TSM Stock: Why SMIC N+3 Still Trails TSMC N6

Taiwan Semiconductor Manufacturing (TSM) received indirect validation of its long-term competitive position from an unexpected source on June 14, 2026: a Semianalysis STEEL Teardown Lab report comparing China's SMIC N+3 process technology to TSMC's N6 node. The report's headline finding — that SMIC's metal pitch is technically smaller than Intel 18A's smallest — is misleading. The substantive finding is that SMIC achieves only TSMC N6 density (a three-year-old TSMC node) via aggressive DUV multi-patterning. For TSM shareholders, the implication is unambiguous: SMIC remains three generations behind TSMC's current leading edge, despite well-funded efforts under export-control conditions.

What the SMIC STEEL teardown actually shows

Semianalysis's STEEL (Teardown Engineering & Evaluation Lab) reverse-engineered Huawei's Kirin 9030 Pro SoC, the most advanced chip currently shipping from SMIC's N+3 process. Key findings:

• SMIC N+3 smallest metal pitch is genuinely smaller than Intel 18A's smallest. This is the cherry-picked metric that drives the headline.
• N+3 reaches TSMC N6 logic density — a node TSMC introduced in 2020. The pitch advantage does not translate to viable production economics because SMIC requires aggressive DUV multi-patterning.
• Process maturity is materially worse at SMIC. Yield rates, defect density, and process control all trail TSMC by significant margins.
• Kirin 9030 Pro performance matches Android flagship phones from approximately three years ago. Current flagship SoCs from Apple, Qualcomm, MediaTek, and Samsung use TSMC N3 nodes.
• Without EUV (export-controlled), SMIC is fundamentally limited to DUV multi-patterning. Each generation requires more lithography steps, more chemistry, more time. The technical and economic gap widens with each node.

The teardown is significant because it provides hard data, not narrative. Industry observers have long suspected SMIC could not match TSMC without EUV access. The STEEL report makes the gap quantitative.

What the Q1 2026 numbers reveal

TSMC's Q1 2026 financial statements (reported in Taiwan dollars) showed revenue of TWD 1.13 trillion (approximately $36 billion at recent exchange rates). Gross profit was TWD 751 billion (approximately $24 billion). Operating income was TWD 657 billion (approximately $21 billion). Net income reached TWD 572 billion (approximately $18 billion). Diluted EPS was TWD 111.55. Free cash flow was TWD 377 billion (approximately $12 billion) (drillr financial statements).

Operating margin: approximately 58%. That is not a contract manufacturer margin profile. That is a near-monopolist margin profile.

Cash and short-term investments stood at TWD 3.38 trillion (approximately $108 billion) against total debt of TWD 1.09 trillion (approximately $35 billion). TSMC has enormous balance sheet capacity to fund continued capex.

The full-year 2025 results showed revenue of TWD 3.85 trillion (approximately $122 billion), operating income of TWD 1.96 trillion (approximately $62 billion), EPS of TWD 333. The trajectory through 2025 was strong, and Q1 2026 continued that momentum.

How the process node leadership translates to economics

TSMC's leadership position is sustained through three reinforcing mechanisms:

N3/N2 capacity allocation. TSMC's N3 (3nm) capacity is essentially fully booked through 2026, with customer allocation reflecting strategic commercial relationships. Apple has primary N3 allocation; Nvidia has large allocation; AMD has growing allocation. The N2 (2nm) capacity for 2027 is also allocated. Customer commitments are multi-year. This is the pricing-power anchor that creates TSMC's 58% operating margin.

Customer-specific design service. TSMC has built engineering capacity to support Apple's silicon design, Nvidia's accelerator design, AMD's CPU/GPU design. The design-service depth creates customer stickiness that survives any individual node generation.

Capex compounding. TSMC's 2026 capex of approximately $44 billion exceeds the combined capex of Samsung Foundry and SMIC by 4-5x. The capex advantage compounds: more node capacity, more cutting-edge research, more strategic options.

The June 14 STEEL teardown reinforces the structural picture. SMIC's N+3 effort represents China's best attempt under export-control conditions, and that best attempt reaches only TSMC N6 density. The foundry moat is intact.

Why the Intel 18A comparison matters strategically

The STEEL teardown also informs the Intel 18A story. Intel 18A is positioned to be competitive with TSMC N3 on technology, but Intel's manufacturing capacity is currently a small share of global foundry output. Even if Intel 18A reaches volume production at parity with TSMC N3, TSMC's customer allocation advantage and capex compounding maintains the leadership position.

The cohort hierarchy as of June 2026:

• TSMC — leading edge, dominant allocation, sustainable margin
• Samsung Foundry — competitive on tech, behind on yield, limited customer wins
• Intel — 18A competitive with TSMC N3, capacity building, requires capital raise to execute
• SMIC — N+3 reaches TSMC N6 only, structural disadvantage without EUV

The June 14 teardown does not change this hierarchy. It validates it with hard data.

What the cohort context shows

Drillr terminal records institutional flow data showing TSM ownership concentration among long-term strategic positions and large quantitative funds. The shareholder base values stability, growth, and pricing power — all of which TSM continues to deliver.

The June 14 STEEL teardown reduces a tail risk for TSM: the concern that Chinese foundry capacity might eventually compete on advanced nodes. The teardown shows the gap remains structural and widens with each generation. Investors looking at TSM with three-to-five year horizons are getting cleaner visibility on the moat than before.

What to monitor through 2026

• TSMC Q2 2026 earnings (expected mid-July) for capex commentary and customer allocation updates.¹
• Apple, Nvidia, AMD specific TSMC capacity commitments and pricing.
• Intel 18A commercial shipment timing and yields.
• Any Chinese government policy shifts affecting SMIC funding or foreign technology access.
• Samsung Foundry customer win announcements (any large customer signs are bearish for TSMC near-term).

What this means for TSM positioning

TSM trades at recent prices around $230 (US ADR), reflecting strong fundamental momentum but ongoing geopolitical risk premium. The forward multiple at approximately 22-25x consensus 2026 earnings is comparable to the 5-year average. The June 14 STEEL teardown is incremental positive that reinforces the structural thesis without changing the valuation calculus.

For investors looking at semiconductor exposure with the cleanest moat profile and the most data-validated competitive position, TSM remains the highest-conviction expression. The teardown provides additional comfort that the process leadership is sustainable through 2027-2028 — and likely beyond.

The risk is not technology. It is geopolitical: Taiwan Strait tensions and US-China relations. The reward is sustained pricing power and customer allocation premium. Investors who can sit with the geopolitical risk are getting one of the highest-quality assets in global equities.

Footnotes

1. Semianalysis, "Is SMIC N+3's Metal Pitch Smaller than Intel 18A's?," STEEL Teardown Lab, June 14, 2026. https://newsletter.semianalysis.com/p/steel-smic-n3-teardown ↩