JCET In-Depth Research

That starting point deeply shaped later strategic choices. A local manufacturing background meant natural emphasis on process, production lines, cost, yield, and delivery, rather than earning high gross margins from licensing or platform fees. The earliest problem it solved was also plain: making back-end chip manufacturing cheaper, more stable, and larger in scale. This kind of company suffers the double pressure of utilization and depreciation in downturns, but when the industry rises and customers outsource more products, operating leverage can quickly amplify profits. JCET was never a "light-asset technology story" from the start, and that has not changed in more than 20 years.

The First Ceiling After Listing

When JCET listed, the way capital markets understood it was almost like a different company from the post-STATS ChipPAC JCET. In its early phase, it was essentially still a scale player in mainland China's OSAT industry, expanding through local manufacturing cost advantages, product iteration, and growth in domestic demand. After the 2000s, as global IDMs and fabless companies increased outsourcing of packaging and testing, the OSAT industry began a genuine global reshuffle. JCET realized that relying only on the domestic market and traditional packaging would make it hard to enter the industry's first tier. That judgment directly led to the cross-border acquisition that changed the company's fate.

The STATS ChipPAC Acquisition Changed the Company's Ceiling

In 2015, JCET, together with the National Integrated Circuit Industry Investment Fund and SMIC Semiconductor, privatized Singapore-listed STATS ChipPAC through a voluntary conditional general offer. The equity transaction consideration was about USD 780 million, corresponding to enterprise value of about USD 1.8 billion. Shanghai Stock Exchange case materials describe the transaction as the first cross-border acquisition of an internationally first-class packaging and testing company by a Chinese A-share listed company, and also the world's largest control-rights acquisition in the integrated-circuit packaging and testing industry over the prior five years. More importantly, this was a transaction to buy customers, process know-how, brand, overseas delivery capability, and global organizational experience, far more than a simple asset purchase.

After the acquisition, JCET's geographic footprint and customer structure changed. It formed a global capacity layout coordinated across China, Singapore, South Korea, and other locations, no longer simply "doing packaging and testing in China". The transaction background materials particularly stressed that STATS ChipPAC had key patents and a mature R&D ecosystem in advanced packaging, fan-out technology, and flip-chip bumping. For JCET, this mattered more than simply adding revenue, because it moved the company into a higher tier of customer relationships and technology tracks. The market's willingness today to believe that JCET can participate in 2.5D/3D, Chiplet, and CPO rests on this acquisition, which transformed it from a domestic champion into a global player.

After the Acquisition Came Long Digestion, Then a Cycle Dividend

The acquisition was not the endpoint. The truly difficult part came in the following years. JCET's net profit attributable to shareholders was only RMB 89 million in 2019; it rose to RMB 1.304 billion in 2020, further jumped to RMB 2.959 billion in 2021, reached RMB 3.231 billion in 2022, then fell to RMB 1.471 billion in the 2023 industry downturn, recovered to RMB 1.610 billion in 2024, and fell again to RMB 1.565 billion in 2025. Revenue followed another path: RMB 23.526 billion in 2019, RMB 26.464 billion in 2020, RMB 30.502 billion in 2021, RMB 33.762 billion in 2022, RMB 29.661 billion in 2023, RMB 35.962 billion in 2024, and RMB 38.871 billion in 2025. This sequence is representative: globalization and platformization of revenue succeeded, but profit remained dragged by the cycle, utilization, product mix, and depreciation.

This is also the most important key to understanding JCET. It has proved that it can expand revenue and win major customers, but it has not proved that it can steadily and high-quality crystallize the advanced-packaging dividend into margins the way TSMC can. 2022 was one of the best years for the cycle and efficiency, while 2025 was the highest-revenue year, but profits did not match. For a manufacturing company that talks about AI upside, this mismatch between revenue and profit is itself the most valuable research point.

Longitudinal Financial Review

One strength of JCET over the past five years is that operating cash flow has long been better than accounting net profit. On a 2021–2025 five-year cumulative basis, operating cash flow was about RMB 28.364 billion, cumulative net profit attributable to shareholders was about RMB 10.836 billion, and OCF/net profit was about 2.62x. This shows that it is not a manufacturing company with "good-looking profit but no cash collection". Operating cash flow in 2025 alone was RMB 4.652 billion, still materially above net profit attributable to shareholders of RMB 1.565 billion. The problem lies on the other side: capex is again swallowing that cash. Cash paid in 2025 for the purchase and construction of fixed assets, intangible assets, and other long-term assets was RMB 6.298 billion, already higher than operating cash flow; management then raised the 2026 fixed-asset investment budget to about RMB 10 billion. Cash creation exists, but free cash flow is being consumed by advanced-packaging expansion.

