⚠️ Editorial note: The open source ecosystem in China operates under a distinct institutional framework — characterized by state-led initiatives, intranet-like boundaries, and top-down governance. Readers should be aware that this context differs from the community-driven open source model common in other regions. The term “open source” as used in Chinese media may refer to practices that diverge from the conventional definition.

China Open Source Daily — 2026-10-01

🏛️ Institutional Unwinding — Higobase (瀚高) Equity Transfer: Shandong SASAC State Capital’s 5-Year Exit Cycle and the Mid-Tier Domestic-Replacement Database Vendor’s Collapsing Financials

1. 自主可控新鲜事 (Autonomous-Controllable Fresh News, WeChat) (2026-09-30) — “老牌信创数据库企业瀚高软件股权挂牌转让 · 信创产业进入分化期”

2. 山东产权交易中心 listing announcement (2026-09-30, listing open through 2026-10-28) — 0.7236% of Higobase’s equity at a floor price of ¥19,197,900

The Shandong Property Rights Exchange Center’s listing announcement places Higobase (瀚高软件, est. 2005, self-identified as “the country’s first domestic open-source database localization enterprise”) on the market for transfer. The sharpest institutional-economics fact of the day is the transaction structure: 0.7236% of equity at a floor price of ¥19,197,900 (implying an overall valuation of approximately ¥2.65 billion), against an internal appraisal of ¥12,379,500 (a premium of over 55%). The transferor is 山东科创新股权投资有限公司 (Shandong Kechuang Innovation Equity Investment Co., Ltd.), a wholly-owned subsidiary of Shandong Ke-chuang Group with the Shandong SASAC as its actual controller — meaning this is a provincial SASAC technology-investment platform executing a scheduled exit, not a market-driven liquidation. Per the piece, Shandong’s provincial regulations specify that such platforms operate on a 拨投结合 (guantou jiehhe) combined capital-injection-plus-equity-investment model with an equity holding period of no more than 5 years, after which funds are rolled over and redeployed. The transaction is therefore the terminal event in a documented 5-year state-capital investment cycle.

Higobase’s institutional label vs. Higobase’s financial reality. The piece documents two institutionally distinct surfaces of the same company. On the institutional-label side, Higobase is documented as the “country’s first domestic open-source database localization enterprise” (est. 2005), a dominant participant in multiple database industry standards, with its V9.0 database management system selected for Shandong’s first-batch-of-high-end-software-products list, and a June 2026 release of an “AI-era intelligent data foundation HigoBase.” On the financial side, the numbers are: 2025 full-year net profit ¥2.97 million; 2026 first-five-months net profit −¥6.2953 million; 2026 first-five-months revenue below ¥40 million; 2025 full-year revenue ¥174 million. Per the piece, “from a slim profit to a loss took only five months.”

The Xinchuang stratification framing. Per the piece: “The Xinchuang market, in the transition from policy-driven to market-driven, has been characterized by top-tier vendors — with ecosystem and capital advantages — accelerating land-grab, while mid-tier vendors’ survival space has been continuously squeezed.” The piece explicitly places Higobase in the mid-tier: “Higobase has both a long history and substantial technical accumulation, but has not entered the first tier in market scale and commercialization capability.” The open-source-database-localization path faces, per the piece, a double squeeze — Oracle/MySQL international ecosystem inertia on one side, and Dameng (达梦) / RENCK (人大金仓) / other mid-tier domestic-replacement vendors competing for the same Xinchuang cake on the other side.

The sharpest institutional-economics object of the day: the Shandong SASAC Guantou Jiehhe 5-year-exit-cycle-plus-Higobase-mid-tier-domestic-replacement-plus-open-source-label-as-market-positioning triple. Per the piece: “The question is whether the ¥2.65 billion overall valuation can obtain transferee recognition in the current primary-market environment.” The piece does not answer — but the answer’s absence is the sharpest institutional-economics observation: the state capital has exited (as its 5-year cycle requires), the mid-tier vendor has negative cash flow, and no buyer at the 55%-premium implied valuation has appeared in the ~4-week window between the 09-30 listing and the 10-28 close — meaning the institutional object (a 20-year-old domestic-replacement database vendor with open-source-label-plus-industry-standard-participation credentials) is documented at its terminal-investment-cycle exit point with no market buyer.

