medium convictionactive · updated 2026-05-12T00:00:00.000Z
Samsung SF2P yields hit 70% + Taylor 90% ready + Apple visit → genuine third foundry leg → order flow splits, not concentrates
Samsung's SF2P (2nm GAA-improved variant) reached 70% yield in January 2026 — potentially above Intel 18A's current range and converging on TSMC N2. The Taylor TX fab is 90% ready for mass production with 2nm initial in Q2 2026 and full mass production by 2027. Apple executives physically visited Taylor. The 'TSMC-or-Intel' framing is now bilateral-by-default but trilateral-in-practice: Samsung is a genuine third leg with US footprint, yields, and active Apple evaluation. Threatens the concentration-on-Intel story.
The chain
1
Samsung SF2P (2nm GAA improved variant) hit 70% yield as of January 2026 — converging toward TSMC N2 early yields and ahead of Intel 18A's mid-2025 vintage range.
From 2026-05-11-autoresearch-samsung-foundry-third-alternative-2026: "Its SF2P (2nm GAA improved variant) reached 70% yield as of January 2026 — potentially above Intel 18A's current range and converging toward TSMC N2 early yields"
From 2026-05-11-autoresearch-intel-18a-yield-vs-tsmc-samsung-2026: "Samsung SF2P has unexpectedly hit 70% yield as of January 2026 — closing the gap to TSMC and ahead of where Intel 18A appears to be today"
2
Samsung's Taylor, Texas fab is 90% ready for mass production (April 16, 2026), with 2nm initial manufacturing in Q2 2026 and mass production by 2027. A second Taylor fab is planned.
From 2026-05-11-autoresearch-samsung-foundry-third-alternative-2026: "The Taylor, Texas fab is now 90% ready for mass production (April 16, 2026), with 2nm initial manufacturing targeted for Q2 2026 and mass production by 2027. Samsung is planning a second Taylor fab"
3
Apple executives physically visited Samsung's Taylor facility — upgrading the relationship from desk research to facility-level diligence. Apple is now evaluating Samsung (US) AND Intel as TSMC diversification alternatives.
From 2026-05-11-autoresearch-samsung-foundry-third-alternative-2026: "Apple executives have physically visited Samsung's Taylor facility, upgrading the evaluation from desk research to facility diligence"
From 2026-05-11-autoresearch-intel-18a-yield-vs-tsmc-samsung-2026: "Apple's primary 2nm capacity remains reserved at TSMC; Intel (via 18A-P) and Samsung (US Texas fab) are both being evaluated as diversification alternatives"
4
The original wiki framing of 'TSMC-or-Intel' understates Samsung's competitiveness. The market now has three leading-edge fabs with US footprint: TSMC Arizona (3nm 2H27), Intel Fab 52 (18A in volume), Samsung Taylor (2nm Q2 2026 initial).
From 2026-05-11-autoresearch-samsung-foundry-third-alternative-2026: "The original thesis framing — 'the world is bilateral: TSMC vs. Intel' — needs revision. Samsung is a genuine third option with US footprint, improving yields, and active Apple evaluation"
5
Order flow consequence: anchor customer commitments (AWS, Microsoft, Apple, Terafab) may split across Intel + Samsung rather than concentrating at Intel. Re-rate magnitude for INTC alone is bounded by the share of new orders Samsung captures.
From 2026-05-11-autoresearch-us-fab-bottleneck-anchor-update-may-2026: "The risk is that Samsung's improved Taylor capacity means the order flow will split rather than concentrate on Intel"
What would falsify this
- Step 1: Samsung SF2P yields regress below 60% in Q2/Q3 2026 reporting.
- Step 3: Apple publicly names Intel only (or TSMC only) for diversification, with Samsung dropped.
- Step 5: All four announced Intel anchors confirm Intel as primary, no Samsung overlap — would reconcentrate the re-rate.
Contradictions / tensions
- 70% yield single data point — needs second-quarter confirmation that Samsung yields are sustainable, not transient.
- Apple visit ≠ order. Apple's primary 2nm capacity is still at TSMC. The diversification commitment is preliminary on all three legs.
Implications
- Beneficiary expansion: Samsung Foundry (005930.KS) as previously-discounted reshoring play.
- Caps the magnitude of the tsmc-saturation-to-intel-anchor-stack re-rate for INTC alone — the trade is alternative-foundry, not Intel-monopoly.
- ASML, AMAT, KLA still win regardless — all three serve all three fabs.
Companies
Concepts
US fab capacity bottleneck → Intel as the alternative foundryTSMC capacity shortfall and 4-year pricing power
Open questions
none