TSMC legal meeting transcript: This year’s capital expenditure will increase to US$64 billion, and the shortage of high-end chips will continue for several years
On Thursday, TSMC announced a strong second-quarter financial report and comprehensively raised its full-year revenue and capital expenditure outlook: it raised its 2026 capital expenditure forecast from US$52 billion to US$56 billion to US$60 billion to US$64 billion, and estimated that the full-year US dollar-denominated revenue growth rate will be slightly higher than 40%, higher than the previous guidance of more than 30%.
On Thursday, TSMC announced a strong second-quarter financial report and comprehensively raised its full-year revenue and capital expenditure outlook: it raised its 2026 capital expenditure forecast from US$52 billion to US$56 billion to US$60 billion to US$64 billion, and estimated that the full-year US dollar-denominated revenue growth rate will be slightly higher than 40%, higher than the previous guidance of more than 30%. At the TSMC legal conference, TSMC Chairman and CEO Wei Zhejia repeatedly emphasized that the AI market demand is strong, especially the demand for high-end chips, which will remain strong overall in the next three to four years. "The long-term industry trend is extremely resilient, and we are witnessing the formation of a new AI industry." Wei Zhejia emphasized that TSMC has strong confidence in the long-term industry wave of AI, so it continues to increase the scale of capital expenditures. "It is currently judged that the scale of capital expenditures in the next three years will further increase significantly compared with the past three years." Interestingly, Wei Zhejia also bluntly stated at this legal meeting that he envied the 86% gross profit margin of his Korean storage peer (SK Hynix). "If we can do about 68%, I'll be satisfied." Despite this, he still emphasized that TSMC will adopt a win-win pricing strategy and will not significantly raise prices to squeeze customer profits. Several highlights of TSMC’s legal meeting 1. Demand side: AI’s long-term prosperity is determined and the market is structurally differentiated 1. Intelligent AI has spawned a new increase in CPUs. The demand for GPUs and various AI acceleration chips has simultaneously strengthened. The gap between supply and demand for high-end chips is huge, and the shortage cycle will continue until at least 2029-2030. 2. The hot and cold differentiation of mature processes: The production capacity of AI supporting power ICs and image sensors is tight, and the demand for mature consumer chips is weak. 3. There is no need to worry about the customer structure, as new manufacturers continue to emerge in the AI track. 2. Comprehensive increase in capital expenditure and global production expansion 1. Increase the annual capital expenditure range. The overall investment scale in the next three years will be significantly higher than in previous cycles, and funds will mainly be invested in advanced processes. 2. Arizona in the United States has invested hundreds of billions of dollars in additional investment to build multiple 2nm and supporting packaging factories; Taiwan continues to build new high-end production lines and simultaneously deploys Japanese and German specialty process factories. 3. The production capacity has flexible deployment capabilities and can flexibly balance the supply of 7/5/3nm production capacity. 3. Clear technical iteration route and build long-term barriers 1. The research and development progress of 1.4nm (A14) exceeds expectations. Mass production will be launched in 2028, and the derivative process will be launched in 2029. The life cycle is expected to exceed 2nm. 2. Promote the introduction of high-NA EUV prudently, and consider the cost and yield before putting it into large-scale production. 3. There is a short-term shortage of advanced packaging production capacity, and we are optimistic about new cost-cutting technologies such as glass substrates; we do not exclude competing packaging solutions, and external production capacity can release front-end wafer orders. 4. Industry competition: OEM barriers are high, so there is no need to worry about peer impact Although Samsung and Intel have financial and policy support, their advanced process verification and mass production cycles take up to five years. It is extremely difficult for customers to switch suppliers, and it is difficult to divert core high-end orders in the short term. 5. Pricing and Profit Ideas Adopting a win-win pricing strategy will not significantly raise prices to squeeze customer profits; 2nm ramp-up and mass production of new overseas factories will continue to dilute gross profit margins, which are fixed factors that suppress mid- to long-term profits. 6. Shareholder dividends will continue to improve in the long run We continue to steadily increase annual and quarterly cash dividends, and have made it clear that dividends will continue to grow in 2027, and the high production expansion cycle still guarantees shareholder returns. The following is a transcript of the TSMC Dharma Talk (AI-assisted translation, some content has been deleted) Su Zhikai (Investor Relations Director/Host) Good afternoon, everyone, and welcome to TSMC's 2026 second quarter financial results conference and conference call. I am Su Zhikai, Director of Investor Relations at TSMC, and I am also the host of today’s meeting. The process of this meeting is as follows: First, Mr. Huang Renzhao, senior vice president and chief financial officer of TSMC, will review the operating status of the second quarter of 2026 and announce performance guidance for the third quarter of 2026; then Mr. Huang Renzhao and TSMC Chairman and CEO Dr. Wei Zhejia will jointly elaborate on the company's core business views. After the speech session, we will open a question-and-answer session for on-site participants and online telephone participants.
