TSMC reports December revenue up 58% YoY to ~$8.4B, pushing its 2024 revenue up 34% YoY to ~$88.02B, driven by AI chip demand from Nvidia, Broadcom, and others
Nikkei Asia :
Context & Ripple Effects
TSMC's December result caps a year in which AI-chip orders from Nvidia, Broadcom, and other customers became a material driver of foundry demand. It follows the company's 40.1% year-over-year revenue growth in the prior quarter, also tied to surging demand for advanced AI chips.
The report is an early marker of a pattern that persisted in later coverage: TSMC continued to post AI-led growth and reported that demand still exceeded supply in 2025. That makes the revenue increase relevant beyond a single monthly update.
First-order effects
- TSMC enters 2025 with substantially higher annual revenue and a clear concentration of incremental demand in AI-chip production for customers including Nvidia and Broadcom.
- Nvidia, Broadcom, and other TSMC customers gain evidence that their AI-chip programs are translating into large manufacturing orders, while TSMC's advanced-capacity allocation becomes more consequential.
Second-order effects
- Strong utilization and AI-led orders increase the pressure on rival chip designers to secure access to leading-edge foundry capacity, rather than treating fabrication as a readily available input.
- The demand signal extends through the AI hardware supply chain: customers' chip volumes can pull on adjacent packaging, memory, and infrastructure spending, though this report does not quantify those effects.
Third-order effects
- If repeated, this pattern strengthens TSMC's position as a key production bottleneck in AI infrastructure, making access to advanced manufacturing a strategic constraint for chip vendors.
- The broader shift is from episodic accelerator launches toward an AI infrastructure capital cycle in which foundry capacity and associated supply-chain availability shape how quickly compute products can scale.
The trend: This is one data point in the AI infrastructure supercycle, where demand for AI compute is transmitting into sustained pressure on leading-edge chip manufacturing capacity.