Another clue that needs separate attention is depreciation. In 2025, fixed-asset depreciation and various amortization items totaled about RMB 4.14 billion, while cash paid for long-term assets that year had already reached RMB 6.30 billion. In other words, new capital investment was materially higher than depreciation recovered from existing assets, and the increase in depreciation is not over. This means that even if revenue keeps growing in 2026–2027, the income statement will face a thicker layer of depreciation expense. For companies like JCET, judging profit trends cannot rely only on revenue and orders; it must also compare the relative speed of new-line utilization ramp-up and depreciation ramp-up.

The balance sheet itself is not fragile. At the end of 2025, total assets were RMB 55.517 billion and equity attributable to shareholders was RMB 28.671 billion; shareholder equity kept growing, and operating cash flow did not deteriorate into negative territory. The real pressure is capital allocation, not solvency. When a company commits a RMB 10 billion 2026 budget to production lines, its valuation must rest on what ASP, gross margin, and customer lock-in that money will bring. If the money only upgrades "technology existence" into "mass-production capability" but cannot quickly convert into orders and profit, valuation will retrace before fundamentals catch up.

Share-Price and Valuation History

JCET's capital-market history can be roughly divided into four stages. The first ran from listing to around 2014, when the market mainly treated it as a domestic OSAT leader and valued it with industry prosperity and domestic substitution. The second began with the 2015 STATS ChipPAC acquisition, which lifted the story from "local champion" to "global platform" and gave the market more room for imagination. The third was 2019–2022, when the industry upcycle and company efficiency improvement overlapped, and profit release drove valuation expansion. The fourth is now, when the market has begun adding a premium for "domestic AI advanced-packaging positioning".

The real warning sign is that the valuation label has changed, but the profit label has not fully caught up. Today's JCET is more often traded as an "advanced-packaging platform stock", but 2025 net profit attributable to shareholders did not set a new high alongside record revenue, and recurring profit fell by double digits. The shift in valuation center reflects not only a change in business quality but also a clear migration in market preference: capital markets are willing to pay upfront for the scarcity expectation around China's advanced packaging. As long as advanced-packaging revenue and orders have not been fully quantified, that premium will be more fragile than traditional OSAT cycle valuation.

Sources: 2021 annual report, 2022 annual report summary, 2025 annual report, 2026 Q1 report; author compilation.

Business Model and Industry Cycle Analysis

Revenue Structure and Profit Engine

In its financial-report segments, JCET does not break "advanced-packaging revenue" and "traditional-packaging revenue" into a directly readable table. This is one of the most important information gaps in this case. The company discloses its main business category as "electronic components"; more useful structural views come from applications and unit shipments. In 2025, communications electronics accounted for 31.4% of revenue, consumer electronics 24.6%, computing electronics 12.8%, industrial and medical electronics 11.7%, and automotive electronics 17.0%. In other words, automotive, computing, industrial, and medical together had already reached 41.5% of 2025 revenue, materially higher than the prior year. This shows that the company is actively shifting its revenue center of gravity away from traditional consumer and communications markets toward compute, automotive, and industrial scenarios.

But on a unit-shipment basis, the story narrows considerably. In 2025, traditional-packaging shipments were 192.13 billion units, up 14.5% year on year; advanced-packaging shipments were 20.09 billion units, up 38.5%; testing business shipments were 14.49 billion units, up 5.8%. By packaged-unit count, advanced packaging accounted for only about 9.5% of total packaging units. This of course cannot directly equal revenue share, because advanced packaging has much higher ASP and value added than traditional packaging. Still, it is enough to show that in a revenue base close to RMB 39.0 billion, advanced packaging looks more like the segment with the strongest marginal pull, not the segment that already covers the whole business. Investors who directly equate "advanced-packaging leader" with "overall profit will become highly elastic" are jumping too fast.

The customer structure is more diversified than many imagine. In 2025, the top five customers accounted for 31.34% of sales, with the largest customer at 10.63%. This supports two judgments: first, JCET is indeed not a single-customer business and is more resilient to customer volatility than many foundries; second, even if a certain type of AI customer is being introduced, it has not yet formed a "single customer drives the whole statement" state in disclosures. This data is very important for studying "domestic AI packaging absorption", because it reminds us not to infer from a few thematic headlines that the whole company is already highly tied to one AI product line.

Cost Structure, Operating Leverage, and Moat

JCET's operating leverage is very heavy. Plants, equipment, process development, quality certification, global supply chain, and upfront customer-introduction costs mean it cannot travel lightly like a software company. Once mature factories are fully loaded, profit can rise quickly; once orders are insufficient and depreciation rises, fixed costs can quickly swallow profit. Q1 2026 profit repair was exactly this mechanism at work: the company publicly stated that mature factories stayed fully loaded, utilization remained high, and it continued to advance advanced-packaging and automotive-electronics layouts, further optimizing product mix, which allowed a sharp profit increase despite a slight year-on-year decline in revenue.