Next, I will hand over the microphone to Mr. Huang Renzhao, Chief Financial Officer of TSMC, who will review the current period's operations and announce next quarter's performance guidance. Huang Renzhao (Senior Vice President, Chief Financial Officer) Thank you Su Zhikai. Good afternoon everyone, thank you all for attending. My report will first sort out the core financial data for the second quarter of 2026, and then announce the performance guidance for the third quarter of 2026. First, the revenue share is broken down by process dimension: the 2-nanometer process contributed 3% of the wafer revenue in this quarter; 3-nanometer, 5-nanometer, and 7-nanometer accounted for 30%, 33%, and 11% respectively; the advanced processes of 7-nanometer and below accounted for 77% of the total wafer revenue. Breaking down the revenue by application platform: high-performance computing (HPC) increased by 20% month-on-month, accounting for 66% of total revenue this quarter; smartphone revenue fell 4% month-on-month, accounting for 22%; Internet of Things increased 4% month-on-month, accounting for 5%; automotive chips increased 15% month-on-month, accounting for 4%; consumer electronics (DCE) increased 5% month-on-month, accounting for 1%. Next, I will introduce the balance sheet to you. As of the end of the second quarter, the company's cash and marketable securities totaled NT$3.5 trillion, equivalent to US$110 billion. Liability side: Current liabilities increased by NT$144 billion month-on-month, mainly due to an increase of NT$58 billion in accounts payable and an increase of NT$48 billion in accrued liabilities and other accounts. In terms of financial turnover indicators: accounts receivable turnover days increased by 3 days month-on-month to 29 days; inventory turnover days increased by 7 days month-on-month to 87 days. The core reason is the ramp-up of 2 nanometer (N2) process production capacity. Cash flow and capital expenditure (CapEx): Cash flow generated from operating activities in the second quarter was NT$783 billion; capital expenditure investment was NT$496 billion; cash dividends distributed in the third quarter of 2025 were NT$156 billion. Overall, cash balances increased by NT$99 billion in the quarter from the previous quarter, with cash reaching NT$3.1 trillion at the end of the quarter. In U.S. dollars, second-quarter capital expenditures totaled $15.7 billion. The above is a complete financial review, and then the performance guidance for this quarter (third quarter) will be announced. Based on the current market operating environment, we estimate that total revenue in the third quarter will range from US$44.6 billion to US$45.8 billion; calculated at the midpoint of the range, it will increase by 12% month-on-month and 37% year-on-year. The exchange rate is assumed to be NT$32 for 1 U.S. dollar. The gross profit margin is expected to be in the range of 65%-67% and the operating profit margin in the third quarter is 56%-58%. Compared with the first quarter, the gross profit margin in the second quarter increased by 1.5 percentage points quarter-on-quarter to 67.7%, slightly exceeding the previous guidance. This was mainly due to cost optimization measures and a slight increase in overall capacity utilization, which was only partially offset by the dilutive effect of gross profit margin caused by the ramp-up of overseas fab production capacity. The median gross profit margin guidance range we just gave for the third quarter was 66%, down 1.7 percentage points from the previous quarter. The core reason is that the large-scale ramp-up of the 2-nanometer process will bring about a gross profit margin dilution of about 3 to 4 percentage points; however, the extremely strong demand for high-end processes, coupled with the continued promotion of cost optimization (including production efficiency improvements and cross-process capacity allocation optimization), will partially offset the downward pressure on gross profit margins. Looking at the second half of 2026, the factors that affect profitability have pros and cons. Let me explain to you here: the large-scale ramp-up of the 2-nanometer process is expected to dilute the gross profit margin by 3 to 4 percentage points in the second half of the year; as the scale of overseas expansion continues to expand, the gross profit margin dilution caused by the ramp-up of mass production of overseas wafer fabs in the next few years is expected to be 2%-3% in the initial stage and will expand to 3%-4% in the later period. Positive aspects: demand for high-end processes continues to be strong; secondly, we continue to rely on mature manufacturing capabilities to increase wafer output scale, and support profitability through optimization of cross-process capacity across the entire factory; finally, exchange rates are uncontrollable external variables and will also have a fluctuating impact on profitability. Next, we will introduce the full-year capital expenditure budget for 2026.