**The institutional-economics pairing with this series’ September 1 and September 2 briefings’ documentation of the National AI Industry Investment Fund’s $8.8B-in-capital consolidation of Moonshot at $35B and DeepSeek at $74B.** This cycle's briefing documents the same state-capital-portfolio surface this series' September 1 briefing documented at the frontier-AI-consolidation-head layer being operationalized at the Xinchuang-tail-exit layer simultaneously — the state-capital portfolio is being rotated from the low-valuation Xinchuang-era tail (Higobase at ~¥2.65B with negative cash flow) toward the high-valuation frontier-AI-consolidation head (Moonshot + DeepSeek at combined ~$85B), and the two institutional surfaces — the Xinchuang-tail-exit and the frontier-AI-consolidation-head — are being operationalized at the same state-capital-portfolio-institutional object at different ends of the same portfolio.

The open-source-label-as-market-positioning reading. Per the piece, Higobase is documented at “国内首家致力于开源数据库国产化的基础软件企业” (the country’s first domestic open-source database localization foundation software enterprise). The piece does not describe Higobase’s open-source codebase, its contribution surface, its upstream community relationship, or its license — it describes Higobase’s open-source label in the context of competing with Dameng/RENCK for Xinchuang procurement market share. From an institutional economics standpoint, this open-source-label / Xinchuang-market-positioning pairing is a structural finding in this series: the “open source” institutional label in a state-directed-capital-context document is being used as a market-positioning brand equity and an industry-standard-participation credential, not as a FLOSS-commons institutional identity. This is the institutional-label-and-institutional-identity distinction that the series’ prior briefings have documented on the Chinese-frontier-lab-open-source-agent surface (September 30 briefing) and the Chinese-municipal-open-source-AI-foundation surface (September 30 briefing) — and this cycle’s briefing documents the same distinction on the mid-tier-domestic-replacement-database-vendor surface for the first time.

Institutional significance: The Shandong SASAC Guantou Jiehhe 5-year-exit-cycle Higobase 0.7236% equity transfer is the first-documented instance in this series of a Chinese-provincial-SASAC-technology-investment-platform being operationalized at the Xinchuang-tail-mid-tier-domestic-replacement-database-vendor exit layer simultaneously with this series’ September 1 briefing’s documentation of the National AI Industry Investment Fund at the frontier-AI-consolidation-head layer — a first-documented two-endpoint-simultaneous state-capital-portfolio-rotation institutional-form observation.

Sources:

🏗️ Institutional Unwinding — The 1,200 SOE Digital Technology Subsidiaries: The 2020 SASAC Directive’s Institutional Wave Entering Its 5-Year-Point Collapse, with Four Explicitly Documented Institutional Failure Modes

2. 自主可控新鲜事 (Autonomous-Controllable Fresh News, WeChat) (2026-09-30) — “背靠大树为啥还是凉了：1200家央国企数科公司的生死局”

The second 自主可控新鲜事 piece documents the collapse of the 1,200 SOE Digital Technology (数科) subsidiaries that the 2020 SASAC directive (一纸文件) triggered across central and state-owned enterprises. Per the piece, at its peak in 2020-2021 the count was 1,200+; by 2025, 注销解散+名存实亡 (registered-but-dissolved + nominally-existing-but-effectively-dead) had eliminated more than half, with 500-600 remaining nominally in operation — of which, per the piece, “真正活得好的两只手数得过来” (the number actually thriving can be counted on two hands).

The four documented institutional failure modes. The piece explicitly names four institutional failure modes, each of which is a Williamson L3-governance-mechanism misalignment between the state-institutional-form template applied to the subsidiary and the tech-business-market-rhythm template required:

  • Dead-mode one — 用行政思维做科技业务 (administrative-thinking applied to tech business). Per the piece: “Many of these digital-technology companies were established not because of market demand but because of policy assessment. Once the 2020 SASAC directive came down, not establishing a digital-technology company was equivalent to being late to transformation, and every level of SOE panicked. Without a business model, register first; without technical accumulation, hang the sign first.” The piece documents the “new bottle, old wine” (新瓶装旧酒) pattern: the group’s original information department was rebranded as a digital-technology subsidiary to satisfy the directive, with the same team, the same business, and cascading duplication across group levels (HQ at level 2, followed by level 3 and level 4 with entirely overlapping business).

  • Dead-mode two — 用短期考核管长期战略 (short-cycle assessment applied to long-term strategy). Per the piece: “The tech industry logic is long-term investment, rapid iteration, patient capital. Making products requires repeated polishing, potentially 3-5 years to see returns. But SOE assessment is one-year-cycle, three-year-term. The leader changes every three years — the money this term invests and the next term reaps the fruit, who would do it?” The consequence, per the piece: long-horizon R&D was halted, self-development investment was compressed, and the subsidiaries converged on low-end outsourcing and system integration — quick-revenue, low-margin, non-replicable.

