On July 27, Changxin Technology, the largest IPO in the history of the STAR Market, officially listed, with its market capitalization surging directly to 3 trillion yuan, becoming the company with the highest market capitalization on the A-share market and the largest technology company by market capitalization in A-share history.
長鑫科技的創立背景源於一項代號為「506」的國產DRAM（動態隨機存取存儲器）造芯計劃，旨在打破國際市場美、韓存儲巨頭的長期壟斷，實現國產存儲芯片的突破。在合肥國資出資、兆易創新董事長朱一明提供技術的合作模式下，雙方一拍即合。此後，長鑫科技在研發與擴產過程中，更獲得數十家投資機構的實質資金支持。從2018年天使輪至2025年6月，長鑫科技已完成九輪融資，累計融資數百億元。至招股書簽署日，其背後共有60名股東。經過長達十年的耐心孵化與陪伴，這些股東共同推動並成就了如今這家國內第一、全球第四的國產DRAM芯片巨頭。
隨著長鑫科技順利登陸資本市場，最早且最堅定重倉的合肥國資、公司在九輪融資及IPO戰投階段接力的各方資本，以及以朱一明為代表的公司創始團隊，成為長鑫科技上市盛宴背後最重要的三方股東。
其中，最早重倉長鑫科技且為最大出資方的合肥國資成為最大贏家，預計浮盈超過萬億元，相當於合肥市2025年GDP（國內生產總值）的70%，地方政府扶持硬科技的合肥模式亦成為業界標杆性案例。長鑫科技的創始團隊也迎來人生中的高光時刻。包括長鑫科技創始人、董事長朱一明在內的十名董事及高級管理人員，人均身家上億；按長鑫科技上市首日收盤價計算，朱一明身家更達779億元。同時，通過實施兩期員工持股計劃，長鑫科技累計授予6760人次股份，覆蓋公司管理人才、業務骨幹、核心技術人員及生產一線關鍵崗位員工等，不少員工已藉由長鑫科技上市成為千萬富翁。
除合肥國資及創始團隊外，長鑫科技的其他股東也均收穫頗豐。IPO前，長鑫科技有60名股東，既有國家大基金二期、國調基金二期等國隊，又有阿里、騰訊、小米等產業資本，還有國壽投資、人保資本等保險資金以及燕創資本、招商局資本、恒旭資本等市場化投資機構。在IPO階段，長鑫科技又獲得了30家戰略投資者的鼎力支持。其中，阿里系通過參與多輪融資與戰略配售，合計持有超30億股，持股比例約為4.51%，上市首日持股市值接近1500億元。
於7月27日上市前一晚，長鑫科技於上海舉辦了名為「長鑄十年、鑫啟未來」的上市前答謝酒會，據現場人士觀察，約有包括戰略投資者、產業合作夥伴在內的300餘名嘉賓出席，朱一明於現場回顧長鑫科技的創業歷程，感慨公司創立之初與國際巨頭存在巨大的技術鴻溝，依靠長期資本支持及海內外技術團隊持續攻關，最終實現中國大陸DRAM規模化量產突破，並向耐心陪伴十年的長鑫股東表達感謝。
「晚宴氣氛熱烈，與會股東眾多，大家對長鑫十年攻堅所取得的發展成果給予充分認可，看好公司未來的發展空間，並對上市後的資本市場表現充滿期待。」一位參加晚宴的人士告訴《財經》。
合肥市國資「長期陪伴、早投硬科技、產業協同」的扶持模式，是地方國資支持硬科技發展的典範樣本，非常契合當下中國資本市場大力倡導的「科技－產業－資本」良性循環，有助於推動科技創新，落實促進中國經濟高質量發展的資本市場邏輯。長江商學院金融學教授、高層管理教育項目副院長周春生對《財經》表示，合肥國資以科技自主自強為核心目標，願意陪伴企業走完完整的產業週期，這種「耐心資本」是硬科技企業最稀缺的資源。
China Life Investment began its investment in 2020 when the industry was at the bottom of its cycle, with a post-A-round valuation of just over RMB 30 billion; over the years, this investment has delivered highly impressive returns. We have consistently aligned with national industrial policies and technological support initiatives by establishing innovation technology funds and dual-carbon technology funds, focusing primarily on hard technology sectors. Investing in CXMT can be seen as the starting point of this trend, and the entire insurance and financial industry has now fully understood and proactively shifted to the main channel of supporting hard technology development,” a China Life Investment insider familiar with the matter told Caijing.