The moat that truly holds, in my view, has three layers. The first is the global manufacturing and delivery network. The Singapore, South Korea, and China capacity left by the STATS ChipPAC acquisition is not just geographic diversification; it is a real threshold across different customers, processes, and validation systems. The second is integrated collaboration across design, packaging, and testing. The value of high-density heterogeneous integration platforms such as XDFOI lies in binding customer design, packaging integration, and testing into a longer development cycle, not in a promotional term. The third is major-customer qualification and mass-production validation. The strength of an OSAT has always been whether international customers are willing to entrust high-ASP, high-yield-requirement chips to it for repeated mass production, not how many technology terms appear in a presentation. JCET disclosed in 2023 that its XDFOI Chiplet series process had entered stable mass production, had shipped multi-chip system-in-package products at the 4nm node for international customers, and had achieved about 2μm minimum line width and spacing on organic redistribution interposers. These are real thresholds.

The moat that does not really hold is treating the "advanced-packaging concept" itself as a barrier. Advanced packaging is not an exclusive track for JCET. TSMC itself is the center of the profit pool in high-end packaging; ASE has already begun measuring advanced-packaging revenue in billions of U.S. dollars and publicly giving 2026 growth targets; Tongfu Microelectronics, because of its deep AMD tie, has more direct capital-market recognition in AI/HPC packaging than JCET. JCET's strengths are breadth of platform, breadth of customers, and breadth of processes. Its weakness is that AI delivery is less concrete in public disclosures than Tongfu's, and it has not broken out advanced-packaging revenue for the market the way ASE has.

Industry Structure, Cycle, Policy, and Geopolitics

OSAT is a mature industry, but advanced packaging is not. Traditional packaging and testing are still deeply affected by the semiconductor cycle through pricing, utilization, customer inventory, and end demand; advanced packaging is growing much faster under the push of AI, HPC, HBM, Chiplet, and CPO. In 2025 industry research, Yole raised its 2030 advanced-packaging market expectation to USD 83.0 billion; SEMI conference materials also viewed high-end 2.5D/3D performance packaging as one of the fastest-growing subfields from 2024 to 2030. This means the industry is splitting internally: traditional OSAT remains cyclical manufacturing, while advanced packaging is developing structural growth. JCET stands at the boundary of these two worlds.

The profit pool is still not on the OSAT side, a common-sense point that must not be forgotten. Comparing JCET's 2025 recurring net margin of about 3.5% with TSMC's Q1 2026 operating margin of 58.1% is not an apples-to-apples business comparison, but it is enough to remind investors that even when advanced packaging is hot, the center of the profit pool is still more often held by the strongest foundries and platform customers. OSATs can capture incremental growth, but they may not capture the thickest layer of profit. For JCET, the truly important question is whether it can obtain sufficiently high ASP, stable utilization, and long customer stickiness in advanced packaging, not whether it "has advanced packaging".

Policy and geopolitics are a double-edged sword for JCET. U.S. rules on advanced computing chips and AI flows introduced the AI Diffusion framework and additional due-diligence measures in January 2025; BIS later announced it would rescind the original AI Diffusion Rule and previewed replacement rules, and in early 2026 the Trump administration again adjusted license reviews for advanced AI chips to China. The rules themselves are changing. The market interprets any action that "restricts China's access to overseas advanced compute" as positive for domestic advanced packaging. That directional judgment makes sense, but the path is not smooth. For JCET, this may bring opportunities as domestic AI-chip companies become more willing to use local packaging, while also making overseas customers, overseas capacity, and cross-border compliance more complex.

Another policy clue comes from internal governance. In 2024, the CR Holdings system entered JCET through Panshi Hong Kong/Panshi Runqi, holding about 22.54%, and triggered commitments related to business overlap with the CR Micro packaging and testing system. CR has committed to resolve business overlap or potential competition with Wuxi CR Advanced Semiconductor and Jiequn Electronics within five years. This is both a plus and a constraint for JCET: the plus is a more stable funding and industrial-resource platform; the constraint is that if future integration is not smooth, capital markets will again demand a governance discount.

Peer Analysis

If the packaging and testing industry is drawn as a group portrait, JCET stands in the middle: neither among the most aggressive nor among the most conservative. ASE has become the global scale king and can already describe advanced-packaging revenue increments concretely; Amkor has become the OSAT with the strongest U.S. attribute and the clearest geographically diversified customer profile; Tongfu Microelectronics has become the Chinese packaging and testing company most like a "pure AI/HPC beta", because its AMD tie is deep enough; Huatian Technology is more like a company with a thicker traditional packaging and testing base that is also upgrading advanced packaging, but with less direct capital-market elasticity than the first two. JCET stands among these companies: more globalized than domestic peers, less explicit in AI orders than Tongfu, and less transparent than ASE in advanced-packaging revenue breakdown.

What JCET should most learn from ASE is that ASE already explains advanced packaging to the market as a separately measurable business, not as a set of technology terms. Reuters reported twice on ASE in 2026, noting that its 2026 advanced-packaging revenue target was above USD 3.5 billion and that its previous target was to double to USD 3.2 billion by 2026. This type of disclosure directly raises valuation quality in the AI cycle, because the market knows what it is buying. JCET's problem is that it has not yet clearly explained "how much revenue, gross profit, and capital return advanced packaging gives me", not that it lacks technology.

