TSMC released its latest Sustainability Report on 20 August. For the first time, the company has set interim absolute emissions-reduction targets for 2030: a 31% reduction in Scope 1 and Scope 2 emissions from a 2025 base year, bringing them back to 2020 levels, and a 15% reduction in Scope 3 emissions. Compared with its previous reliance on interim renewable-energy commitments and a long-term net-zero target, this is the first time TSMC has defined a clear pathway towards net zero.
In May, the Taiwan Climate Action Network (TCAN) published its 2026 TSMC Climate Performance Assessment. The assessment found that TSMC lacked a medium-term absolute emissions-reduction pathway covering Scopes 1–3 and aligned with its 2050 net-zero target. It also highlighted the low share of renewable electricity at the company’s Taiwan fabs, shortcomings in meeting energy-efficiency targets, and limited climate-policy engagement.
The new targets, strengthened Scope 3 accounting, disclosure of the renewable-electricity share at Taiwan fabs, progress in process energy efficiency, and reference to a just transition respond to some of TCAN’s recommendations and deserve recognition.
In his opening message, TSMC Chairman C.C. Wei highlights the International Court of Justice’s landmark advisory opinion on states’ obligations to address climate change:
“This marks the first time in history that climate responsibility has officially risen to the level of a legal obligation, underscoring that the climate crisis is an urgent and existential threat to humanity.”
Wei further states that TSMC takes its responsibilities seriously: the company aligned its net-zero roadmap with Science Based Targets initiative (SBTi) absolute emissions targets this year, is allocating more of its earnings to long-term sustainability and ESG initiatives, and is collaborating with stakeholders to drive collective social progress. If TSMC is to live up to this standard, it must address four major gaps.
1. Renewable-energy commitments remain insufficient, and on-site solar-plus-storage should be integrated into fab expansion
For the first time, TSMC reports that renewable energy accounted for only 11.4% of electricity use at its Taiwan fabs in 2025. This was a significant increase from TCAN’s estimate of 6.2% for the previous year, but it remained below the company-wide average of 20.1%. Semiconductor production at TSMC’s Taiwan fabs therefore remains heavily dependent on a power system dominated by gas- and coal-fired generation.
TSMC states that it has signed 7.3 GW of renewable-energy contracts, which are expected to supply 15.9 TWh annually once all projects become operational. However, it has not disclosed the annual commissioning schedule for these projects or the proportion available to its Taiwan fabs.
Based on Taipower’s forecast for growth in semiconductor electricity demand, TCAN estimates that TSMC’s Taiwan electricity consumption could rise to approximately 40 TWh by 2030. Meeting the company’s RE60 commitment under this expansion scenario would still leave a renewable-electricity shortfall of about 8 TWh. If this gap is not filled with domestic renewable energy, TSMC’s expansion could deepen Taiwan’s reliance on gas-fired power and weaken energy resilience.
Even the most basic form of self-generation shows a striking disparity. TSMC Arizona has installed a 14.6 MWp solar system over its fab parking facility, while the combined on-site solar capacity of all TSMC fabs in Taiwan is only 8.45 MW. TSMC should expand rooftop and parking-lot solar at existing sites and integrate solar PV and energy storage into the design of all new and expanded fabs.
2. Progress at the 3 nm node shows that the five-year energy-efficiency target is achievable
TSMC’s stated target is to “Double energy efficiency after five years of volume production for each process technology.” However, energy efficiency for the 5 nm process had improved by only 0.6 times by the fifth year of volume production, falling short of the target.
The latest report shows that energy efficiency for the 3 nm process was already 0.5 times higher in the second year of volume production, outperforming the original target of a 0.2-times improvement and delivering half of the improvement required to double efficiency within five years. This indicates that the goal is not evidently beyond technical reach.
TSMC should provide Taiwan’s Energy Administration and Taipower with data on capacity, electricity use per unit of production, volume-production learning curves and energy-efficiency scenarios for each process node. These data should also be incorporated into reviews of the company’s energy-use plans.
TSMC should further disclose projected electricity demand under three scenarios: no further efficiency improvement, continuation of existing improvement trends, and achievement of its five-year energy-efficiency target. This would prevent the Ministry of Economic Affairs’ electricity supply-and-demand projections from reflecting only high electricity growth under current technology while overlooking the energy-saving potential embedded in TSMC’s own commitments.
3. Energy transition advocacy cannot hide behind industry associations
InfluenceMap’s LobbyMap has rated TSMC’s climate-policy engagement as “very low,” awarding the company a C grade. In the latest report, TSMC states only that it continues to work with the Taiwan Semiconductor Industry Association (TSIA) and the SEMI Energy Collaborative (EC), together with government agencies, power utilities and industry partners, to review regulations and remove barriers to renewable-energy supply. It does not disclose which policy proposals it advanced, how many resources it committed or what outcomes this engagement produced.
More importantly, publicly available information contains no clear TSMC position on recent renewable-energy controversies and regulatory reforms, including amendments to environmental-impact-assessment rules for renewable-energy projects and unreasonable restrictions imposed by local governments on solar development.