合肥國資重倉
ChangXin Technology 能夠創立並成長為如今國內第一、全球第四的國產 DRAM 芯片巨頭，離不開其此前多輪融資中的 IPO 前股東支持，尤其是合肥國資，持股最早且最多。
「存儲晶片是典型的資本密集、技術密集和長週期產業，早期投入大、量產爬坡時間長，單靠短週期資金很難承擔從0到1的全週期風險。合肥國資在長鑫發展過程中發揮的作用，不只是出資，更重要的是提供長期、連續的資本支持，並協同融資、人才和產業鏈資源，幫助企業跨過早期建設和量產爬坡階段。」耀途資本創始合夥人楊光對《財經》表示。
回顧長鑫科技的創立歷程，2016年6月13日，由合肥國資旗下的合肥產投出資144億元、朱一明旗下的兆易創新出資36億元，合計出資180億元作為長鑫科技一期建廠啟動投資，長鑫科技由此正式成立。其後，合肥國資在長鑫科技創立後一路注資扶持，助力長鑫科技投入巨資進行DRAM晶片研發及擴產。粗略估計，截至IPO發行後，合肥國資通過合肥清輝集電（持有130.4億股，持股比例19.5%，第一大股東）、合肥長鑫集成（持有70.48億股，持股比例10.54%，第二大股東）、合肥產投壹號（持有11.11億股，持股比例1.66%，第七大股東）、合肥建長股權（持股9.013億股，持股比例1.35%，第十一大小東）、合肥產投高成長（0.38億股，0.057%）等多家主體合計持有長鑫科技約221.38億股，合計持股比例約為33.1%，是長鑫科技背後最大持股主體。
隨著長鑫科技順利實現IPO，合肥國資經過十年的陪跑，終迎來收穫期。按照長鑫科技上市首日收盤價49元/股計算，合肥國資持有的長鑫科技股權市值超過了1萬億元，達到10847.6億元。
值得注意的是，根據官方數據，合肥市2025年全年的GDP約為14210億元。這意味著，僅對長鑫科技一筆投資獲得的收益，就約相當於合肥市一年GDP的70%。
在長鑫科技上市當晚，作為合肥國資所屬三大平台公司之一的合肥產投集團發文，對長鑫科技上市進行了表態。合肥產投集團表示，2016年，在合肥市委、市政府決策部署下，合肥產投集團投資長鑫科技，助力其成長為中國規模最大、技術最先進、布局最全的動態隨機存取存儲器研發設計製造一體化企業。未來合肥產投集團將一如既往地支持長鑫科技發展，共同助力合肥市打造具有全球影響力的集成電路產業高地。
朴拙資本執行合夥人苗天一對《財經》表示：「合肥地方國資以長期耐心資本入局長鑫科技，陪伴硬科技從0到1成長，是國內地方產業招商與科技自主戰略結合的典型範式。」地方國資跳出短期財務回報訴求，以城市產業集群建設、供應鏈安全為核心目標，承擔存儲製造重資產、長週期、高虧損的早期風險，彌補市場化財務資本不敢投、熬不住長週期研發爬坡的短板。但該模式也對地方財政、產業研判能力提出極高要求。
Nankai University Institute of Financial Development Dean Tian Lihui told Caijing: “The core of the Hefei state-owned capital model lies in abandoning short-term profit-seeking thinking and acting as a ‘patient capital’ to accompany enterprises through high-risk phases such as technology R&D and production capacity ramp-up. Unlike simple financial bailouts, Hefei precisely selects industry sectors, ties up industrial chain resources, and provides phased dynamic funding, thereby ensuring capital efficiency while avoiding administrative interference with market pricing. The key to the model’s success lies in positioning state-owned capital as a ‘strategic investor’ rather than a ‘controlling shareholder,’ respecting market-oriented corporate operations, while amplifying technology spillover effects through industrial synergy.”