At TSMC, the scale of capital expenditures has always been highly tied to growth opportunities in the next few years. Relying on our leading and differentiated technical strength, the company is able to fully undertake the long-term structural growth needs of the three major industries of 5G, artificial intelligence, and high-performance computing. The current structural demand from customers continues to rise, including the incremental demand brought by the emerging intelligent AI market. Therefore, we have decided to increase our full-year capital expenditure budget for 2026 to US$60 billion to US$64 billion, and continue to increase investment to match customer growth needs. No matter whether the industry is in an up cycle or a down cycle, we will collaborate deeply with equipment suppliers in advance to plan production capacity in advance; just like customers will contact us in advance to plan wafer production capacity. We predict that the capacity expansion plan will not cause equipment supply bottlenecks. About 70%-80% of the full-year capital expenditure in 2026 will be invested in advanced process research and development and production line construction; about 10% will be used for specialty processes; and the remaining 10%-20% will be allocated to advanced packaging, testing, mask manufacturing and other supporting businesses. Even if we maintain high-scale capital expenditures in 2026, we will continue to create profitable growth for shareholders; at the same time, we will insist on steadily increasing cash dividends per share every year and every quarter. The company's total cash dividends in 2025 will be NT$467 billion, a year-on-year increase of 28.6%, and shareholders will receive a total dividend of NT$18 per share throughout the year; the dividend per share will increase to NT$24 in 2026, a year-on-year increase of another 33%; we expect cash dividends per share to continue to grow in 2027. Next I will hand over the microphone to Wei Zhejia, Chairman and CEO. Wei Zhejia (Chairman and CEO) In the long term, consumer and price-sensitive end markets are suppressed by rising component prices and macroeconomic uncertainty, putting demand under pressure. Therefore, we remain cautious when formulating our business plans, while focusing on core business fundamentals and continuing to consolidate our competitive advantages. But even so, the demand for artificial intelligence is still extremely hot. The long-term industrial wave of artificial intelligence continues to generate demand for massive computing power, which directly drives the demand for high-end chips to continue to rise. Our customers and their downstream customers (mainly cloud service providers) continue to send us strong demand signals and optimistic market expectations, so we maintain a high degree of confidence in the long-term industry trend of AI. Relying on differentiated technical strength and extensive customer base, we have now raised our expectations: the year-on-year growth rate of US dollar-denominated revenue in 2026 is slightly more than 40%. Let’s talk about the acceleration of industry demand brought about by intelligent AI. The AI market is iterating at an extremely fast pace. The rise of intelligent AI has once again increased the value of CPUs in AI data centers. In addition to AI acceleration chips, it has further boosted the overall demand for chips. We believe this is a major benefit to TSMC: whether customers use x86, Arm or RISC-V architecture CPUs, almost all of them are TSMC’s cooperative customers. We have deeply collaborated with CPU customers to provide the most advanced manufacturing processes and supporting production capacity to help customers seize opportunities in the intelligent AI market. Next, we will introduce TSMC’s global production capacity expansion strategy. In order to match the long-term structural demand growth of the semiconductor industry, TSMC continues to plan production capacity in depth with customers and their downstream terminals. The lead time for high-end process research and development, customer product introduction, and mass production is extremely long, so we can clearly grasp our customers' product routes and mass production plans for the next few years. This is crucial: the complete cycle from new technology research and development, supporting production capacity construction to large-scale mass production takes more than 5 years. At the internal level, TSMC has established a standardized production capacity assessment system to assess market demand from top to bottom combined with bottom-up. This work continues to be promoted throughout the year. After comprehensive assessment, we increased capital expenditure and expanded production capacity to support the future growth of our customers. Relying on the strong support of leading US customers, US federal, state, and local governments, we hereby officially announce that TSMC will invest an additional US$100 billion in Arizona to build multiple 2-nanometer and below logic wafer production lines, and simultaneously support advanced packaging factories to meet the continued strong order demand of leading US customers in the next few years.
