JCET is the largest mainland OSAT and the name that comes up first on almost any advanced packaging shortlist in China. This piece reads its H1 2026 half-year report rather than commentary about it, because the mix shift inside those numbers is what actually tells you which programmes the line is tuned for.
The headline is a margin story, not a volume story. Revenue grew about 5%; profit grew about 79%. That gap is the whole point.
The half-year numbers
For the first half of 2026 JCET reported revenue of 19.53 billion RMB, up about 5.0% year-on-year and described as a record for the period. Net profit attributable to shareholders was 845 million RMB, up 79.4%, and ex-items net profit was 808 million RMB, up 84.7%. Second-quarter revenue was 10.36 billion RMB, up 11.7% year-on-year and 12.9% sequentially, with attributable net profit of 554 million RMB, up 107.3%.
Operating cash flow was 2.96 billion RMB, up 26.8%. Reported gross margin in secondary coverage is around 15%, with the second quarter cited near 15.7% — low by semiconductor standards, which is what an assembly-and-test model looks like.
The mix shift is the real signal
Three segment figures explain the profit growth:
- Computing electronics revenue grew 40.4% year-on-year and became the largest revenue source at 30.1% of the total
- Automotive electronics revenue grew 25.0% to 2.16 billion RMB, or 11.1% of revenue
- Test revenue grew 9.4%
Why profit moved faster than revenue
A five-point revenue increase producing a seventy-nine-point profit increase only happens when the revenue that grew carries a much better margin than the revenue that did not. The practical reading for a programme is straightforward: the line is being tuned around computing and automotive work, and other work competes for attention against that.
Technology platform status
JCET's self-developed XDFOI chiplet high-density heterogeneous integration platform is stated in the half-year report as having reached volume production. On co-packaged optics, the company states that a silicon photonics engine developed on that platform has completed customer sample delivery and verification.
On FC-BGA, the company cites years of work and engineering and volume production capability for very large body size and very thin profiles. On storage, it covers DRAM and NAND Flash with 32-layer stacking and hybrid heterogeneous stacking capability.
Note the wording carefully. Volume production and sample delivery are different commitments, and the company's own phrasing distinguishes them. Ask which category your route falls in.
Capacity and capital
JCET announced a 7.8 billion RMB investment in a high-end advanced packaging and test facility in Shanghai Lingang, with a wholly owned subsidiary established as the project entity and registered capital of 4.0 billion RMB. Full-year 2026 fixed-asset investment is budgeted at roughly 10 billion RMB, directed mainly at advanced packaging, computing and automotive electronics.
The Lingang automotive electronics facility entered production in March 2026 and reached volume production in the second quarter. First-half R&D spend was 1.03 billion RMB, about 5.3% of revenue.
Capacity under construction is not capacity available today. If your programme depends on a line that is still being built, treat that as a schedule input, not a capability statement.
Which programmes fit JCET
JCET is the right first call when the package route is genuinely undecided — when you know you need heterogeneous integration but not yet whether that lands on 2.5D, fan-out or a stacked route. Breadth is the asset.
It is a weaker fit when your programme is small and undifferentiated. Broad capability attracts broad demand, and queue position follows commercial weight. Ask what support model applies at your volume band and get it in writing.
