China's OSAT Firms Ride AI Demand Price Hikes, but Capex Surge and Earnings Quality Raise Concerns
nashnova research
All five listed Chinese OSAT firms posted sequential revenue growth in Q3, entering the first industry-wide price-hike cycle since the pandemic chip shortage — yet elevated non-recurring gains and surging capex are putting real earnings quality under scrutiny.
An industry-wide price hike — where is it coming from?
Morgan Stanley analysts report that ASE (日月光投控) and King Yuan (京元電子) have raised quotes by 10% to 30%, with hikes spreading to second-tier firms. This means → it is not a company-specific move but a sector-wide repricing cycle, the first since pandemic-era chip shortages.
CITIC Securities attributes the hikes to three forces converging: stronger-than-expected customer restocking, surging metal prices, and years of razor-thin margins building up cost pressure.
How big is the cost squeeze? In 2025, gold futures rose over 50%, silver over 150%, and copper over 50%. In plain terms = packaging uses large quantities of metals — with input costs up this much, holding prices flat means losing money.
How much is AI actually pulling demand?
Yole Group estimates the advanced packaging market reached $55 billion in 2025 and will exceed $120 billion by 2031 — more than doubling in six years.
Shanxi Securities analysts Chen Yulu and Zhang Tian note that global advanced packaging capacity is tight, with mainland China facing a "definitive gap." This means → it is not a question of "might fall short" — current capacity plans simply do not cover projected demand.
Soochow Securities analyst Chen Haijin reports that overseas clients are accelerating order and capacity shifts to mainland China, with leading chip designers evaluating a "China-for-China" strategy. This reflects a geopolitical reshaping of the OSAT supply chain.
JCET's profit surged — what is behind the numbers?
JCET — TrendForce's third-largest global OSAT firm in 2024 — posted Q3 revenue of RMB 11.43 billion, up 13.53% year-on-year. Net profit hit RMB 773 million, up 59.99% YoY. Computing electronics (including AI infrastructure) accounted for 30.9% of nine-month revenue, up 54% YoY.
But the effective tax rate dropped from roughly 19.5% to about 8.6% over the first three quarters. In plain terms = a sharply lower tax bill propped up net profit — strip that benefit out, and the growth rate looks less impressive.
The board has raised its 2026 investment plan from RMB 11.01 billion to RMB 18.2 billion. Nine-month long-term asset spending reached RMB 8.9 billion — roughly twice operating cash flow. This means → the company is spending far beyond what it generates internally; expansion relies on external financing.
TongFu's profit tripled — how much was "real"?
TongFu Microelectronics — the fourth-largest global OSAT firm — posted Q3 revenue of RMB 8.84 billion, up 24.95% YoY. Net profit reached RMB 1.34 billion, nearly tripling.
The critical figure: non-recurring items — mainly fair-value gains on financial assets — totaled RMB 713 million, or about 53% of net profit. In plain terms = more than half the profit came not from packaging chips but from investments rising in value — gains that may not repeat next quarter.
Stripping those out, recurring net profit was RMB 630 million, up 76.30% YoY. Huachuang Securities analyst Yang Yue notes that TongFu is AMD's primary OSAT partner, with Suzhou and Penang plants handling CPU and GPU work.
Smaller firms see volatile profits — could consumer demand drag them down?
Hua Tian Technology — ranked sixth by TrendForce — grew Q3 revenue 36.13% YoY to RMB 6.26 billion, but net profit fell 9.19% YoY to RMB 287 million. The swing factor: non-recurring items flipped from a gain of roughly RMB 488 million in Q2 to a loss of RMB 106 million in Q3. Excluding those, recurring profit surged 230.82% YoY — a clear operating improvement.
Consumer-side risk is building: IDC preliminary data shows Q3 global smartphone shipments fell 22.0% YoY — the steepest single-quarter decline on record.
Omdia chief analyst Hou Haijin says memory prices peaked in Q3 and cost pressure is passing through to end products. This means → AI demand is lifting high-end packaging, but weakening consumer electronics could erode the order base for mid-tier OSAT firms.
What does this expansion cycle ultimately need to prove?
The sector sits at the intersection of AI-driven capacity expansion and a price-hike cycle. Beneath the volume-and-price headline, actual profit quality varies widely.
Non-recurring gains and tax-rate windfalls have meaningfully inflated reported earnings — peel those layers away, and some firms' underlying profitability looks far less striking than the headline numbers suggest.
The critical test over the next one to two years: can rapid capex expansion be absorbed by sustained high utilization? Put simply = firms are spending heavily to build new lines — if orders do not keep pace and utilization drops, today's expansion becomes tomorrow's depreciation burden.
市场有风险，内容仅供研究参考，不构成投资建议。