不過，各方專家亦提醒，合肥國資的投資模式不能簡單複製。
“地方國資支持硬科技有其必要性，但這種模式成立有幾個前提。政府更適合做耐心資本和產業生態的組織者，技術路線、產品和經營決策仍要交給專業團隊，並接受客戶和市場檢驗。能否複製，不在於照搬資金和政策，而在於當地有沒有產業基礎、專業決策能力、匹配的人才儲備以及陪伴企業穿越週期的能力。否則，也可能演變為同質化招商和重複建設。”楊光表示。
田利輝也提醒，需警惕盲目複製合肥模式的風險，硬科技投資需與地方產業基礎及專業研判能力相匹配，若缺乏技術理解與退出機制設計，易陷入「只投不退」困局。合肥經驗啟示我們，地方國資應做產業生態的「建築師」，而非簡單的「金主」。
周春生同樣認為，合肥模式儘管存在諸多優勢，但複製難度極高。投資有風險，如果失敗造成損失，問責邊界應如何確定？如果盡職免責容錯機制不到位，難以造就敢於承擔風險的國資投資者；過於強調容錯，又難免帶來道德風險。
「合肥模式從京東方開始起步，持續了十年時間，多數地方政府很難十年堅持做同一件事，並不是單純燒錢就能做成，因此很難被其他地方複製，未來城市產業發展會出現分化，基礎好的城市會越來越強。」一位投行人士告訴《財經》。
各方資本接力
事實上，雖然在合肥國資的鼎力支持下，長鑫科技項目得以啟動。但作為一個需要持續燒錢的重資產存儲芯片行業，尤其在前幾年週期低谷、全行業大面積虧損時，單靠合肥國資並不足以推動這個國產造芯計劃成功突圍。長鑫科技能夠成功登陸資本市場，背後離不開數十家機構和產業資本在其後數輪融資中的支持。
According to available information, from the start of its angel round in 2018 to its final financing round before listing in June 2025, CXMT completed nine financing rounds over eight years, accumulating a total funding scale of tens of billions of yuan. Prior to its IPO, CXMT had a total of 60 shareholders, including state-backed funds such as National Integrated Circuit Industry Fund II and China Reform Fund II, industrial capital such as Alibaba, Tencent, and Xiaomi, insurance capital such as China Life Investment and PICC Capital, and market-oriented investment institutions such as Yanchuang Capital, Jishi Capital, and Hengxu Capital.
值得注意的是，長鑫科技在IPO前的融資大部分來自國資系機構。「在長鑫科技融資早期，不少國資系機構響應國家投資硬科技的號召，在長鑫科技尚處虧損之際堅定入局。如今來看，結果證明，彼時的投資是完全正確的。」一位業內人士對《財經》表示。
中國人壽旗下的國壽投資於2020年參與了長鑫科技A輪融資，截至IPO前，國壽投資持有長鑫科技4.76億股股份，佔IPO前總股本比例0.79%。一位了解情況的國壽投資相關人士對《財經》講述：「即使當時最大膽的設想與現在的盛況也相去甚遠，以目前的市價，這是一筆近50倍收益的投資。」
Regarding the background of investing in CXMT at that time, a representative from China Life Investment told Caijing, “At that time, the state had clearly proposed policies to support the real economy through finance, and the hard technology sector became a key focus for support. At that time, China’s chip industry was in the process of building from 0 to 1, and the demand for independent and controllable industrial capabilities was very prominent. There were very few enterprises in China capable of tackling the DRAM chip field. With strong support from the Hefei municipal government, CXMT had been operating stably for four years, accumulating solid technical expertise and mass production experience, which had been fully validated by the market. The enormous demand for domestic substitution, combined with industry development opportunities, strengthened our investment confidence.”
