Semiconductor Packaging and Testing

What packaging and testing does

A finished wafer from a foundry is not yet a sellable product. An outsourced semiconductor assembly and test (OSAT) firm cuts the wafer into individual dies, mounts and connects each die, seals it in a package, and tests the finished part. JCET describes its own service as a one-stop finished-chip service covering microsystem integration, design simulation, wafer probing, chip and device packaging, final testing, product certification and direct shipping (2026 interim report, p. 7). Its filing lays out the chain position plainly: design companies hand a design to a wafer maker, the wafer maker produces the wafer, and the OSAT firm packages and tests the chip before the customer sells it into a finished device (p. 9).

The older techniques are wire bonding and leadframes. The newer ones are flip-chip, wafer-level packaging, fan-out, system-in-package (SiP) and 2.5D/3D integration, where several dies are packed into one device. JCET lists wafer-level packaging, 2.5D/3D packaging, SiP, flip-chip and wire bonding among its technologies (p. 7), and Tongfu Microelectronics lists 2D+ multi-die stacking, silicon interposers and embedded bridges under its VISionS platform (2026 interim report, p. 12).

Why it matters now

Packaging used to be the cheap, trailing step. It now decides how much performance a system gets out of its chips: chiplets from different process nodes are combined in one package, memory is stacked next to logic, and high-bandwidth links run through the package rather than the circuit board. Yole Group figures quoted in JCET filing put the global advanced packaging market at about US$53.1bn in 2025, rising to about US$62bn in 2026, with wafer-level packaging passing half of the total for the first time, and above US$79bn by 2030 (JCET p. 9).

That is why the stage attracts competitors from both sides: JCET warns that wafer foundries and integrated device makers are accelerating their own advanced packaging work, while new capacity in traditional packaging sharpens price competition (JCET p. 22).

Where China stands

TrendForce 2024 revenue ranking, reproduced by Evertiq, puts Taiwan ASE first at US$18.54bn, nearly 45% of the top ten combined, followed by US-based Amkor at US$6.32bn, then JCET at US$5.00bn, up 19.3%, and Tongfu Microelectronics fourth at US$3.32bn, up 5.6% (TrendForce top ten OSAT ranking, 2024). The top ten totalled US$41.56bn, up 3%.

JCET filing cites ChipInsights data showing 2025 global OSAT revenue reached a record ¥333.2bn, with the top three firms holding more than 52% of it, and JCET remaining third in the world and first in mainland China (JCET p. 9). The Chinese position is therefore strong in scale but second-tier in the ranking: two Chinese firms in the top four, none at the top, and the leader roughly three times JCET size on 2024 revenue.

Economics

OSAT margins are thin. JCET first-half 2026 revenue of ¥19.53bn against cost of sales of ¥16.57bn implies a gross margin of about 15.2% (our calculation from JCET pp. 6, 13). Tongfu packaging and testing segment earned a 15.12% gross margin in the same half, up 0.73 points (Tongfu p. 18). Both firms are shifting mix toward higher-value work: JCET computing revenue rose 40.4% and auto revenue 25.0% in the half (JCET p. 8), while Tongfu says its mid-to-high-end product revenue rose markedly and its advanced packaging share kept climbing (Tongfu pp. 12, 17).

What to watch

• Advanced packaging share. JCET XDFOI chiplet platform is in volume production and Tongfu VISionS platform is shipping into high-performance computing and inference chips. How much of each firm revenue comes from advanced packaging, which neither filing breaks out cleanly, is the number that decides whether margins widen.
• Capacity races. JCET approved a ¥7.8bn high-end advanced packaging plant in Shanghai Lingang (JCET p. 16), and Tongfu is expanding across Nantong, Suzhou, Penang, Hefei and Xiamen (Tongfu p. 13).
• Customer concentration. Tongfu says it is AMD main OSAT supplier, covering more than 80% of AMD relevant products (Tongfu p. 14). A single-customer shock is the clearest downside for the number four firm.

Our view: China OSAT strength is scale, cost and domestic demand, not the leading edge. The gap to ASE is a technology and customer gap in advanced packaging, and the two Chinese firms are spending to close it. Whether that spending lifts margins above the mid-teens, or just buys share in a price war, is the question for the next two years.