← Sazabi Research

SMIC (0981.HK) — Q2 2026 Results: The Margin, Not the Revenue

First $3B quarter; gross margin 25.3% beat; the repricing from cyclical foundry to strategic asset

The Take

This quarter isn't about the revenue headline — it's the margin. Gross margin moving 20.1% → 25.3% → guided 26–28% is the real story, and it's why the stock rallied on results day. The +36% YoY revenue is nice, but the operating leverage from running at 93.7% utilization is the durable driver. SMIC is being repriced by the market from "cyclical mature-node foundry" toward "strategic leading-edge asset."

1. Business Model Summary

SMIC is mainland China's largest and most advanced pure-play wafer foundry, manufacturing integrated circuits on behalf of fabless customers. It is the only Chinese foundry able to mass-produce leading-edge logic (7nm-class CPUs/GPUs), which makes it a genuine scarcity asset in a world where advanced-node capacity is concentrated in TSMC (TSM) and Samsung.

• Revenue model: per-wafer foundry fees, driven by shipments × average selling price (ASP) × product mix.
• Wafer mix (Q2 2026): 12-inch wafers 78.2% of revenue (up from 76.4% in Q1), 8-inch 21.8%.
• Geography: China = 90.2% of revenue — the "China-for-China" localization trade in one number.
• Growth drivers: AI demand spillover from memory into logic/analog/power; mature-node capacity crunch as overseas foundries reallocate to AI/HBM; export-control displacement pulling foreign clients back to China.

2. End-Market & Customer Concentration

SMIC's Q2 2026 revenue by application:

End market	% of revenue
Consumer electronics	44.2%
Smartphones	16.9%
Industrial & automotive	16.5%
Computer/tablet, IoT & other	~22.4%

Consumer + smartphone ≈ 61% of revenue means SMIC remains levered to the Chinese consumer cycle. On the customer side, Huawei is the single most strategically important client on the advanced-node side, but SMIC does not disclose a formal customer-concentration table in the quarterly release. The broader point: demand is diversified across thousands of fabless customers, while strategic value is concentrated in a handful of leading-edge wins.

3. Financial Statement Analysis

Q2 2026 vs. Q1 2026 vs. Q2 2025

Metric	Q2 2025	Q1 2026	Q2 2026	YoY	QoQ
Revenue ($M)	2,209	2,505	3,006	+36.1%	+20.0%
Gross profit ($M)	450	504	761	+69.1%	+51.0%
Gross margin	20.4%	20.1%	25.3%	+4.9pp	+5.2pp
Net profit attrib. ($M)	~132	197	479	+261.7%	+142.7%
Net margin	~6.0%	~7.9%	~15.9%	—	—
Wafer shipments (M, 8" eq)	2.39	2.51	2.87	+20.1%	+14.4%
Utilization	92.5%	93.1%	93.7%	+1.2pp	+0.6pp
Monthly capacity (M, 8" eq)	—	1.078	1.097	—	+1.7%
Capex ($M)	—	1,563	1,836	—	+17.5%

The Margin Story — What Actually Matters

Depreciation grew +11.4% QoQ while revenue grew +20%. When a foundry runs near full and adds capacity slower than revenue, the fixed-cost base (depreciation is roughly a third of foundry cost of goods sold) spreads thinner. That is structural operating leverage, not a one-off. Net margin went from ~6% a year ago to ~16% this quarter — a full 10 points of operating leverage in twelve months.

Balance Sheet & Returns

• Debt/equity: 0.37, with net debt/equity of just 0.03 — effectively net cash on the balance sheet.
• Debt/EBITDA: ~2.6x (net debt/EBITDA ~0.23x) — leverage is a non-issue.
• Current ratio: 1.87 — ample liquidity.
• Returns: ROE ~4.2%, ROIC ~2.3% (trailing). This is the crux of the bear case: SMIC earns very little on its invested capital because it is still in a heavy reinvestment phase.

Valuation

• H-share price: ~HK$68.40 (Aug 13 close), rallying on results.
• Trailing P/E (TTM): ~113x — rich in absolute terms, and the single biggest bear argument.
• Analyst consensus: Buy; 12-month target ~HK$86.80 (~+32% upside).
• The honest read: you are not paying for the current P&L. You are paying for (a) the only Chinese foundry with 7nm-class logic at volume, (b) the China-for-China localization trade, and (c) state backing (Big Fund now the third-largest holder). It is a strategic-asset multiple, not a value multiple.

4. Q3 Guidance — Deceleration That Isn't Weakness

5. Risk Factors (Next 12 Months)

• Capex intensity: $1.84B in Q2 ≈ 61% of revenue and rising. Heavy depreciation will structurally cap gross-margin upside — you cannot run 30%+ GM while reinvesting this aggressively.
• US export controls / entity-list risk: still caps equipment access for leading-edge nodes. The 7nm story is real but fragile.
• China demand cyclicality: ~61% consumer + smartphone exposure means a domestic demand pullback hits hard.
• Valuation: at ~113x trailing earnings and ~4% ROE, any growth stumble gets punished brutally.
• Geopolitics: escalation of US-China tech restrictions would hit both demand (export displacement) and supply (tooling).

6. Catalysts (Next 12 Months)

• AI spillover broadening: demand is migrating from memory into logic, analog, power, and MCU — the second leg of an AI cycle, still early.
• Mature-node pricing power: "panic" over mature-node capacity crunch is driving orders and price hikes back to Chinese foundries.
• SMNC consolidation: the ~$6B acquisition of the remaining 49% of SMNC consolidates scale.
• Leading-edge wins: Huawei advanced-node orders + Big Fund backing reinforce the strategic-asset narrative.
• Margin line of sight: 25.3% → guided 26–28% is now explicit, not aspirational.

7. Competitive Landscape

• TSMC (TSM): global foundry leader, ~60%+ share, monopolizes leading edge (3nm/2nm). The benchmark SMIC is measured against.
• Samsung Foundry: #2 in leading-edge capacity.
• Hua Hong (1347.HK): China #2, mostly mature nodes; Q2 record revenue $717.5M, utilization 102.8%.
• GlobalFoundries (GFS), UMC (UMC): mature-node peers globally reallocating capacity toward AI/HBM, creating the mature-node shortage SMIC benefits from.
• SMIC share: ~5.3% of global foundry revenue (Q4 2025, Counterpoint) — small globally, dominant domestically.

8. Related Tickers

• TSM (TSMC): the global leading-edge benchmark; SMIC's valuation is partly a China-discount/premium trade against TSM.
• 1347.HK (Hua Hong): the direct domestic peer — reported the same day, same AI-driven upcycle, utilization over 100%.
• SMICY (ADR) / 688981.SH (STAR A-shares): the same company's ADR and A-share listings; A-shares trade at a large premium to H-shares.
• UMC, GFS: mature-node foundry peers whose AI/HBM reallocation is driving orders back to Chinese fabs.

9. Investment Thesis

Bottom Line

Strong quarter, and the strength is in the right place — margin, not just volume. The market is reading it correctly: SMIC is repricing from "cyclical mature-node foundry" toward "strategic leading-edge asset." The risk is that this premium is already largely in the price, and the capex/depreciation math means gross-margin upside has a hard ceiling around the high-20s for now. Treat it as a momentum + scarcity story with a cyclical tail risk, not a valuation story.