On 4 June, Taiwan Semiconductor Manufacturing Company (TSMC) will hold its annual shareholders’ meeting. Investors have much to applaud: a commanding position in advanced chips, a surging share price, and aggressive capital expenditure to meet the artificial intelligence boom.
But beneath this success lies a question that should no longer be confined to sustainability reports. Can the world’s most important chipmaker deliver a credible climate transition plan while its electricity demand is rising sharply?
Taiwan Climate Action Network’s 2026 TSMC Climate Performance Assessment Report finds that TSMC has made important commitments: net zero emissions for global operations by 2050, zero emissions growth in 2025, a return to 2020 greenhouse gas levels and RE60 by 2030, RE100 by 2040, and a Science Based Targets initiative-aligned plan in 2026. Yet commitments are not delivered.
The first warning sign is that TSMC’s reduction pathway is drifting away from its net zero pledge. Its 2025 goal requires product-level greenhouse gas intensity to fall 10 per cent from 2020 levels. Yet in 2024, that indicator was 19 per cent higher than in 2020. TSMC’s Self-determined Reduction Plan under Taiwan’s carbon fee framework also appears inconsistent with its 2030 commitment. The company has promised that emissions will return to the 2020 level by 2030. But under the disclosed plan, Taiwan-site Scope 1 and 2 emissions would decline only to 12.47 million tons of CO2e by 2030 — still far above the 9.47 million tons recorded in 2020.
The second warning sign is TSMC’s renewable energy gap. The company has made progress: its renewable energy use rose from 7.6 per cent in 2020 to 20.1 per cent in 2025, while overseas manufacturing sites and offices have achieved 100 per cent renewable energy use for eight consecutive years, but the renewable energy share of its main Taiwan fabs may still be only around 6.2 per cent per cent. RE60 cannot be judged only by global totals. It must be judged by whether renewable energy is actually reaching the fabs that produce the world’s most advanced chips.
The scale of the challenge is large. TSMC has disclosed 7.3 gigawatt s (GW) of renewable energy power purchase agreements. If approximated as 2GW of offshore wind and 5.3GW of solar photovoltaic (PV), that portfolio would generate about 14.04 terawatt-hours (TWh) a year. Compared with TSMC’s 2025 electricity use of about 28.76TWh, that would cover only about 48.8 per cent of demand. By 2030, electricity demand could exceed 42TWh.
RE60 would then require more than 25TWh of renewable energy. Even if the current portfolio is delivered, TSMC could still face a shortfall of more than 10TWh.
The third warning sign is energy efficiency. TSMC has set a goal to double energy efficiency five years after mass production for each process technology. But by the fifth year of 5nm mass production in 2024, energy efficiency had improved by only 0.6 times. Its Self-determined Reduction Plan also relies too heavily on low-additionality measures. Lighting accounts for only about 2.5 per cent of fab electricity use, yet LED replacement represents about 10 per cent of listed measures. Process equipment accounts for roughly half of fab power use, but only 2.6 per cent of measures address it. A credible climate transition plan for a leading-edge semiconductor company cannot be built around lightbulbs.
Some may argue that TSMC’s supply-chain position makes these concerns secondary. The opposite is true. Major customers including Nvidia, AMD, Apple, Microsoft and Google are strengthening Scope 3 and product-carbon-footprint management. AMD has set a 2030 target to reduce supplier carbon intensity by 25 per cent from 2024 levels. If these expectations become procurement criteria, TSMC’s electricity mix and product carbon footprint could affect order allocation and pricing power.
Location and grid carbon intensity can also reshape advanced-chip emissions. Analysis cited in the Taiwan Climate Action Network’s report suggests that if both Intel 18A and TSMC 2nm reached 90 per cent yield, Taiwan-made TSMC 2nm emissions could be 47 per cent higher than Intel 18A produced in Oregon, largely because of Oregon’s lower-carbon electricity mix.
This is also a national security question. A disruption in the Strait of Hormuz would expose Taiwan to unprecedented fossil-fuel supply risks, making further dependence on gas-fired power even more dangerous. If Taiwan cannot secure enough local renewable energy for its most strategic industry, pressure will grow to meet semiconductor demand with additional gas-fired generation. That would deepen dependence on imported fuel, expose Taiwan’s grid to geopolitical and price shocks, and force Taiwan into a trade-off between economic security and energy security. It would also increase risks for the global semiconductor supply chain.
Shareholders should press TSMC to make three commitments.
First, align interim emissions targets with the net-zero pledge. TSMC should publish a year-by-year 2025–2035 pathway for absolute Scope 1, 2 and 3 reductions, with clear accountability and verification mechanisms. It should also explain how the Self-determined Reduction Plan under Taiwan’s carbon fee framework will be brought into line with its 2030 commitment to return emissions to 2020 levels.
Second, close the domestic renewable energy gap. TSMC should move beyond global renewable energy totals and disclose how much renewable electricity will actually be delivered to its Taiwan fabs. The company should provide a year-by-year plan for the 7.3GW Power Purchase Agreement (PPA) portfolio, including grid-connection timelines, expected generation and the share available to Taiwan’s advanced-node production. Only renewable energy generated and delivered in Taiwan can meaningfully reduce fossil-fuel supply risks, ease pressure for new gas-fired power, and strengthen semiconductor supply-chain resilience.
Third, reset its energy-efficiency strategy. TSMC should establish node-level electricity key performance indicators, kWh per mask layer and energy-per-wafer-pass indicators, work more closely with ASML and other key equipment suppliers, and shift its Self-determined Reduction Plan away from low-impact facility upgrades toward process equipment, extreme ultraviolet lithography (EUV) systems and fab utilities.
TSMC does not need weaker growth. It needs a credible climate transition plan equal to its strategic weight. For the company at the heart of the AI age, renewable energy is no longer just a climate metric. It is a condition of competitiveness, supply-chain trust, energy security and Taiwan’s national security.
Chia-Wei Chao, Ph.D., is a researcher and policy advisor specializing in sustainability transitions and industrial ecology. He earned his doctorate in 2013 from the Graduate Institute of Environmental Engineering at National Taiwan University. Since 2022 he has served as Research Director of the Taiwan Climate Action Network — a coalition of five NGOs accelerating Taiwan’s path to net zero — where he leads the research unit in producing high-quality policy analysis on energy transition, industrial decarbonization, climate adaptation, and just transition. Chao also serves on the National Climate Change Committee, and the Carbon Fee Rate Review Committee.