“在2021年，DRAM市場長期被國際巨頭壟斷、半導體國產化前景尚不明朗的背景下，我們在寧波縣市區國資、當地民營資本等支持下，毅然選擇重倉支持長鑫科技，投資13.2億元，這也是我們公司有史以來最大的一筆投資。”參與長鑫科技B輪融資的寧波燕創合夥人、投資負責人郭成威對《財經》表示，“支撐這一決定的，正是我們對長鑫科技作為實現國產DRAM從無到有歷史性突破的先行者身份的高度認可。”
Regarding the specific reasons for investing in CXMT, Guo Chengwei provided four points: First, the DRAM chip market has vast potential; second, CXMT is one of the few domestic companies to achieve mass production of general-purpose DRAM, presenting significant opportunities for domestic substitution, further enhanced by support from national and local industrial capital, aligning with its long-term investment strategy in hard technology; third, CXMT possesses robust technical and industrial capabilities with a clear commercialization pathway; fourth, the company’s governance structure is well-designed and its management team is highly capable—supported by Hefei state-owned capital and the National Integrated Circuit Industry Investment Fund, the core technical and operational team led by Zhu Yiming has established a governance structure that combines government investment backing with market-oriented management, achieving an excellent integration of industrial guidance and market mechanisms.
In 2021, China Merchants Capital, which participated in the Series B financing of CXMT, stated that it made a strategic investment in CXMT in 2021 and supported CXMT in breaking through key technologies and accelerating capacity expansion through continuous capital and industrial resource empowerment. China Merchants Capital will work with CXMT to promote the expansion of its advanced capacity and the technological breakthrough of next-generation memory technology, continuously consolidating CXMT’s competitive position in the global industry.
Securities firm-backed capital, Huian Jiaye, the private equity subsidiary of Huian Securities, has invested in Changxin Technology through a combination of direct investment via a special fund and a fund-of-funds approach. It stated, “Huian Jiaye, grounded in the national strategy for semiconductor self-reliance and the local industrial cluster advantages in Anhui, has provided long-term support to Changxin Technology to help it navigate industry cycles, effectively supporting the breakthrough of China’s domestic DRAM industry while achieving dual returns in both fund value and industrial value.”
長鑫科技上下游產業鏈上的相關產業資本也紛紛投資長鑫科技。
其中，阿里系是長鑫科技最大的外部投資機構之一，旗下阿里雲計算在上市前最後一輪Pre-IPO融資中，以2.63元/股的價格投入61億元，為該輪最大的產業投資方。IPO後阿里雲計算持有23.19億股，持股比例為3.47%，是長鑫科技第六大股東；同時，阿里網絡持有6.76億股，持股比例1.01%。此外，阿里系還通過旗下阿里雲飛天參與了長鑫科技IPO階段的戰略配售，獲配0.18億股，阿里系合計持股數達到30.13億股，持股比例約為4.51%。按上市首日收盤價49元/股計算，阿里系持有市值達到1476億元。
「長鑫科技作為全球近十年唯一實現規模化量產的新晉 DRAM 廠商，打破了海外廠商的技術壟斷，其技術稀缺性、車規級存儲與智能汽車產業鏈的協同價值，是恒旭敢於重倉的核心邏輯。」參與長鑫科技 B 輪融資的恒旭資本表示。恒旭資本係上汽集團（國有企業）旗下通過市場化改制成立的私募股權投資機構。
進入 IPO 階段，長鑫科技又獲得 30 家戰略投資者的鼎力支持，鎖定期為 12 個至 36 個月。其中，國隊主要代表——全國社會保障基金理事會旗下 36 個產品認購 75 億元，國調基金二期認購 1 億元，國隊合計認購 76 億元，佔戰略配售總額的五成以上。人保財險、中國人壽、中郵人壽、泰康人壽等大型保險資金亦在列。長鑫科技的戰投還包括多家產業鏈夥伴，如中微半導體、拓荆科技、瀾起科技、滬硅產業等半導體公司。下游終端的產業鏈戰略合作企業亦參與其中，例如阿里雲飛天、上海灝裕信息（騰訊）、深圳三快網絡（美團）等互聯網公司；武漢壹捌壹零（小米）、傳音控股等消費電子公司；蔚來動力、奇瑞汽車等汽車企業。
Among them, NIO Power subscribed to RMB 158 million, with a lock-up period of 18 months. It is understood that NIO’s manufacturing facility is also located in Hefei, just a few hundred meters away from CXMT’s factory, and NIO is a cornerstone strategic partner of CXMT for DRAM. Both parties will collaborate on existing automotive-grade LPDDR4X and LPDDR5X products. On the evening of July 26, Li Bin, founder of NIO, attended the pre-IPO thank-you dinner hosted by CXMT, and when discussing the collaboration with CXMT, he said, “The cooperation is progressing smoothly, and the partnership with CXMT helps stabilize NIO’s supply chain.”
