Foundry Power in 2030: Why China Will Lead in Capacity but Taiwan Will Still Rule the Market
The semiconductor industry is undergoing a transformation.
By 2030, China is expected to become the world’s largest semiconductor manufacturing base by volume, surpassing Taiwan in total foundry capacity measured in wafers per month. But this physical lead will not translate into dominance over the semiconductor market’s most lucrative segment. The global center of gravity in chip manufacturing is splitting into two regimes: volume versus value.
China, with a flood of state-backed investment and trailing-edge fab buildouts, is racing to secure supply chain sovereignty by expanding 28nm and larger nodes. Taiwan, meanwhile, is consolidating its position at the top of the value stack — not by outproducing, but by outpricing. TSMC’s (NASDAQ: TSM) grip on 3nm, 2nm, and eventually sub-2nm nodes ensures that Taiwan will continue to generate the majority of global foundry revenue, even as China pulls ahead in wafer capacity.
This divergence — of physical scale versus financial yield — is not a contradiction. It is the defining structure of the modern semiconductor economy. What follows is a detailed breakdown of how capacity, node strategy, and company alignment are evolving across regions, using a unified framework of 8-inch wafer equivalents to compare apples to apples. From 2019 to 2030, this lens reveals not just growth but strategic intent. To keep up with the latest in the important semiconductor market and what stocks to own, we invite you to subscribe to our Gilder’s Technology Report investment newsletter and weekly updates.
The 2025 Landscape: Volume Favors China, Value Favors Taiwan
According to Table 1, the global semiconductor foundry landscape in 2025 is bifurcated by design and destiny. On one side sits the volume buildout — a mass of capacity in mature nodes like 65nm and 45nm, driven by auto ICs, power management chips, and microcontrollers. On the other sits the value front, advanced logic nodes at 5nm and below where margins soar and complexity concentrates. When measured in 8-inch equivalents, China has already taken the lead in physical volume at mature nodes, with 1,100 kwpm of ≥65nm capacity — far outpacing Taiwan, Korea, or the United States. Taiwan, by contrast, leads in advanced-node throughput, with nearly 600 kwpm at ≤10nm alone.
The regional divergence is more than technical — it is geopolitical. Taiwan and Korea continue to invest in high-value node transitions, while China adds capacity at legacy nodes as a hedge against export control regimes and to support domestic substitution. The United States remains modest in volume but increasingly strategic in node targeting through the CHIPS Act and Intel’s resurgence.
The Silent Accumulation: Seven Years of Wafer Buildout
According to Table 2, the evolution from 2019 through 2025 tells a clearer story of who is building and why. China’s foundry capacity nearly doubled in six years, adding over 600,000 8-inch wafer equivalents per month. Taiwan, while still expanding, did so more slowly — yet added disproportionate value by targeting smaller geometries. Korea’s expansion has been methodical, centered on Samsung’s ramp of GAA at 3nm. Meanwhile, U.S. growth has been largely backloaded, with early CHIPS Act funding still maturing into production output.
The most significant fact in this table is not the growth itself, but where it occurs. China is not chasing bleeding-edge nodes — it is amassing the foundation of the analog, auto, industrial, and power markets. It is building the chip equivalent of steel and concrete, not jewelry.
Revenue and Reality: The Density Differential
According to Table 3, the revenue generated by each wafer varies by an order of magnitude depending on process node. A 3nm wafer at TSMC can command over $17,000. A 65nm display driver wafer might earn $700. This is why physical capacity and financial dominance are diverging: TSMC, despite producing fewer wafers than China, earns over 60% of global foundry revenue. The industry is now split into two economies — volume and value.
The investor takeaway is stark: volume is not value. And the control of high-margin nodes creates far more strategic leverage than the number of wafers alone.
Forecasting the Future: Toward 2030
According to Table 4, the forecast through 2030 shows China continuing to grow capacity faster than any other nation. While Taiwan and Korea will expand advanced nodes, their total wafer output will rise more slowly. China’s wafer expansion is not a bubble — it is a rebalancing of the industrial base.
If current growth rates hold, China will surpass Taiwan in total installed capacity sometime between 2027 and 2028. But whether it surpasses in value will depend on tooling access, packaging advances, and the ability to scale high-yield processes at 14nm and below.
The Corporate Map: Foundry Share by Company
According to Table 5, capacity leadership at the company level still belongs to TSMC, though SMIC is closing the gap in physical volume. Samsung remains a strong second player in advanced logic, while Intel Foundry begins its long climb into relevance. UMC, GF, and Vanguard maintain consistent positions as mature-node anchors.
The rise of SMIC and the reemergence of Intel Foundry signal that while the top remains stable, the middle tier is in motion — especially in China.
Dissecting the Node Strategy: Who Builds What
According to Table 7, the node-level breakdown by company reveals the actual battlefield. TSMC has more capacity at ≤10nm than the entire rest of the world combined. SMIC and Hua Hong dominate at 65nm and 45nm. Intel (NASDAQ: INTC), for all its ambition, remains small in absolute terms but is positioned entirely at advanced logic.
The table makes clear that China’s strength is in breadth, not depth. SMIC, Hua Hong, and Nexchip have built out a distributed architecture of trailing-edge logic nodes that support everything from set-top boxes to automotive MCUs. But it is TSMC and Samsung that still control the engines of AI, HPC, and advanced smartphone SoCs.