We believe that this additional investment will further improve the domestic semiconductor industry chain in the United States, strengthen the resilience of the supply chain, and at the same time create a large number of high-paying high-tech jobs in the United States. At the same time, in the next few years, we will build 13 new high-end process and advanced packaging factories in Taiwan, China, and continue to increase local production capacity in Taiwan. In summary, TSMC’s semiconductor technology and manufacturing capabilities will continue to support the development of the global semiconductor industry and help customers achieve technological innovation. Next, we will introduce the current progress of 3nm production capacity expansion: We are steadily advancing the construction of three new 3nm factories around the world, located in Taiwan, Arizona, and Japan, to undertake the continued strong demand for 3nm orders in the next few years. In addition to building new factories, the Taiwan factory continues to renovate 5-nanometer equipment to expand 3-nanometer production capacity; at the same time, it relies on global factory manufacturing optimization capabilities to comprehensively improve production line production efficiency and increase wafer output. We simultaneously promote cross-process production capacity deployment and achieve flexible interoperability of 7nm, 5nm, and 3nm production lines. In summary, we use multiple means and synchronize efforts in multiple factories around the world to do our best to meet the production capacity needs of all customers. Let’s talk about the mature process strategy: TSMC’s overall mature process strategy remains unchanged. Our primary goal is to fully ensure customer supply and continue to expand mature process production capacity for high value-added segments, rather than reduce it. For example: Japan's JASM Factory No. 1 is expanding its mature process production capacity for CMOS image sensors; Germany's ESMC factory is increasing its mature process production capacity for automobiles and industrial chips. In the current market, except for specific tracks such as power management ICs and CMOS image sensors, the demand for other general mature processes is weak. Therefore, TSMC’s mature process resources will continue to focus on high value-added and strategic tracks to ensure the production capacity required for customer growth. Next, we will introduce the technological progress of A14 (1.4nm). As mentioned just now, the complexity of high-end process research and development continues to increase; for new technologies such as A14, the complete cycle from research and development, supporting factory construction to large-scale mass production takes 5 to 7 years, and there are no shortcuts. A14 is a second-generation nanosheet transistor process that achieves a complete generation technology leap compared to 2nm, taking into account performance improvement and power consumption optimization to meet the market's continued demand for high-performance, low-power computing power. Compared with 2nm, chip speed is increased by 10%-15% at the same power consumption; power consumption is reduced by 25%-30% at the same speed; chip density is increased by nearly 20%. The A14 research and development progress is fully in line with the plan. The internal test chip device performance has reached nearly 90% of the target value, and the 256-megabit static random access memory (SRAM) yield is close to 90%. Smartphone and HPC/AI track customers have shown a high willingness to cooperate and continue to in-depth docking of technical solutions. Customer tape-out progress is ahead of schedule; trial production will be launched in 2027, and large-scale mass production will be officially launched in 2028. Based on the continuous iteration technology route, we simultaneously launched A13 and A12 as extension technologies of the A14 process family. A13 is further optimized on the basis of A14, achieving a chip area reduction of more than 6% through 97% high-precision optical shrinkage; relying on collaborative optimization of the design process, performance and power consumption have been upgraded again; A13 design rules are backward compatible with A14 to ensure seamless migration of customers' intellectual property rights. A12 integrates innovative Super.