創始團隊主導
除了合肥國資及各方資本，長鑫科技背後還有一支重要的股東力量——以朱一明為核心的創始團隊。在公司股權分散、無實際控制人的股權結構下，他們擁有公司的經營主導權。
截至IPO前，包括朱一明在內的10名董事、高級管理人員，合計持有公司20.12億股，持股數量均超過380萬股。
According to the prospectus, before its listing, ChangXin Memory Technologies implemented two employee stock ownership plans. In the second phase, the board of directors of ChangXin granted Zhu Yiming 1.536 billion shares as recognition for his decade-long efforts in advancing China's domestic DRAM industry, at a grant cost of only HK$0.108 per share of registered capital.
Among them, Zhu Yiming holds a total of 1.59 billion shares, company director, president, and core technical staff Cao Kanyu holds 212 million shares, and executive vice president Zhu Wenju holds 53 million shares; the three rank first, second, and third in shareholding volume. Calculated based on the closing price of NT$49 per share on the first day of listing, the market value of shares held by these ten individuals is close to NT$98.6 billion. Zhu Yiming’s personal market value of shares reaches NT$77.9 billion. Meanwhile, Zhu Yiming has voluntarily committed that within ten calendar years after the third anniversary of Changxin Technology’s listing, he will use 768 million shares under his ownership for employee incentives to motivate the company’s core personnel.
公開資料顯示，朱一明於1997年畢業於清華大學物理系，畢業後赴美國紐約州立大學攻讀電子工程碩士，隨後進入矽谷從事存儲器晶片開發工作。2005年朱一明回國創立存儲晶片設計公司兆易創新。2016年在合肥國資支持下創立長鑫科技，並於2018年辭去兆易創新總經理職務，專注長鑫科技發展。
When CXMT was founded, China's DRAM industry was completely blank. Samsung Electronics, SK Hynix, and Micron Technology had already entered mass production of 10-nanometer-class DRAM, while CXMT could only start based on the outdated technical framework left by Qimonda, with a process gap of more than two generations. Zhu Yiming led CXMT to abandon a step-by-step iteration approach and instead adopted a "leapfrog R&D" strategy, directly achieving mass production of a process technology platform from the first to the fourth generation.
The success of the ChangXin project, beyond favorable conditions such as the opportunity for domestic substitution, hinges on another crucial factor: people. The entrepreneurial spirit embodied by Zhu Yiming has been vital to ChangXin’s development—whether in strategic vision, dedication, or critical decision-making at pivotal moments, all of which profoundly influenced ChangXin’s rapid start and growth. “During ChangXin’s early stages, Zhu Yiming placed great emphasis on intellectual property, leveraging IP related to Qimonda and proactively securing a large number of defensive patents—a move that proved extremely critical,” said a representative from China Life Investment, speaking to Caijing. He also assembled a highly experienced technical team and managed to retain these talents in Hefei—an achievement that is not easy. These talents chose to stay largely due to Zhu Yiming’s personal charisma and influence. The company’s sustained investment in and deep commitment to talent acquisition, along with its continuous attraction of top-tier industry professionals, has been a key reason for its success.
“他們非常堅韌，且極具創業初心與激情。團隊核心成員大多深耕行業多年，長期紮根產業一線，專注於DRAM技術攻關與量產落地。能在行業低谷期耐得住寂寞，這種堅持非常難得，是一支兼具技術實力、實幹精神和戰略視野的成熟創業團隊。”該國壽投資相關人士對《財經》表示。
Guo Chengwei also told Caijing that the management team of ChangXin Technologies is impressive. First, they demonstrate a profound sense of industrial patriotism and commitment to long-termism, diligently focusing on developing domestic memory storage without being swayed by short-term gains, showing remarkable resolve. Second, they possess strong professional capabilities, with core members holding extensive global semiconductor industry backgrounds and excelling in technology R&D, factory operations, and market expansion. Third, they are pragmatic and candid, with a broad perspective, objectively recognizing challenges and risks, placing high importance on talent incentives, and fully aligning the core team. At the same time, they maintain a clear mindset as a follower, steadily building their technology and business strategies—making them a manufacturing team worthy of “strategic” long-term support.
