TSMC supply squeeze pushes chip giants toward Intel and Samsung
Taiwan Semiconductor Manufacturing Co. (TSMC) reported a 77% increase in quarterly earnings but is struggling to keep pace with demand for high-end AI chips. This supply shortage is prompting major chip designers to explore alternative manufacturing partnerships with competitors like Intel and Samsung.
Key Takeaways
- Revenue grew 34% year-over-year while operating margins exceeded 60% due to aggressive AI accelerator and CPU demand.
- TSMC's 2nm process now accounts for 3% of wafer revenue, with advanced nodes generating over 75% of total sales.
- Apple and Fortinet have inked preliminary deals with Intel's foundry services for custom chip manufacturing.
- Samsung secured a multi-year AI chip deal with Tesla and is reportedly in talks with AMD and Alphabet for server CPUs and TPUs.
Why It Matters
The structural supply-demand imbalance at TSMC is breaking the company's virtual monopoly on advanced logic. While TSMC remains the performance leader, the sheer lack of available wafers is forcing tier-one designers to treat Intel and Samsung as primary alternatives rather than secondary backups. For the streaming and data center ecosystem, this diversification may eventually lower the cost of hardware by providing manufacturers more leverage in price negotiations. However, in the immediate term, supply constraints will continue to bottleneck the deployment of AI infrastructure. Watch the yield rates of Intel’s 18A process as they enter high-volume manufacturing; achieving parity with TSMC is the final hurdle for these rivals to secure long-term contracts.
Additional Context
The global foundry landscape is shifting as TSMC’s capacity reaches a firm ceiling. Per TrendForce reporting from July 2026, TSMC's share of global foundry revenue reached 72.2% in late 2025, but the industry's focus has moved from raw wafer starts to advanced packaging. Bottlenecks in Chip-on-Wafer-on-Substrate (CoWoS) packaging have become the primary constraint for AI hardware, with capacity reportedly sold out through 2026. This shortage has allowed Samsung to pivot its strategy, pitching itself as a 'one-stop' provider that combines logic, memory, and packaging. In July 2026, Samsung reportedly landed orders for Meta’s MTIA accelerators and Anthropic’s 2nm AI chips, helping the company target a return to foundry profitability by the fourth quarter. Simultaneously, Intel has accelerated its production roadmap to capitalize on the shortfall. Per Intel's June 2026 update, the 18A-P process — a performance-optimized variant of its 18A node — has entered risk production, promising a 9% performance boost over its predecessor. Intel is also the first to utilize High-Numerical Aperture (High-NA) Extreme Ultraviolet (EUV) lithography for volume production, an investment the U.S. government supported through an equity stake in 2025. This technological head start is designed to lure customers like Apple, who previously relied almost exclusively on TSMC for leading-edge nodes. TSMC is responding with its largest-ever expansion beyond Taiwan. Per company updates from January 2026, the first fab in its $165 billion Arizona complex is now mass-producing 4nm chips, with 3nm production scheduled to start in the second half of 2027. Despite these efforts, NVIDIA CEO Jensen Huang noted in February 2026 that demand for AI wafers is so high that TSMC would need to double its current capacity within the next decade to fully satisfy the market. This persistent gap ensures that for the next several cycles, the primary competition in the chip industry will be for manufacturing slots, not just architectural performance.
Read full article at finance.yahoo.com
Get this in your inbox → Subscribe
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source