The difference between Tongfu Microelectronics and JCET lies in narrative purity. Sell-side and industry media have long defined Tongfu as one of AMD's largest packaging and testing suppliers, and some materials even state that AMD orders account for more than 80% of its orders. Even if that number needs cautious treatment, it at least shows that Tongfu's AI/HPC logic is easier for the market to grasp. In Q1 2026, Tongfu revenue rose 22.8% year on year and net profit rose 224.6%, a reporting elasticity naturally better suited to thematic trading. JCET's Q1 profit elasticity was also strong, but its revenue was down year on year, suggesting that it currently looks more like "manufacturing efficiency improvement plus structural upgrade" than a "single AI major-customer breakout".

Huatian Technology reminds investors that the mainland OSAT line has more than one company recovering. Huatian's 2025 revenue was RMB 17.214 billion, up 19.03% year on year, and net profit attributable to shareholders was RMB 711 million, up 15.30%; Q1 2026 revenue was RMB 4.80 billion, up 34.49%, also showing a clear repair. Compared with JCET, Huatian is smaller, has a heavier traditional packaging and testing base, and is weaker in global customer chains and overseas footprint, but the market has also given it high multiples. This shows that A-share pricing for the packaging and testing sector has broadly been lifted by "AI plus advanced packaging plus mainland substitution", rather than rewarding only JCET.

The true cross-boundary competitors remain TSMC and some foundries' in-house packaging operations. JCET can benefit from partial substitution demand where mainland customers cannot access TSMC-like capacity, but at the highest-end and most profitable packaging layer, it still faces the reality that foundries are moving downstream in the value chain. Researching JCET requires comparison with OSATs, but also recognition that some of the profit pools it is competing for do not naturally belong entirely to OSATs.

Sources: JCET 2025 annual report and 2026Q1 report, Google Finance, Reuters, Yahoo Finance Japan, company/media public disclosures; author calculations.

JCET's niche is therefore closer to a "globalized Chinese OSAT leader" than a "single AI packaging high-elasticity target". It fills a very specific gap: an advanced packaging and testing platform with both China domestic delivery capability and international customer validation plus an overseas capacity base. It is most directly competing for high-end order share in global OSAT. The players that could truly compete for its profit pool are not only Tongfu and Huatian, but also foundries' in-house advanced packaging and customers' captive lines. If the industry enters a price war, its resilience in traditional packaging and testing is stronger than that of smaller players; if the industry enters scarcity competition in high-end packaging, it may still not be stronger than the best foundries. That is its real position.

Current Fundamentals and Valuation Analysis

The Last Four Quarters and What the Market Is Trading

JCET's rhythm over the last four quarters shows that the profit inflection came before a major revenue inflection. In Q4 2025, revenue was RMB 10.20 billion, up 1.4% quarter on quarter, and net profit attributable to shareholders was RMB 610 million, up 26.6% quarter on quarter. In Q1 2026, revenue was RMB 9.171 billion, down 1.76% year on year, but net profit attributable to shareholders reached RMB 290 million, up 42.74% year on year, and operating cash flow was RMB 1.779 billion, up 55.44%. This looks more like the combined effect of improved mature-factory utilization, better product mix, expense control, and a higher share of some high-value-added businesses, rather than a typical "broad demand explosion".

The themes currently traded by the market are therefore not singular. The first layer is earnings recovery, the second is the AI advanced-packaging narrative, and the third is capex expansion. The 2026 fixed-asset investment budget of about RMB 10 billion, focused on advanced-packaging production lines, is enough by itself for the market to reclassify the company from an "earnings recovery stock" to a "growth investment stock". But viewed the other way, that same RMB 10 billion of capex makes valuation hard to explain only by profit recovery; capital markets are effectively betting early that the company can turn these investments into higher-margin orders over the next two to three years.

Technology also supports this expectation. JCET disclosed in 2023 that the XDFOI Chiplet process had entered stable mass production and achieved 4nm multi-chip system-level packaging shipments for international customers. Market and media coverage later continued to link the company with Chiplet, HBM, CPO, and optical-engine directions. The problem is that the public reporting lacks the most important next step: revenue share, gross-margin difference, order backlog, and profit margin after full utilization. The truth of these directions themselves is not really the dispute. Without this step, the market is still trading "believing it will happen", not "verifying that it has happened".

Bull-Bear Divide

The first bullish evidence is that structural upgrade has already shown up in the statements. In 2025, automotive electronics, computing electronics, industrial electronics, and medical electronics together rose to 41.5% of revenue, and advanced-packaging unit shipments grew 38.5% year on year, faster than traditional packaging; in Q1 2026, the company achieved 42.7% profit growth despite no revenue growth. This set of data is enough to support the judgment that advanced packaging and high-value-added markets are lifting the company's profit elasticity.