值得注意的是，隨著合肥國資、朱一明創始團隊及數十家機構資本支持長鑫科技順利上市，公司無實際控制人問題也受到外界關注。長鑫科技副總裁、董事會秘書袁園在上市路演中提到，公司上市後股權結構將進一步分散，前五大股東持股比例均不超過30%，不存在單一持股比例超過50%的股東，上市後公司仍將保持無實際控制人的控制權架構。
合肥國資、公司創始團隊、機構和產業資本三方，誰才能決定長鑫科技未來的最終走向？據了解，作為長鑫科技最大投資方的合肥市國資委與合肥市經開區國資委更多扮演財務投資者的角色，其均表示，不謀求合肥清輝集電和長鑫科技實際控制權。同時，朱一明雖為合肥清輝集電執行事務合夥人，並不能實際控制該公司。
ChangXin Technology's status as having no actual controller is closely related to the rights check-and-balance agreements with its largest shareholder, Hefei Qinghui Jidian Partnership. ChangXin Technology has previously stated that the rights check-and-balance mechanism of Hefei Qinghui Jidian aligns with the investment original intent of the two local state-owned platforms in Hefei, which is “to aim for technological self-reliance and strength, not to seek control,” while also reflecting a strong emphasis on the operational autonomy of the management team.
Zhou Chunsheng told Caijing that Changxin Technology’s governance structure of “no actual controller, with state-owned and founding team shares power and checks” is an innovative model of large-scale hard-tech state-owned equity participation, fundamentally balancing the strategic demands of state capital with the operational autonomy of the founding team. Zhou Chunsheng summarized that the advantage of the above checks and balances mechanism lies in: state capital providing support through resources, credit, and strategy instead of administrative intervention, while the founding team retains full control over technology and operations; this simultaneously solidifies state capital as the “ballast” and activates market-driven “original momentum,” avoiding the efficiency losses typical of traditional state-owned enterprises.
Miao Tianyi told Caijing that the above model has four advantages: First, the separation of strategic authority and operational authority allows state-owned capital to anchor industrial security and large-scale investment without interfering in daily operations or technology pathways, safeguarding the founding team’s leadership and avoiding short-term political achievements disrupting long-cycle R&D; Second, multi-layered partnership checks and balances prevent any single entity from controlling the voting rights of the largest shareholder, curbing unilateral intervention and enhancing market-oriented governance; Third, risk and return are aligned, with state-owned capital bearing the risk at the bottom of the cycle while management ties performance to long-term equity returns, creating a framework of shared risk and shared benefits; Fourth, compliance is aligned with IPO requirements, clarifying the boundaries of state-owned capital’s rights and responsibilities while balancing asset preservation and appreciation with market-oriented operations of science and technology enterprises.
However, this mechanism also carries risks. Zhou Chunsheng summarized that the storage industry has characteristics such as strong cyclical fluctuations, and business decisions requiring rapid responses—such as price adjustments and capacity allocation—are hindered by multi-party consensus mechanisms, easily leading to missed opportunities; during industry downturns, the inefficiency in decision-making will be further highlighted. In the future, as the industry enters a mature phase, conflicts may arise between state-owned capital’s demands for value preservation and appreciation and industrial security, versus management teams’ focus on market expansion and short-term performance, potentially turning the balance structure into internal friction and affecting governance stability.
In Miao Tianyi’s view, a structure without actual controllers lacks a defense center when facing hostile takeovers or external capital博弈, with equity stability relying on long-term mutual understanding among all parties; diverse stakeholders have conflicting interests, raising governance costs through prolonged coordination; and complex multi-layered structures increase the difficulty of equity penetration and transparency of rights and responsibilities.
本文來自微信公眾號「讀數一幟」（ID：dushuyizhi007），作者：康國亮 張建鋒