Jiangsu Changjing's IPO Filing Shows China's Chip Push Is Also About Mature Power Devices

China's semiconductor capital story is often reduced to the race for AI accelerators and advanced nodes. Jiangsu Changjing Electronics Technology's newly accepted IPO filing points to a less glamorous but commercially important part of the same campaign: power semiconductors, discrete devices and power-management chips that sit inside cars, industrial systems, servers, energy equipment and consumer electronics.

The Shenzhen Stock Exchange's official project record shows Changjing was accepted on June 10 for a main-board IPO. The filing lists planned fundraising of 1.4904 billion yuan, Huatai United Securities as sponsor and the issuer's industry as computer, communications and other electronic equipment manufacturing. That acceptance does not mean approval. It does mean another domestic chip company has entered the public review system at a time when Beijing's semiconductor priority is spreading well beyond the highest-end logic chips.

Changjing's prospectus describes a company built around discrete semiconductors and power-management ICs, including diodes, triodes, MOSFETs, IGBTs and discrete-device wafers. It calls the company a comprehensive semiconductor enterprise using both Fabless and IDM models. The distinction is important. A pure fabless designer can move faster with less fixed investment, but a company with some manufacturing and packaging-test control can argue for better supply security, process coordination and cost discipline.

The filing shows why this segment deserves attention. Changjing says it has mass-production capacity across more than 10,000 product specifications and positions itself among China's most complete domestic suppliers of discrete devices and power-management ICs. It also cites industry association evidence that in 2024 its diode and triode sales ranked third in China and second among mainland Chinese enterprises, while its CSP MOSFET market share ranked third in China and first among mainland enterprises.

These are not frontier AI chips. They are the semiconductors that convert, distribute, regulate and protect electrical power. That makes them easy to overlook and hard to replace. Electric vehicles need power devices. Solar and storage systems need them. Industrial automation, robotics, servers and consumer electronics need them. If China wants supply-chain resilience in advanced manufacturing, it needs companies that can make dependable components at scale, not only firms that can tell a story about large models.

The planned use of proceeds makes the industrial logic clearer. Changjing's prospectus says funds will support a project for annual packaging and testing capacity of 10 billion industrial- and automotive-grade power devices, R&D for MOSFETs, IGBTs and third-generation semiconductor technology, power-management IC R&D and working capital. The capital needs are therefore tied to capacity, product upgrading and balance-sheet support rather than brand marketing.

That is both a strength and a risk. Power semiconductors can benefit from long replacement cycles and growing demand in vehicles, factory equipment and energy systems. But the field is also exposed to price competition, inventory swings and customer qualification cycles. A company expanding packaging-test capacity has to keep utilization high enough to absorb depreciation and labor costs. If downstream demand slows, capacity can become a margin problem.

The financial profile is stronger than many early-stage chip issuers. Changjing reports revenue of 2.260 billion yuan, 2.678 billion yuan and 2.985 billion yuan over the reporting period, while net profit attributable to shareholders after excluding non-recurring items rose from 124 million yuan to 218 million yuan and then 334 million yuan. Those figures give the IPO a different texture from loss-making AI-chip candidates. The company is not asking investors only to fund a future technology dream; it is presenting a profitable industrial platform that wants more capital.

That does not remove the need for scrutiny. Investors should examine customer concentration, inventory discipline, average selling prices, the profitability of each product line and how much of the growth comes from market share rather than a temporary cycle. Power devices can be cyclical. They can also face brutal substitution when customers qualify multiple suppliers. The promise of localization is real, but localization does not guarantee pricing power.

The Fabless-plus-IDM model also deserves a careful reading. Vertical integration can improve supply reliability and technical feedback. It can also raise fixed costs and management complexity. For a company selling thousands of specifications, the challenge is not only designing products; it is controlling manufacturing economics, yield, quality and delivery across a wide customer base. The more Changjing moves into automotive and industrial products, the more reliability and certification will matter.

The timing is useful for China's capital markets. The mainland IPO system has been trying to balance support for hard technology with tighter review of quality. Changjing's filing gives exchanges and investors a cleaner case than many speculative issuers: a semiconductor business with revenue, profit, domestic market ranking claims and a clear use of funds. That may make it easier to analyze, but not automatically easier to value.

For Jingpost's semiconductor coverage, Changjing's filing is a reminder that China's chip drive has two layers. One layer is symbolic and strategic: AI accelerators, advanced compute and the effort to reduce reliance on foreign vendors. The other is operational and deep: power devices, analog chips, packaging, testing, wafers and the components that keep industrial systems running. The second layer is less glamorous, but it may offer a more measurable route from policy priority to revenue.

The question for Changjing is whether public capital can help it move up the value chain without weakening returns. A successful listing would strengthen its ability to fund packaging-test capacity, R&D and working capital. A poor execution cycle would leave it exposed to the usual semiconductor traps: inventory, price erosion, underused capacity and customer bargaining power.

China's semiconductor ambitions will not be judged only by who builds the fastest AI chip. They will also be judged by whether companies like Changjing can supply the uncelebrated devices that make electrification, industrial automation and computing infrastructure reliable. The Shenzhen acceptance is only the first step. The harder test is whether a profitable domestic power-chip platform can use the public market to become larger without becoming less disciplined.