The second bullish evidence is that JCET's technology and customer thresholds are real. The globalized customer base brought by the STATS ChipPAC acquisition, plus XDFOI entering stable mass production and completing 4nm multi-chip shipments, show that JCET has the ability to truly make complex packaging work and make it stable. For an OSAT, this is more valuable than a laboratory demo.

The first bearish evidence is also very solid: in 2025, revenue hit a historical high, but net profit attributable to shareholders did not, and recurring net profit fell 11.51% year on year. If the advanced-packaging and AI narrative were already strong enough, it should have appeared first in the income statement. At least through the 2025 annual report, that step had not happened.

The second bearish evidence comes from cash flow and capex. In 2025, operating cash flow was RMB 4.652 billion, while long-term asset spending was RMB 6.298 billion, making free cash flow tight; the 2026 budget rises further to RMB 10 billion. If new-line utilization and ASP improvement are even slightly slow, depreciation and interest will pressure profits first. For capital-intensive manufacturers, the truly dangerous moment is always "the technology has been built, but cash has not yet returned", not "the technology cannot be built".

The third bearish evidence is the information structure. ASE can discuss 2026 advanced-packaging revenue targets above USD 3.5 billion, Tongfu's AMD tie is a public capital-market narrative, and Huatian at least puts its repair rhythm into quarterly reports. JCET still mainly discloses application mix, unit-shipment growth, and technology nodes, without disclosing advanced-packaging revenue share, HBM-related revenue, domestic AI customers, or order size. This makes the market willing to assign it a high valuation when optimistic, but also makes that valuation easier to take back when doubts rise.

Cash-Flow Look-Through, Peer Valuation, and Absolute Valuation

Start with cash-flow look-through. From 2021 to 2025, operating cash flow/net profit attributable to shareholders was about 2.62x, so cash collection was not poor; but cash paid in 2025 for fixed assets, intangible assets, and other long-term assets was RMB 6.298 billion, while depreciation and various amortization items were about RMB 4.14 billion. If the latter is roughly treated as a proxy for maintenance capex, 2025 owner earnings were about RMB 510 million, implying an owner-earnings P/E of about 252x at the current share price and a yield of only about 0.4%. Even using the 2021–2025 average operating cash flow to smooth the cycle, owner-earnings valuation remains far away from headline P/E. This gap is clearly above 30%, so in subsequent valuation I prioritize owner earnings and capex constraints rather than only accounting profit.

Next, relative valuation. JCET's current static P/E is about 81.6x and P/B is about 4.45x. Horizontally, A-share peers are also expensive: Tongfu is about 72.8x forward P/E, and Huatian about 66.8x P/E; global leader ASE is about 51x P/E, and Amkor about 46x P/E. The issue is that an expensive peer group does not automatically prove JCET is cheap. Instead, this looks more like a sector valuation center broadly lifted by AI and advanced-packaging themes. JCET's discount to Tongfu should be justified by less direct AI-customer elasticity; its premium to ASE/Amkor should be justified by mainland substitution and domestic scarcity. But once cash flow and capex are included, the current share price has already used those arguments heavily.

I use three yardsticks for absolute valuation: first, a medium- to long-term normalized valuation anchored on 2027 net profit attributable to shareholders and P/B; second, an owner-earnings anchor to test whether current investment can turn into real cash two years later; third, the tension between "AI advanced-packaging rerating" and "traditional packaging and testing cycle valuation", allowing a wider valuation range rather than a single point. Within this framework, I do not deny the optimistic scenario exists, but I am unwilling to use it directly as sufficient justification for the current price.

This is a valuation scenario analysis under the research framework and does not constitute investment advice. Sources: company annual reports, quarterly reports, earnings briefings, and industry comparisons; author estimates.

Expectation Gap and Margin of Safety

The market's current embedded expectation is no longer just "profit recovery" but "advanced packaging will materially rewrite the company's overall earnings structure". The three indicators most likely to create expectation gaps are utilization after advanced-packaging capacity comes online, the true share of advanced-packaging revenue in total revenue, and whether the 2026–2027 depreciation increase is covered by ASP and yield improvement. If in the next annual report or interim report the company still discloses only unit counts and application mix, without a clearer advanced-packaging revenue/gross-margin basis, the market will begin to question which business it is paying a premium for.

From a margin-of-safety perspective, the current price has no discount to conservative-case intrinsic value, so the margin of safety is zero. The most fragile assumption across the three scenarios is that "advanced-packaging capacity expansion smoothly turns into high-ASP orders". If this assumption is cut by 30%, my base-case intrinsic value would fall from RMB 52–55 to about RMB 41–46. If earnings merely stand still over the next three years, expected annualized return at the current price is left with only a very low dividend and sentiment volatility, which is not a reliable margin of safety. My conclusion on margin-of-safety adequacy is: none.