Next, I will hand over the microphone to Mr. Huang Renzhao, Chief Financial Officer of TSMC, who will review the current period's operations and announce next quarter's performance guidance. Huang Renzhao (Senior Vice President, Chief Financial Officer) Thank you Su Zhikai. Good afternoon everyone, thank you all for attending. My report will first sort out the core financial data for the second quarter of 2026, and then announce the performance guidance for the third quarter of 2026. First, the revenue share is broken down by process dimension: the 2-nanometer process contributed 3% of the wafer revenue in this quarter; 3-nanometer, 5-nanometer, and 7-nanometer accounted for 30%, 33%, and 11% respectively; the advanced processes of 7-nanometer and below accounted for 77% of the total wafer revenue. Breaking down the revenue by application platform: high-performance computing (HPC) increased by 20% month-on-month, accounting for 66% of total revenue this quarter; smartphone revenue fell 4% month-on-month, accounting for 22%; Internet of Things increased 4% month-on-month, accounting for 5%; automotive chips increased 15% month-on-month, accounting for 4%; consumer electronics (DCE) increased 5% month-on-month, accounting for 1%. Next, I will introduce the balance sheet to you. As of the end of the second quarter, the company's cash and marketable securities totaled NT$3.5 trillion, equivalent to US$110 billion. Liability side: Current liabilities increased by NT$144 billion month-on-month, mainly due to an increase of NT$58 billion in accounts payable and an increase of NT$48 billion in accrued liabilities and other accounts. In terms of financial turnover indicators: accounts receivable turnover days increased by 3 days month-on-month to 29 days; inventory turnover days increased by 7 days month-on-month to 87 days. The core reason is the ramp-up of 2 nanometer (N2) process production capacity. Cash flow and capital expenditure (CapEx): Cash flow generated from operating activities in the second quarter was NT$783 billion; capital expenditure investment was NT$496 billion; cash dividends distributed in the third quarter of 2025 were NT$156 billion. Overall, cash balances increased by NT$99 billion in the quarter from the previous quarter, with cash reaching NT$3.1 trillion at the end of the quarter. In U.S. dollars, second-quarter capital expenditures totaled $15.7 billion. The above is a complete financial review, and then the performance guidance for this quarter (third quarter) will be announced. Based on the current market operating environment, we estimate that total revenue in the third quarter will range from US$44.6 billion to US$45.8 billion; calculated at the midpoint of the range, it will increase by 12% month-on-month and 37% year-on-year. The exchange rate is assumed to be NT$32 for 1 U.S. dollar. The gross profit margin is expected to be in the range of 65%-67% and the operating profit margin in the third quarter is 56%-58%. Compared with the first quarter, the gross profit margin in the second quarter increased by 1.5 percentage points quarter-on-quarter to 67.7%, slightly exceeding the previous guidance. This was mainly due to cost optimization measures and a slight increase in overall capacity utilization, which was only partially offset by the dilutive effect of gross profit margin caused by the ramp-up of overseas fab production capacity. The median gross profit margin guidance range we just gave for the third quarter was 66%, down 1.7 percentage points from the previous quarter. The core reason is that the large-scale ramp-up of the 2-nanometer process will bring about a gross profit margin dilution of about 3 to 4 percentage points; however, the extremely strong demand for high-end processes, coupled with the continued promotion of cost optimization (including production efficiency improvements and cross-process capacity allocation optimization), will partially offset the downward pressure on gross profit margins. Looking at the second half of 2026, the factors that affect profitability have pros and cons. Let me explain to you here: the large-scale ramp-up of the 2-nanometer process is expected to dilute the gross profit margin by 3 to 4 percentage points in the second half of the year; as the scale of overseas expansion continues to expand, the gross profit margin dilution caused by the ramp-up of mass production of overseas wafer fabs in the next few years is expected to be 2%-3% in the initial stage and will expand to 3%-4% in the later period. Positive aspects: demand for high-end processes continues to be strong; secondly, we continue to rely on mature manufacturing capabilities to increase wafer output scale, and support profitability through optimization of cross-process capacity across the entire factory; finally, exchange rates are uncontrollable external variables and will also have a fluctuating impact on profitability. Next, we will introduce the full-year capital expenditure budget for 2026.
