Who's actually winning the foundry race?
TSMC holds ~two-thirds of the market and the yield lead; Intel is betting its comeback on 18A; Samsung trails on yield. Our read on a race where one player's dominance keeps compounding. (Our opinion, not investment advice.)
On paper the 2nm race looks like a three-way tie — all three leaders reached 2nm-class volume in 2025. In practice it is not close. TSMC holds roughly two-thirds of the foundry market and, more importantly, the yield and capacity lead, so the biggest customers queue for its lines and fund its next node. That is a compounding advantage: volume pays for the next fab, which wins the next generation of customers.
The challengers are running different bets. Intel is wagering its turnaround on 18A and its PowerVia backside-power lead, trying to win external foundry customers it has never really had — a real technology story attached to an unproven business one. Samsung has the technology and the capital but has trailed on yield, the single number that decides whether customers trust a node. China's SMIC is capable but capped by its lack of EUV.
Our read is that node parity on a slide is not the same as winning, and the metric that matters most is mundane: yield-backed volume share. Until a challenger demonstrably closes the yield gap and pulls marquee customers off TSMC, the leader's dominance is more likely to widen than narrow. We track share, capex and node together precisely because the race is decided by their interaction, not by who announces the smallest number.
This is our interpretation of the public data, not investment advice.