Risks, Catalysts, and Horizontal-Vertical Synthesis

Risk Analysis

Mismatch between advanced-packaging expansion and depreciation is the primary risk I consider medium probability and high impact. Capex cash outflow already reached RMB 6.298 billion in 2025, and the 2026 budget rises further to RMB 10 billion. If new-line utilization, ASP, or yield ramps slower than expected, incremental depreciation will pass through the income statement first. Observable indicators are simple: cash paid for long-term assets, depreciation and amortization, recurring net margin, and mature-factory utilization. If profit growth is materially weaker than depreciation growth for two consecutive quarters, the market will immediately reclassify it from a "growth stock" back to a "cyclical manufacturer".

Insufficient AI order landing is the second risk, with medium probability and high impact. JCET has already proved technical capabilities such as 4nm, multi-chip, 2μm line width and spacing, and the XDFOI platform, but public disclosure still lacks quantified views on advanced-packaging revenue share, HBM-related revenue, and core domestic AI accelerator-card customers. If what appears in the future is only samples, validation, and small trial orders rather than platform-level orders, the "AI beta" currently paid for in valuation will quickly evaporate. Observable indicators include separate disclosure of advanced-packaging revenue, customer certification progress, and continued improvement in computing-electronics share alongside gross margin.

Export controls and overseas-customer compliance are the third risk, with medium probability and medium-high impact. One of JCET's strengths is overseas capacity and an international customer base, but that also means it can more easily be drawn into rule changes. BIS introduced the AI Diffusion framework and due-diligence requirements in January 2025, later announced the rescission of the original rule and preparation of new rules, and in 2026 continued adjusting license-review logic. Repeated rule changes have substantive effects on overseas-customer signing rhythm, product classification, end use, and delivery paths. Observable indicators include new BIS rule texts, customers' requirements for supply-chain location and compliance documents, and product-mix changes at overseas factories.

Integration after acquisitions and control-right changes is the fourth risk, with medium probability and medium impact. JCET is digesting the acquisition of an 80% stake in SanDisk Semiconductor (Shanghai), entering a new governance cycle after CR Holdings became controlling shareholder, and handling commitments to resolve potential business overlap within the CR Micro system. Each item alone is not fatal, but if a capex peak, customer-structure shift, and internal integration all occur in 2026–2027, execution complexity will rise markedly. Observable indicators include acquired-asset profit delivery, progress in resolving horizontal competition, whether related-party transactions rise, and board and senior-management stability.

Catalysts and Tracking Dashboard

Positive catalysts are clear, and the key is whether there is a leap "from concept to data". The strongest positive catalyst would be the company's first disclosure of advanced-packaging revenue share, advanced-packaging gross margin, or a major domestic AI-chip customer entering mass production, rather than another mention of XDFOI. Second, if H2 2026 shows a combination of continued revenue growth and a declining capex/operating-cash-flow ratio, the market will be more willing to believe this expansion does not only bring depreciation. Further out, if post-CR industrial integration gives a clearer path for customer and resource collaboration, valuation quality would also improve.

Negative catalysts are equally straightforward. If revenue grows for two consecutive quarters while margin falls instead of rising, the market will question the quality of new capacity; if operating cash flow cannot cover a meaningful share of capex in 2026–2027, valuation will be compressed first; if domestic AI packaging absorption remains stuck at "technology reserve" and "sample bring-up" without order and revenue support, thematic capital will withdraw quickly.

For these indicators, the most important issue is the combination, not any single number. If automotive, computing, and industrial/medical share rises without improving recurring net margin, the quality of the structural upgrade is insufficient; if capex stays high and OCF/net profit stays low, expansion is eating shareholder returns. The public sources most worth tracking remain the company's annual reports, quarterly reports, earnings briefing records, and the SSE roadshow platform.

Horizontal-Vertical Synthesis

Placed back on its own timeline, JCET has truly proved two capabilities. The first is the ability to break through ceilings. Starting from old Jiangyin electronics-industry assets, it reached top-three global OSAT status by daring to do two difficult things at critical moments, rather than merely telling stories: first, building scale; second, swallowing STATS ChipPAC through a highly complex transaction structure in 2015, transforming itself from a regional manufacturer into a global manufacturing platform. The second is the ability to stand back up after cycles. When profit was only RMB 89 million in 2019, many viewed it as a post-acquisition-syndrome company; in 2021–2022, it used profit recovery to prove that in OSAT, as long as utilization and structure are positioned right, operating leverage remains large.

But placed back into the industry cross-section, JCET's real advantage today is a more practical combination: global customer qualification, international manufacturing layout, advanced-process track record, and scarcity of Chinese domestic supply. It is not the imagined absolute moat of being "irreplaceable". These advantages are enough to make it one of the most important mainland candidates for an advanced-packaging platform, but not enough to let it capture the industry's thickest profit layer as steadily as a top foundry. Its weakness is a gap between disclosure basis and capital-market expectation, not technological backwardness: the market is already pricing it as a "domestic AI advanced-packaging platform", while the financial statements still speak mainly through application mix, shipment counts, and the overall income statement.