At TSMC, the scale of capital expenditures has always been highly tied to growth opportunities in the next few years. Relying on our leading and differentiated technical strength, the company is able to fully undertake the long-term structural growth needs of the three major industries of 5G, artificial intelligence, and high-performance computing. The current structural demand from customers continues to rise, including the incremental demand brought by the emerging intelligent AI market. Therefore, we have decided to increase our full-year capital expenditure budget for 2026 to US$60 billion to US$64 billion, and continue to increase investment to match customer growth needs. No matter whether the industry is in an up cycle or a down cycle, we will collaborate deeply with equipment suppliers in advance to plan production capacity in advance; just like customers will contact us in advance to plan wafer production capacity. We predict that the capacity expansion plan will not cause equipment supply bottlenecks. About 70%-80% of the full-year capital expenditure in 2026 will be invested in advanced process research and development and production line construction; about 10% will be used for specialty processes; and the remaining 10%-20% will be allocated to advanced packaging, testing, mask manufacturing and other supporting businesses. Even if we maintain high-scale capital expenditures in 2026, we will continue to create profitable growth for shareholders; at the same time, we will insist on steadily increasing cash dividends per share every year and every quarter. The company's total cash dividends in 2025 will be NT$467 billion, a year-on-year increase of 28.6%, and shareholders will receive a total dividend of NT$18 per share throughout the year; the dividend per share will increase to NT$24 in 2026, a year-on-year increase of another 33%; we expect cash dividends per share to continue to grow in 2027. Next I will hand over the microphone to Wei Zhejia, Chairman and CEO. Wei Zhejia (Chairman and CEO) In the long term, consumer and price-sensitive end markets are suppressed by rising component prices and macroeconomic uncertainty, putting demand under pressure. Therefore, we remain cautious when formulating our business plans, while focusing on core business fundamentals and continuing to consolidate our competitive advantages. But even so, the demand for artificial intelligence is still extremely hot. The long-term industrial wave of artificial intelligence continues to generate demand for massive computing power, which directly drives the demand for high-end chips to continue to rise. Our customers and their downstream customers (mainly cloud service providers) continue to send us strong demand signals and optimistic market expectations, so we maintain a high degree of confidence in the long-term industry trend of AI. Relying on differentiated technical strength and extensive customer base, we have now raised our expectations: the year-on-year growth rate of US dollar-denominated revenue in 2026 is slightly more than 40%. Let’s talk about the acceleration of industry demand brought about by intelligent AI. The AI market is iterating at an extremely fast pace. The rise of intelligent AI has once again increased the value of CPUs in AI data centers. In addition to AI acceleration chips, it has further boosted the overall demand for chips. We believe this is a major benefit to TSMC: whether customers use x86, Arm or RISC-V architecture CPUs, almost all of them are TSMC’s cooperative customers. We have deeply collaborated with CPU customers to provide the most advanced manufacturing processes and supporting production capacity to help customers seize opportunities in the intelligent AI market. Next, we will introduce TSMC’s global production capacity expansion strategy. In order to match the long-term structural demand growth of the semiconductor industry, TSMC continues to plan production capacity in depth with customers and their downstream terminals. The lead time for high-end process research and development, customer product introduction, and mass production is extremely long, so we can clearly grasp our customers' product routes and mass production plans for the next few years. This is crucial: the complete cycle from new technology research and development, supporting production capacity construction to large-scale mass production takes more than 5 years. At the internal level, TSMC has established a standardized production capacity assessment system to assess market demand from top to bottom combined with bottom-up. This work continues to be promoted throughout the year. After comprehensive assessment, we increased capital expenditure and expanded production capacity to support the future growth of our customers. Relying on the strong support of leading US customers, US federal, state, and local governments, we hereby officially announce that TSMC will invest an additional US$100 billion in Arizona to build multiple 2-nanometer and below logic wafer production lines, and simultaneously support advanced packaging factories to meet the continued strong order demand of leading US customers in the next few years.
