Taiwan Semiconductor Manufacturing Company reported a strong quarterly performance, with its stock rising. The company announced it plans to increase capital expenditures to between $52 billion and $56 billion in 2026, up from $41 billion in 2025.
This spending plan indicates sustained demand for artificial intelligence chips. Other companies are expected to gain from this dynamic.
TSMC's customers, who are seeking more manufacturing capacity, will benefit. Nvidia, whose graphics processing units train large language models, is a clear beneficiary due to its CUDA software advantage. Broadcom is also set to gain as it helps design custom AI chips; Citigroup analysts see potential for its AI revenue to increase fivefold in two years. Advanced Micro Devices holds a strong position in data center CPUs and its GPUs may gain in the AI inference market.
ASML stands to gain significantly as TSMC's increased spending will require more advanced semiconductor manufacturing machines. ASML is the sole producer of the extreme ultraviolet lithography machines needed. TSMC's push for advanced chip architecture also sets up ASML to eventually sell its newer, more expensive High-NA EUV platform.
Memory makers are another beneficiary group. AI chips require high-bandwidth memory, a specialized DRAM type. The DRAM market is currently tight with rising prices, and TSMC's expansion may keep supply constrained. This benefits DRAM companies like Micron Technology, SK Hynix, and Samsung, which are already seeing high demand and price increases.
Cloud computing and neocloud companies are also positioned to benefit. Major cloud providers Amazon, Microsoft, Alphabet, and Oracle are spending aggressively on data center capacity for AI services. On an earnings call, TSMC CEO C.C. Wei stated he received evidence from cloud service providers that AI is helping their businesses and that infrastructure spending is driving strong returns. This assurance influenced TSMC's capex decision.
Smaller neocloud providers like CoreWeave and Nebius Group, which rent computing power for AI workloads, are also seeing high demand, though they operate with more leverage than the larger cloud companies.