This determines the nature of the current valuation. It is more a case of bringing the future forward than simply rewarding past success. The 2025 annual report and 2026 Q1 report both contain optimistic signals: the structure is changing, profit is repairing, and advanced-packaging expansion is very determined. Yet the same documents also contain harder constraints: attributable and recurring profit did not keep pace with record revenue, free cash flow is being consumed by capex, and advanced-packaging revenue share has not been quantified. In manufacturing investment, such contradictions cannot be waved away with the phrase "good track". The point capital markets most easily misjudge is exactly this: mistaking the business with the greatest marginal change for the business that already dominates the overall income statement.

The most important variable over the next year is new-line utilization and cash flow, not technology releases; over the next three years, it is whether advanced-packaging revenue share and gross margin finally enter public disclosure; over the next five years, it is whether JCET can turn "scarce mainland advanced-packaging asset" into a genuinely higher-ROE, steadier-free-cash-flow business. If the company begins to steadily disclose advanced-packaging revenue/profit metrics, operating cash flow again covers most capex, and CPO/HBM/Chiplet orders no longer stop at the sample and validation stage, it will become a better investment target. Conversely, if 2026–2027 repeatedly show a combination of high capex, higher depreciation, flat margins, and still-opaque AI share, the core judgment of this research should be overturned, or at least the valuation framework must be revised down overall.

Bull and Bear Arguments

• Bull argument one: JCET has grown from a Chinese domestic OSAT into a top-three global player; the China-Singapore-South Korea footprint and international customer chain left by the STATS ChipPAC acquisition are among the hardest assets for mainland peers to replicate.

• Bull argument two: In 2025, automotive, computing, industrial, and medical electronics revenue share rose to 41.5%, while advanced-packaging units grew 38.5% year on year, showing that high-value-added businesses are indeed accelerating.

• Bull argument three: In Q1 2026, revenue edged down year on year but attributable net profit rose 42.74% and operating cash flow rose 55.44%, showing that utilization and product-mix improvement have begun to release profit elasticity.

• Bull argument four: After CR Holdings' entry, the company's capital platform is more stable, and it has made a five-year commitment to resolve potential business overlap with the CR Micro system.

• Bear argument one: 2025 revenue set a historical high, but net profit attributable to shareholders fell 2.75% year on year and recurring net profit fell 11.51%; the AI narrative has not yet fully penetrated the income statement.

• Bear argument two: 2025 long-term asset cash outflow was RMB 6.298 billion, and the 2026 budget is about RMB 10 billion, equivalent to 6.4x 2025 net profit attributable to shareholders, creating very heavy free-cash-flow pressure over the next two years.

• Bear argument three: Public disclosures include technology nodes, application mix, and unit shipments, but do not break out advanced-packaging revenue share, HBM-related revenue, domestic AI major customers, or order backlog size.

• Bear argument four: Based on 2025 figures, the current share price implies a static P/E of about 81.6x and P/B of about 4.45x, while 2025 ROE was still in the mid-single digits, meaning valuation has clearly brought future delivery forward.

Pre-mortem

The first scenario most likely to make me lose 50% after three years is "capacity expansion comes first, orders arrive later". Suppose JCET invests close to RMB 10 billion-level capex in 2026–2027 as planned, but main orders in 2.5D/3D, HBM-related, and CPO directions land slower than expected, new-line utilization remains below 60% for a long time, incremental depreciation rises another RMB 1.0–1.5 billion per year, and recurring net margin is pushed back below 3% from around 3.5% in 2025. At that point, the market would reclassify it from an advanced-packaging growth stock back into a traditional cyclical OSAT, compressing valuation from nearly 80x static P/E today to 30–40x. A share price back in the RMB 35–40 range would not be surprising.

The second scenario is "the main AI orders do not land in the layer it wants most". TSMC continues to keep the highest-end packaging profits inside the foundry system, Tongfu takes more direct GPU/HPC incremental volume through its AMD tie, and JCET wins some optical-engine, Chiplet, automotive-compute, and storage-related orders, but these are more edge increments than core platform orders. Revenue could still grow in 2027, but growth would come from broader cyclical repair rather than a high-margin AI packaging breakout. Capital markets would discover that the premium they paid over the past two years was too early. P/B falling back to 2.5–3x and the share price returning to RMB 40–50 could still produce losses close to a halving.

Final Research Conclusion

JCET is a company very worth studying, but not a stock suited to a one-line narrative bet. Its most valuable feature is that, in China's semiconductor manufacturing chain, it is one of the few back-end platforms with real global customer qualification, global delivery capability, and advanced-packaging mass-production experience. Its hardest problem comes from exactly the same place: this is a business that is extremely capital-intensive, highly utilization-dependent, and deeply tied to depreciation and cash-flow timing. Today's market price is valuing "future JCET" onto "current JCET". Once investors admit that, the discussion becomes much simpler: this is a purchase of a Chinese advanced-packaging platform option that may deliver, or may arrive late, rather than an already-delivered AI packaging profit machine.

