TSMC(TSM) - 2024 Q4 - Earnings Call Transcript

Financial Data and Key Metrics Changes - Fourth quarter revenue increased by 14.3% sequentially in NT, driven by strong demand for 3 nanometer and 5 nanometer technologies [9] - Gross margin rose by 1.2 percentage points sequentially to 59%, reflecting higher capacity utilization and productivity gains [10] - Operating margin increased by 1.5 percentage points sequentially to 49% [10] - Full year 2024 revenue increased by 30% in US dollar terms to $90 billion, with gross margin up 1.7 percentage points to 56.1% [17] - Full year EPS increased by 39.9% to TWD 45.25, and ROE increased by 4.1 percentage points to 30.3% [17] Business Line Data and Key Metrics Changes - 3 nanometer process technology contributed 26% of wafer revenue in Q4, while 5 nanometer and 7 nanometer accounted for 34% and 14%, respectively [11] - High-Performance Computing (HPC) revenue increased by 19% quarter-over-quarter, accounting for 53% of Q4 revenue [12] - Smartphone revenue increased by 17% to account for 35% of Q4 revenue, while IoT decreased by 15% to 5% [12] - For the full year, HPC revenue increased by 58% year-on-year, while smartphone revenue increased by 23% [13] Market Data and Key Metrics Changes - Advanced technologies (7 nanometer and below) accounted for 74% of wafer revenue in Q4, up from 58% in 2023 [11][12] - The Foundry 2.0 industry grew by 6% year-over-year in 2024, slightly lower than previous forecasts [31] - TSMC forecasts the Foundry 2.0 industry to grow by 10% year-over-year in 2025, supported by robust AI-related demand [32] Company Strategy and Development Direction - TSMC plans to invest between$ 38 billion and $42 billion in capital expenditures for 2025, focusing on advanced process technologies [27] - The company aims to capture growth from industry megatrends such as 5G, AI, and HPC [27][37] - TSMC is expanding its global manufacturing footprint, with new fabs in the US, Japan, and Europe to meet customer demand [39][42][43] Management's Comments on Operating Environment and Future Outlook - Management expressed confidence in outperforming industry growth, forecasting a revenue increase of close to mid-20s percent in US dollar terms for 2025 [33] - AI-related demand is expected to double in 2025, with AI accelerators projected to be the strongest driver of HPC platform growth [34][36] - Management acknowledged potential challenges from inflationary costs and ramp-up expenses associated with new fabs [24][25] Other Important Information - TSMC's effective tax rate for 2025 is expected to be between 16% and 17% [19] - The company generated TWD 1.8 trillion in operating cash flow and TWD 870 billion in free cash flow in 2024 [18] Q&A Session Summary Question: TSMC's US Future Strategy - Management discussed the potential for investing in the latest technology nodes in the US, emphasizing the need for close proximity to R&D for ramping new technologies [54][60] - The company maintains open communication with the US government and is focused on customer needs for overseas fab expansion [62] Question: Gross Margins Outlook - Management indicated that reaching over 60% gross margin is possible if utilization remains high, but noted that US fab costs are higher due to several factors [66][72] Question: AI Business Opportunities - Management confirmed ongoing collaboration with memory suppliers for HBM controllers, with high volume expected in the next year [88] Question: Impact of US Export Restrictions on China Business - Initial assessments suggest that the impact of new US export restrictions on TSMC's China business is manageable [93] Question: Arizona Fab Update - The first fab in Arizona is in volume production, with the second fab's construction nearly complete and plans for a third fab underway [106] Question: Long-term Gross Margin Target - Management reiterated that the long-term gross margin target remains at 53% and higher, citing various factors affecting profitability [121] Question: CoWoS Capacity and Non-AI Applications - Management acknowledged current high demand for AI-related applications but indicated that non-AI applications will also adopt CoWoS technology in the future [125]