Taiwan’s power consumption is to increase faster at an annual compound growth rate of 2.5 percent over the next five years, mainly driven by an eightfold surge in electricity usage by artificial intelligence (AI) applications, the Energy Administration said yesterday.
The latest forecast represented an increase from the 2 percent growth estimated last year for the period from last year through 2029.
The annual growth rate from this year to 2033 is expected to reach 2.8 percent as new data centers for AI applications become operational, the administration said in a report.
Photo: CNA
The increase in demand comes as Taiwan is expected to undergo a strong economic recovery this year, the semiconductor industry expands capacity amid an AI boom and electric vehicles become more popular, the report said.
AI applications including AI-focused data centers would consume about 2.24 gigawatts (GW) of power in 2028, expanding eightfold from 240 megawatts last year, the administration said.
The latest forecast factored in the expected deployment of new data centers, likely from Apple Inc, Nvidia Corp and Amazon Web Services, in addition to existing centers operated by Google, Microsoft Corp and Nvidia.
Photo: Lin Jin-hua, Taipei Times
“Because of AI, Taiwan has to add 12 to 13 percent more capacity during peak load during the 2024-2029 period,” Energy Administration Director-General Yu Cheng-wei (游振偉) told a news conference in Taipei yesterday.
The administration identified this year to 2028 as the critical period when electricity consumption would spike due to rapid uptakes of AI data centers, large-language AI training models and supercomputer centers, which have not been included in the administration’s forecast before.
Peak load demand is expected to expand to about 4.21GW at night from 3.73GW this year, the administration said.
To ensure stable and sufficient supply, Taiwan Power Co (Taipower, 台電) would install more natural-gas generators, enhance electricity grid resilience by revamping electrical substations and expand green energy storage systems, Yu said.
He urged the public to support CPC Corp, Taiwan’s (台灣中油) deployments of new liquefied natural gas (LNG) terminals to regasify LNG before sending it to power stations.
The construction of new LNG terminals was suspended amid environmental concerns.
Taipower projected its reserve margin would improve to 13.3 percent in 2028 from 8.5 percent this year.
That figure is expected to climb to 23.8 percent in 2033, indicating that its electricity capacity would outgrow demand, or capacity adequacy, the state-run utility said.
Asked if the administration would consider restarting or delaying the shutdown of nuclear power plants, Yu said the administration would respect the legislature’s decision on the matter.
Renewable energy is to account for 20 percent of the total power supply in November 2026, rather than the government’s 2025 target, the third time the government has pushed back the goal.
Yu blamed a surge in power consumption for the delays, but said the deployment of renewable energy generation remains on track.
By the end of 2029, green energy is projected to make up about 25 percent of the total power supply, down from 27 percent estimated last year, largely due to increases in total power supply, the report said.
SECOND-RATE: Models distilled from US products do not perform the same as the original and undo measures that ensure the systems are neutral, the US’ cable said The US Department of State has ordered a global push to bring attention to what it said are widespread efforts by Chinese companies, including artificial intelligence (AI) start-up DeepSeek (深度求索), to steal intellectual property from US AI labs, according to a diplomatic cable. The cable, dated Friday and sent to diplomatic and consular posts around the world, instructs diplomatic staff to speak to their foreign counterparts about “concerns over adversaries’ extraction and distillation of US AI models.” Distillation is the process of training smaller AI models using output from larger, more expensive ones to lower the costs of training a powerful new
Micron Technology Inc is a driving force pushing the US Congress to pass legislation that would put new export restrictions on equipment its Chinese competitors use to make their chips, according to people familiar with the matter. A US House of Representatives panel yesterday was to vote on the “MATCH Act,” a bill designed to close gaps in restrictions on chipmaking equipment. It would also pressure foreign companies that sell equipment to Chinese chipmaking facilities to align with export curbs on US companies like Lam Research Corp and Applied Materials Inc. The bill targets facilities operated by China’s ChangXin Memory Technologies Inc
Singapore-based ride-hailing and delivery giant Grab Holdings’ planned acquisition of Foodpanda’s Taiwan operations has yet to enter the formal review stage, as regulators await supplementary documents, the Fair Trade Commission (FTC) said yesterday. Acting FTC Chairman Chen Chih-min (陳志民) told the legislature’s Economics Committee that although Grab submitted its application on March 27, the case has not been officially accepted because required materials remain incomplete. Once the filing is finalized, the FTC would launch a formal probe into the deal, focusing on issues such as cross-shareholding and potential restrictions on market competition, Chen told lawmakers. Grab last month announced that it would acquire
The artificial intelligence (AI) boom has triggered a seismic reshuffling of global equity markets, with Taiwan and South Korea muscling past European nations one by one. With its stock market now valued at nearly US$4.3 trillion, Taiwan surpassed the UK, Europe’s biggest market, earlier this month, data compiled by Bloomberg showed. South Korea is about US$140 billion away from doing the same. The tech-heavy Asian markets have shot past Germany and France in the past seven months. The shift is largely down to massive gains in shares of three companies that provide essential hardware for AI: Taiwan Semiconductor Manufacturing Co (TSMC, 台積電),
RSS