I am willing to recognize its industry position, and I am willing to recognize that Q1 2026 profit repair was of higher quality than full-year 2025. But in mid-2026, I am not willing to treat an OSAT with RMB 1.565 billion of 2025 net profit attributable to shareholders, still-weak owner earnings, and capex set to rise materially over the next two years as a target that can be comfortably held under a growth-stock framework. What would truly change my mind are three specific things, not a richer concept set: clear disclosure of advanced-packaging revenue, faster cash return after the capex peak, and core high-end packaging orders moving from "technically feasible" to "profit visible".

【Company Profile Score】

• Fundamental quality: Medium

• Growth: Medium

• Moat: Medium

• Financial stability: Medium

• Management credibility: Medium

• Valuation attractiveness: Low

• Risk level: High

• Suitable investor type: Cyclical / unsuitable for ordinary investors

【Investment Rating】

• Rating: Watch

• One-line investment thesis: The advanced-packaging option is real, but not yet enough to offset heavy capital intensity and front-loaded valuation.

• 【Ideal/Fair Buy Price】28–36 CNY Basis: This corresponds to applying about a 20% margin of safety to my conservative intrinsic-value range of RMB 35–45.

• Acceptable holding price: 47–63 CNY

• Clearly overvalued price: 70–78 CNY

• Current price classification: Clearly overvalued

• Is it worth waiting for a better price: Yes. A more suitable trigger would be the share price falling below RMB 36, while at least seeing quantified disclosure begin for advanced-packaging revenue/gross margin, or a material decline in the capex/operating-cash-flow ratio. The opportunity cost of waiting is that short-term themes may continue to push valuation higher, but such a rise does not provide a better risk-reward ratio.

• Target holding period: 3–5 years

• Expected annualized return: Conservative about -18%; base about -9%; optimistic about -5% (estimated over three years)

• Maximum loss risk: 40%–50%; triggers are the two pre-mortem scenarios above, both centered on "high capex failing to translate into high-quality orders and cash return"

• Signals for reassessment: if recurring net margin is below 3% for two consecutive quarters; if utilization is clearly insufficient after advanced-packaging expansion while depreciation keeps rising; if operating cash flow in 2026–2027 cannot cover major capex for an extended period; if domestic AI/high-end storage-related orders remain stuck at sample or validation stage for a long time; if progress in resolving horizontal competition within the CR system is materially below the committed pace.

【Valuation Range】

• current: 71.38 (as of the close on 2026-06-11)

• bear (conservative · ideal buy range): [28, 36]

• base (reasonable · acceptable holding range): [47, 63]

• bull (optimistic · above clear overvaluation line): [70, 78]

Key Data Table

Sources: 2025 annual report, 2026 Q1 report, public earnings-briefing information, and author calculations.

Research Uncertainties

• Public disclosure remains insufficient for advanced-packaging revenue, gross margin, HBM-related revenue, and domestic AI customer order size. This is the largest quantitative blind spot in this report.

• The specific deployment structure, phased commissioning schedule, and depreciation start timing for RMB 10 billion of 2026 capex remain broad in public disclosures.

• The profit elasticity, customer synergies, and integration pace after consolidation of SanDisk Semiconductor (Shanghai) can currently only be inferred indirectly from transaction and management statements.

• U.S. export controls and AI-chip-related rules changed multiple times in 2025–2026, so the compliance environment remains in flux.

• This report did not separately retrieve and incorporate the real-time yield on China's 10-year government bond for a spread comparison in the margin-of-safety analysis. The margin-of-safety judgment is therefore mainly based on equity cash return and corporate valuation, rather than a full asset-allocation perspective.

References

• JCET 2025 annual report, financial and cash-flow notes.

• JCET 2026 first-quarter report.

• JCET 2025 annual-report press release and 2026Q1 operating commentary.

• Shanghai Stock Exchange case study on the JCET-STATS ChipPAC acquisition.

• Shanghai Stock Exchange listing information, company history, and share-capital materials.

• Announcements and annual-report excerpts related to CR Holdings' entry and horizontal-competition commitments.

• Public disclosures on XDFOI mass production and 4nm multi-chip system packaging.

• Recent public data on ASE, Amkor, Tongfu Microelectronics, and Huatian Technology, plus Reuters/Google Finance/Yahoo Finance materials.

• Public BIS and Federal Register documents on AI Diffusion and advanced-computing export controls.

Other Securities Mentioned in the Report

• 3711.TW: Global OSAT leader, used to compare quantified advanced-packaging revenue disclosure and the global valuation center.

• AMKR.US: U.S. OSAT representative, used to compare North American supply-chain attributes and AI/automotive packaging valuation.

• 002156.SHE: The mainland peer with the most direct AI/HPC story, especially reflected in its AMD tie.

• 002185.SHE: Mainland packaging and testing recovery sample, used to compare traditional packaging and testing base with the pace of advanced-packaging upgrades.

• 2330.TW: Representative of foundry in-house advanced packaging, reminding that the profit pool does not naturally belong to OSATs.

• 688981.SHG: China's core wafer-foundry target, representing the other end of domestic front-end and back-end collaborative competition.

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.