TSMC Raises 2026 Capex to $64 Billion Amid Sustained AI Chip Demand
The tech sector commits $1 trillion to AI infrastructure, intensifying the global race for compute capacity and specialized hardware.
The story
The technology sector is projected to invest an unprecedented $1 trillion into artificial intelligence infrastructure, marking what Gartner analysts describe as humanity's largest infrastructure project. This colossal expenditure is driving up hardware and software prices across the industry, directly impacting enterprise customers and data center equipping strategies.
Mirroring this trend, Taiwan Semiconductor Manufacturing Company (TSMC) has increased its 2026 capital expenditure budget to between $60 billion and $64 billion, up from prior estimates of $52-56 billion. The foundry cited stronger-than-expected structural demand for AI and high-performance computing as the primary driver, with 70% to 80% of this spending allocated to leading-edge process technologies. This uptick reflects long-term customer commitments and signals a sustained, aggressive build-out phase for the foundational layers of AI.
The build-out
| Project | Who | Scale | Where |
|---|---|---|---|
| Naver AI Data Center Expansion | Nvidia, Naver, Brookfield | $1 billion Nvidia investment, up to $9 billion from Brookfield | South Korea (for sovereign AI infrastructure in Asia, Europe, Middle East) |
| South Korea National AI Initiatives | South Korea government, Samsung Electronics, SK Group, U.S. tech firms | $950 billion in new initiatives, including over $500 billion in partnership value | South Korea |
Supply & policy signals
Enterprise SSDs (eSSDs) becoming higher priority due to shift from AI training to inference workloads.
Implication: Increased demand for faster storage solutions, with Samsung Electronics and SK Hynix positioning products to meet hyperscaler needs for reduced server space and energy.
Chinese memory-chip maker CXMT's market debut.
Implication: CXMT surged 530% in Shanghai, becoming China's most valuable listed chipmaker, underscoring the strategic importance of AI memory amid Beijing's localization drive.
Reporting + analyst voices: grounded via Google Search at publish time.