SMIC and Hua Hong Both Hit Record Q2 Revenue; OSAT Profit Pass-Through Enters Verification Phase
Nashnova编辑部
SMIC and Hua Hong posted record quarterly revenue simultaneously in Q2, giving downstream OSAT firms a dual-foundry demand signal — but between orders and profits, four gates still stand in the way.
Two foundries hitting records at once — what does that tell us?
SMIC Q2 revenue: +20% QoQ, +36% YoY. Hua Hong: +9% QoQ, +27% YoY. Both are all-time quarterly highs.
This means → the upstream strength is broad-based, not one company landing a one-off order.
Q3 guidance holds up: SMIC implies ~3% QoQ growth, Hua Hong ~8%. In plain terms = Q2 is probably not a one-time peak — upstream momentum has at least one more quarter of runway.
How does foundry strength reach OSAT firms — and why the lag?
After wafers are fabricated, they go through customer sign-off, logistics, and production scheduling before entering packaging and testing. The typical lag is one to two quarters.
This means → the Q2 foundry boom maps to an order and utilization improvement window for OSAT firms in Q3–Q4.
But "more orders" does not automatically equal "better profits" — four gates stand between the two.
Orders to profits — what are the four gates?
Gate 1 — Demand: order volume must rise before fixed costs can be spread.
Gate 2 — Pricing: if the new orders are all low-priced, revenue grows but per-unit profit may not.
Gate 3 — Mix: the share of AI, HPC, and advanced packaging — assembling chips in more complex, higher-value configurations — must increase to lift average revenue per unit.
Gate 4 — Utilization: new equipment starts depreciating the moment it is capitalized; only when production volume climbs enough to cover that depreciation does profit truly improve. In plain terms = all four gates need to move up together for margins to follow the upstream wave.
JCET: is pricing power already showing up?
Q2 adjusted net profit midpoint: RMB 560 million, +112% QoQ, +28% YoY — profit growth far exceeds what simple capacity expansion can explain.
This means → at least two of pricing, product mix, and utilization are improving simultaneously. This is not just "more volume, more money."
Full-year capex budget: ~RMB 10 billion. The Lingang advanced OSAT project totals RMB 7.8 billion. This reflects a deliberate push into higher-value packaging — but the bigger the investment, the higher the bar for follow-on orders and cash returns.
The most telling metrics over the next two quarters: whether gross margin, adjusted profit margin, and operating cash flow improve in lockstep.
Weicea Tech: why does revenue jump 63% while profit rises only 23%?
Q2 revenue: +63% YoY. Net profit: +23% YoY. Revenue growth far outpaces profit growth.
In plain terms = the equipment has been purchased and depreciation has begun, but large-scale orders haven't arrived yet — a textbook case of "depreciation leads, the profit inflection follows."
High-end automatic test equipment — expensive machines that verify whether chips meet spec — has long delivery cycles. Customer projects require engineering validation and pilot runs, so the company must buy and commission gear quarters before volume orders arrive.
Analysts expect domestic AI chip volume ramps to begin boosting revenue and gross margin more visibly from Q4. This means → Q2–Q3 margin improvement will be limited; the real inflection is one to two quarters away.
The whole industry is expanding — what's the biggest risk?
OSAT capex is running hot: JCET ~RMB 10 bn annual budget, TongFu Microelectronics private placement up to RMB 4.4 bn, SiCarrier IPO project RMB 11.4 bn, Yongsi Electronics Phase II RMB 11.1 bn + convertible bonds RMB 1.46 bn. These figures use different accounting bases and should not be summed, but the direction is unanimous: 2.5D/3D packaging, heterogeneous integration — combining chips of different functions like building blocks — and high-end testing.
Capex payoff comes in three stages: cash burn during construction → depreciation pressure at early ramp → profit acceleration only after stable volume production, if orders and pricing are strong enough.
This means → the biggest risk from an industry-wide build-out is supply growth outpacing demand from 2027 onward. If customer qualifications and domestic AI chip shipments fall short, depreciation and engineering costs will squeeze margins simultaneously.
Upstream foundry strength is a leading indicator, not an unconditional promise of OSAT profitability. Q3–Q4 utilization, gross margin, and operating cash flow are the proving window for whether this pass-through actually lands.
Content is for reference only, not financial advice.