Power Company statistics containing annual installed capacity and generation data for each type of renewable energy; in 2024, solar PV installed capacity was 14,281 MW and generation was 14.9 billion kWh.↩
30. Taipower: Preliminary Investigation Results for the May 13 Outage — ; Ministry of Economic Affairs: Review Report on the May 13 and May 17 Outages — Official materials explaining the mistaken operation of switch No. 3541 in the May 13 incident, which caused an instantaneous reduction of about 2.2 GW in supply capacity and affected about four million households, and summarizing the May 17 incident and follow-up review.↩
31. Presidential Office Press Release: Taiwan Carbon Solution Exchange Inaugurated — ; Taiwan Stock Exchange 2023 Annual Report; Anue: Taiwan Carbon Solution Exchange Expected to Launch at End of July, With North-South Division of Operations — The carbon exchange was inaugurated on August 7, 2023; planned capital was NT$1.5 billion, initial paid-in capital NT$1 billion, with NT$600 million invested by the Taiwan Stock Exchange and NT$400 million by the National Development Fund. The TWSE annual report also records that the first batch of international carbon credit transactions totaled 88,520 metric tons CO₂e, involving 27 companies, or 45 when financial holding company subsidiaries are included.↩
32. KPMG Taiwan: Carbon Pricing Trend Analysis (2025) — Analysis of market dynamics after Taiwan’s carbon fee system came into force in 2025, including domestic carbon credit pricing, Formosa Plastics at NT$3,000 per metric ton and Hanbao Farm at NT$3,000-4,000 per metric ton, and the challenge of low trading volume.↩
33. Official EU Carbon Border Adjustment Mechanism Page — CBAM entered its transitional period in October 2023 and will be fully implemented in 2026, covering six major product categories: steel, cement, aluminum, fertilizers, electricity, and hydrogen.↩
34. Reccessary: Taiwan Energy Policy 2025 Outlook — Report on Lai Ching-te’s 2024 National Day address announcing the policy direction of a “second energy transition”: diversified green energy, deep energy conservation, and advanced energy storage.↩
35. Bureau of Energy, Ministry of Economic Affairs Statistics: Power Generation Mix by Fuel Type — ; Ministry of Environment Energy Information Platform: Electricity Structure; Economic Daily News: Ministry of Economic Affairs Says Renewable Energy Share Expected to Reach 20% Starting November 2026 — According to Bureau of Energy statistics, in 2024 gas accounted for about 42.4%, coal about 39.3%, renewables about 11.5% to 11.6%, and nuclear about 4.2%; the Ministry of Environment Energy Information Platform shows renewables at 13.1% of national total generation in 2025, while Taipower system power purchase and generation mix charts commonly use a figure around 12.7%, reflecting different definitions. In May 2025, the Ministry of Economic Affairs said the share is expected to reach 20% starting in November 2026 and about 30% by 2030.↩
36. Presidential Office Press Release: Tsai Ing-wen’s 2021 Earth Day Remarks — Tsai Ing-wen declared for the first time as president that “the 2050 net-zero transition is a goal for the whole world, and it is also Taiwan’s goal,” laying the policy foundation for the National Development Council’s subsequent net-zero roadmap.↩
37. United Daily News Group Vision Project: Interview with Hsu Huang-hsiung — A team led by Hsu Huang-hsiung, distinguished research fellow at Academia Sinica’s Research Center for Environmental Changes, analyzed Taiwan temperature data from 1911 to 2020 and found that Taiwan has warmed 1.6°C over the past century, winter has shortened by nearly half, and under the worst-case scenario the number of winter days may fall to zero after 2060.↩
38. CSRone Sustainability Think Tank: Interview with Wang Chung-ho — Wang Chung-ho, adjunct research fellow at Academia Sinica’s Institute of Earth Sciences, has long tracked sea-level changes around Taiwan and notes that sea-level rise around Taiwan is faster than the global average; this article adopts a more conservative risk description when citing the interview to avoid simplifying different research scenarios into absolute conclusions.↩
39. Central Weather Administration Climate Change Information Platform — Taiwan’s historical climate observation database, containing century-scale temperature, rainfall, and extreme weather event records from weather stations, including statistics on the increasing number of days above 35°C in Taipei.↩
40. BBC Chinese: Taiwan’s Worst Drought in 56 Years (2021) — Report on the 2021 drought in central and southern Taiwan, when reservoir levels fell below 10% and technology manufacturers including TSMC activated emergency water-truck measures.↩
41. National Science and Technology Center for Disaster Reduction: Typhoon Morakot Disaster Record — Official disaster archive recording cumulative rainfall of 2,884 millimeters at Alishan Station during Typhoon Morakot, the highest record in Taiwan’s meteorological observation history.↩
42. The Reporter: Investigation of the Destruction of Xiaolin Village — In-depth investigation into the collapse of Xianfu Mountain at Xiaolin Village and the full context of the 491 deaths, including geological causes and failures in the warning system.↩
43. Environmental Information Center: 2024 National Climate Change Science Report — Report on key findings from the latest science report led by Hsu Huang-hsiung: extreme rainfall events that occur once every 50 years may become once-in-10-year events, and days above 36°C may increase by 75 days.↩
44. Environmental Protection Administration Greenhouse Gas Emissions Statistics — Taiwan’s official greenhouse gas emissions database, containing national emissions inventories over time, sectoral emissions, and per-capita emissions data.↩