We believe that this additional investment will further improve the domestic semiconductor industry chain in the United States, strengthen the resilience of the supply chain, and at the same time create a large number of high-paying high-tech jobs in the United States. At the same time, in the next few years, we will build 13 new high-end process and advanced packaging factories in Taiwan, China, and continue to increase local production capacity in Taiwan. In summary, TSMC’s semiconductor technology and manufacturing capabilities will continue to support the development of the global semiconductor industry and help customers achieve technological innovation. Next, we will introduce the current progress of 3nm production capacity expansion: We are steadily advancing the construction of three new 3nm factories around the world, located in Taiwan, Arizona, and Japan, to undertake the continued strong demand for 3nm orders in the next few years. In addition to building new factories, the Taiwan factory continues to renovate 5-nanometer equipment to expand 3-nanometer production capacity; at the same time, it relies on global factory manufacturing optimization capabilities to comprehensively improve production line production efficiency and increase wafer output. We simultaneously promote cross-process production capacity deployment and achieve flexible interoperability of 7nm, 5nm, and 3nm production lines. In summary, we use multiple means and synchronize efforts in multiple factories around the world to do our best to meet the production capacity needs of all customers. Let’s talk about the mature process strategy: TSMC’s overall mature process strategy remains unchanged. Our primary goal is to fully ensure customer supply and continue to expand mature process production capacity for high value-added segments, rather than reduce it. For example: Japan's JASM Factory No. 1 is expanding its mature process production capacity for CMOS image sensors; Germany's ESMC factory is increasing its mature process production capacity for automobiles and industrial chips. In the current market, except for specific tracks such as power management ICs and CMOS image sensors, the demand for other general mature processes is weak. Therefore, TSMC’s mature process resources will continue to focus on high value-added and strategic tracks to ensure the production capacity required for customer growth. Next, we will introduce the technological progress of A14 (1.4nm). As mentioned just now, the complexity of high-end process research and development continues to increase; for new technologies such as A14, the complete cycle from research and development, supporting factory construction to large-scale mass production takes 5 to 7 years, and there are no shortcuts. A14 is a second-generation nanosheet transistor process that achieves a complete generation technology leap compared to 2nm, taking into account performance improvement and power consumption optimization to meet the market's continued demand for high-performance, low-power computing power. Compared with 2nm, chip speed is increased by 10%-15% at the same power consumption; power consumption is reduced by 25%-30% at the same speed; chip density is increased by nearly 20%. The A14 research and development progress is fully in line with the plan. The internal test chip device performance has reached nearly 90% of the target value, and the 256-megabit static random access memory (SRAM) yield is close to 90%. Smartphone and HPC/AI track customers have shown a high willingness to cooperate and continue to in-depth docking of technical solutions. Customer tape-out progress is ahead of schedule; trial production will be launched in 2027, and large-scale mass production will be officially launched in 2028. Based on the continuous iteration technology route, we simultaneously launched A13 and A12 as extension technologies of the A14 process family. A13 is further optimized on the basis of A14, achieving a chip area reduction of more than 6% through 97% high-precision optical shrinkage; relying on collaborative optimization of the design process, performance and power consumption have been upgraded again; A13 design rules are backward compatible with A14 to ensure seamless migration of customers' intellectual property rights. A12 integrates innovative Super.