A memory chip is only as good as the controller that talks to it. For years, Chinese storage brands bought that brain from someone else. Jiangbolong — Longsys — just shipped its own 5nm version by the hundreds of millions.
Key takeaways
- Longsys (江波龙, 301308.SZ; HK: 09976) reported 2025 revenue of ¥22.766 billion (≈ US$3.16 billion), up 30.36% year on year.
- Net profit attributable to shareholders reached ¥1.423 billion (≈ US$198 million), up 185.41%; recurring profit rose 674% to ¥1.289 billion.
- The company said its self-developed controller chips achieved mass deployment exceeding 140 million units, and that its 5nm UFS 4.1 controller entered scaled shipment with major smartphone makers.
- Its Hong Kong-listed shares reported 2025 revenue of HK$25.2 billion and net profit of HK$1.58 billion.
- Gross margin was 19.40%, up just 0.35 points — thin, reflecting a products business exposed to volatile memory pricing.
The two 2025 revenue figures above — ¥22.766 billion from the A-share (Shenzhen) annual report and HK$25.2 billion from the separate H-share (Hong Kong) filing — differ because of reporting standards and currency, not because one is the "real" revenue and the other is not.
From modules to muscle
Longsys, founded in Shenzhen, is best known as a storage-products company: embedded storage, solid-state drives (SSDs), memory cards, USB drives, portable SSDs, and DIMM modules, sold under the FORESEE, Lexar (雷克沙), and Zilia brands. Its customers span consumer electronics, servers and data centers, industrial equipment, and automotive. By revenue it is one of China's largest storage application players, and it is dual-listed, with A-shares in Shenzhen and H-shares in Hong Kong.
The 2025 annual report shows the top line reaching ¥22.766 billion (≈ US$3.16 billion), up from ¥17.46 billion in 2024 — a 30.36% rise — while the Hong Kong filing reported HK$25.2 billion in revenue and HK$1.58 billion in net profit. Net profit attributable to shareholders jumped to ¥1.423 billion (≈ US$198 million), more than double the prior year, with recurring profit up 674% to ¥1.289 billion.
The 5nm controller milestone
The strategic headline is not the memory itself but the logic around it. According to statements in the company's annual-report disclosure, Longsys's self-developed controller chips — the chips that manage data movement, error correction, and wear leveling inside a storage device — reached cumulative mass deployment of more than 140 million units across the product line. More pointedly, the company said its 5nm-process UFS 4.1 controller had reached deep ecosystem cooperation with several major wafer fabs and been designed into the supply chains of leading smartphone makers, with UFS 4.1 products entering scaled shipment.
UFS (Universal Flash Storage) is the fast embedded memory inside premium phones. A 5nm controller is a serious piece of logic design, and building it in-house moves Longsys from assembling bought-in memory into owning a core semiconductor component. As the company described it, UFS 4.1-class products are moving into volume in 2026.
Why a memory assembler building chips matters
Storage has always been a two-part business: the NAND flash memory (still largely sourced from global memory makers) and the controller plus firmware that make it useful. The controller is where differentiation and margin live. By developing its own, Longsys reduces reliance on third-party controller vendors and captures more of the value of every drive it ships.
The 2025 results also reflect the memory cycle. Industry pricing recovered from a deep 2023 downturn — Longsys swung from a ¥828 million loss in 2023 to profit in 2024 and a much larger profit in 2025 — so part of the surge is cyclical, not purely structural. Prudent readers should separate the two: the controller achievement is strategic; the profit jump is strategy plus a favorable market.
The global-brands question
Longsys is unusually international for a Chinese storage firm. Its Lexar (雷克沙) brand has long served consumer channels outside China, and the Zilia brand extends its footprint in select regions. That global exposure is a double-edged sword: it diversifies demand but also subjects the company to trade-policy and supply-chain risk on both the component and the customer side. The annual report shows the top five suppliers accounted for more than half of procurement spend, a reminder that the memory underneath remains concentrated among a few global suppliers.
For the broader China semiconductor narrative, Longsys illustrates a quieter kind of progress: not a breakthrough node or a flagship AI accelerator, but a products company climbing from modules toward in-house silicon. It is the unglamorous middle of the supply chain — and often where localization actually compounds.
Honest limitations
Figures are taken from Longsys's 2025 annual report (A-share and H-share filings) and financial-press summaries; they were not independently audited here. The claim of more than 140 million controller units deployed and the 5nm UFS 4.1 controller's customer integrations are company statements in the disclosure and were not verified against teardown or third-party shipment data. The 2025 profit surge is partly cyclical, tied to memory pricing recovery, and should not be read as purely structural gain. The article covers financial and strategic signals only; it does not assess the technical performance of Longsys's controllers versus leading independent controller vendors, nor quantify export-control or tariff exposure on its global sales.
What readers can do now
- Separate cycle from strategy: watch whether Longsys's gross margin and profit hold if memory prices soften in 2026 — that tests the durability of the gain.
- Track controller self-sufficiency as a metric: rising in-house controller share means more captured value and less vendor dependence.
- Use Longsys as a read on China's storage-brands globalization — its Lexar and Zilia performance abroad is a proxy for how far Chinese memory products travel.