If TSMC is serious about removing supply barriers, it should publicly support renewable-energy targets for 2030 and 2035 that are sufficient to deliver its RE100 commitment. It should also state its positions on grid investment, gas expansion and the risk of fossil-fuel lock-in, and publish a review of whether the climate-policy positions of its industry associations are aligned with the Paris Agreement’s 1.5°C goal.
4. The carbon fee must become more than a passive cost, and TSMC’s Self-determined Reduction Plan should be revised
Taiwan completed its first carbon-fee collection cycle in 2026. The semiconductor industry paid approximately NT$2.2 billion, the highest amount of any sector. Based on publicly available data, TSMC is estimated to be the country’s largest corporate payer. The company has also identified carbon fees and taxes as a transition risk with a medium level of impact.
However, TSMC’s report focuses mainly on estimating that, if the carbon fee rises to NT$1,200–1,800 per metric ton by 2030, the financial impact would be equivalent to approximately 1% of revenue. It does not explain how carbon pricing would change the company’s decarbonisation investment, equipment procurement or technology choices. An impact of “about 1% of revenue” indicates only that the financial burden may be limited; it does not demonstrate that the carbon fee has become an effective incentive for emissions reduction.
There is also a gap between TSMC’s public voluntary commitments and its legally binding Self-determined Reduction Plan under Taiwan’s carbon-fee regime. Now that TSMC has adopted new 2030 targets, it should proactively apply to revise the plan. The updated plan should incorporate annual milestones for 2027–2030; renewable-electricity procurement for Taiwan; on-site solar and energy storage deployment; process energy-efficiency improvements; and associated investment.
This would prevent TSMC from operating a dual-track system—one set of commitments for investors and another for obtaining a preferential carbon-fee rate from regulators. Incorporating interim emissions-reduction and renewable-electricity commitments into the approved plan would make them legally accountable, trackable and enforceable.
TSMC must complete a credible transition plan before its IFRS S2 disclosure deadline
Under the Financial Supervisory Commission’s timetable, TSMC will be required to apply the IFRS Sustainability Disclosure Standards from the 2026 financial year. By the end of March 2027, it must report the relevant information alongside its annual financial statements, including its climate strategy, any existing transition plans, targets, capital allocation and progress.
TCAN calls on TSMC to strengthen three areas over the next seven months to meet the IFRS S2 climate-related disclosure requirements:
- Present a domestic renewable-energy procurement plan capable of closing the gap to RE60. This should include clear renewable-energy shares and a commitment to “Double energy efficiency after five years of volume production for each process technology” in the environmental-impact-assessment process for new fabs.
- Disclose how carbon fees affect net-zero investment decisions. This should include the implications of different carbon-fee trajectories for decarbonisation by TSMC and its supply chain.
- Publish its positions and results on key climate policies. These should include clear positions on recent amendments to the Energy Management Act, mandatory rooftop solar, targets for the next offshore-wind allocation round, and why a renewable-energy share of at least 30% in Taiwan by 2035 is necessary for TSMC to fulfil its RE60 and emissions-reduction targets.
Appendix 1. Comparison of TCAN’s Climate Performance Assessment and TSMC’s Latest Sustainability Report
| Gaps identified by TCAN | Progress in the latest Sustainability Report | Remaining shortcomings |
|---|---|---|
| No medium-term absolute emissions-reduction pathway aligned with the 2050 net-zero target. | New 2030 targets: Scope 1 and Scope 2 emissions reduced by 31%, and Scope 3 emissions reduced by 15%. | The targets use the peak-emissions year of 2025 as the base year; no annual pathway for 2027–2030 has been disclosed, and the targets have not yet been validated by SBTi. |
| Renewable electricity at Taiwan fabs was only about 6.2%, far below the 100% achieved overseas; the 7.3 GW contracts lacked clear delivery information. | Renewable electricity at Taiwan fabs rose to 11.4%, or 2.94 TWh; once fully operational, the 7.3 GW contracts are expected to supply 15.9 TWh annually. | Some 88.6% of electricity used at Taiwan fabs remains unmatched by renewable energy. Annual commissioning schedules and domestic supply volumes have not been disclosed, and the projected 2030 shortfall remains about 8 TWh. |
| The target to “Double energy efficiency after five years of volume production for each process technology” was not achieved for the 5 nm process, while the Self-determined Reduction Plan focused heavily on LED replacement rather than major electricity-consuming equipment. | Energy efficiency for the 3 nm process was 0.5 times higher in the second year of volume production, outperforming the target of 0.2 times. | Energy savings offset only about 28% of the annual increase in electricity consumption. Process-specific targets, investment figures and quantified emissions-reduction benefits remain undisclosed. |
| Climate-policy engagement was passive, with a LobbyMap rating of C. | The report provides more information on policy exchanges, advocacy platforms and participation in industry associations. | TSMC has not disclosed its positions on renewable energy, gas expansion, grid investment or carbon pricing, nor has it published an alignment review of its industry associations. |
| The legally binding Self-determined Reduction Plan diverged from the company’s public voluntary commitments and lacked complete measures and renewable-energy timelines. | The Sustainability Report presents more ambitious 2030 emissions-reduction, renewable-energy and energy-efficiency measures. | TSMC should revise its Self-determined Reduction Plan to incorporate the new targets and annual measures, and link performance to eligibility for preferential carbon-fee rates so that its commitments carry legal accountability. |