  • Dead-mode three — 用僵化制度应对灵活市场 (rigid institutions applied to an agile market). Per the piece: “The tech industry competes on talent. But digital-technology subsidiaries apply SOE salary systems — total-wage-volume-capped, slow promotion. Internet mega-firms offer fresh graduates 200,000-300,000 RMB, senior engineers one million per year. What can a digital-technology subsidiary offer? At most half the market rate.” The piece explicitly documents the “talent-training-base” effect: employees spend 2 years accumulating experience, then leave. And on the approval side: “SOE layered approval — a project takes so long to complete its process that by the time it’s approved the market opportunity has already closed.”

  • Dead-mode four — 把数字化转型当实习生练手 (digital transformation as an intern-training exercise). Per the piece, many digital-technology subsidiaries’ leaders are cross-industry transfers, rotation-trainees, or emergency-appointments — people “who don’t understand what digital transformation is” placed at the helm of teams “composed of either information-department-transferred old employees or fresh graduates with 3-5 years of experience.”

The sharpest institutional-economics object of the day: the “digital transformation is a power-and-interest redistribution, not a system upgrade” framing. Per the piece: “数字化变革的本质是打破既有利益格局，重构商业模式，改变管理决策机制——这不是上个系统那么简单，这是一场权力和利益的重新分配。数字化不是请客吃饭，不是谁都能来练手的。” The piece’s framing is that the digital-technology subsidiary is not merely an outsourced IT-services vendor but a redistribution of power-and-interests within the SOE group — meaning the subsidiaries that fail are not failing because they lack technical capability but because they are structurally positioned inside a power-and-interest-redistribution process that they cannot resolve. This is a Williamson L1-social-embedding observation: the subsidiary is embedded in an SOE-group power structure that determines its access to resources and its ability to reallocate them, and the state-capital-plus-SOE-digital-technology institutional architecture cannot override the L1 power structure to convert the subsidiary into a market-driven tech business.

The survivor-model documentation. The piece names three survivors that “make it”:

  • 宝信软件 (Baosight Software, under China Baowu Steel Group): ¥12.9 billion revenue, ¥2.554 billion net profit, +16.82% growth
  • 联通数科 (Unicom Digital Technology, under China Unicom): ¥11.9 billion revenue, +21% growth, marketized-revenue 3-year compound growth +27.1%
  • 中远海科 (COSCO Shipping Technology, under COSCO Shipping Group): ¥829 million revenue, +22.15% growth

Per the piece, the pattern is: survivors either go deep on internal-industry services (Baosight for steel industry, Unicom for telecom industry) or break out to be industry-digitalization pioneers (COSCO Shipping for shipping logistics industry) — none of the three competes head-to-head with Huawei or Alibaba on generic cloud/AI.

The institutional-economics pairing with this series’ September 1 briefing’s documentation of the National AI Industry Investment Fund’s state-capital portfolio consolidation. The two placements together document the same state-directed-capital-plus-SOE-digital-technology institutional architecture on both its formation side (National AI Industry Investment Fund consolidating Moonshot/DeepSeek at combined ~$85B, as this series’ September 1 briefing documented) and its unwinding side (1,200 SOE digital-technology subsidiaries collapsing to 500-600 nominally-alive-with-most-losing-money, and mid-tier domestic-replacement database vendors like Higobase being exited at the 5-year-cycle endpoint with negative cash flow). The institutional-economics reading is that the state-directed-capital architecture’s institutional-form has both a formation side (frontier-AI consolidation, Xinchuang top-tier vendor concentration) and an unwinding side (mid-tier domestic-replacement vendor exits, SOE digital-technology subsidiary collapse), and the two sides are being documented simultaneously at the same institutional-architecture surface.

Institutional significance: The 自主可控新鲜事 “背靠大树为啥还是凉了” 1,200 SOE Digital Technology subsidiary collapse piece is the first-documented instance in this series of the state-directed-capital-plus-SOE-digital-technology institutional architecture being operationalized at the 2020-SASAC-directive-1200-to-500 collapse layer simultaneously with this series’ September 1 briefing’s documentation of the state-directed-capital architecture at the National AI Industry Investment Fund’s frontier-AI-consolidation-head layer — a first-documented two-side-simultaneous state-directed-capital-architecture institutional-form observation.

Sources:

🔍 Commentary

Two institutionally dense placements on the same unsolved institutional object this series’ prior briefings have been documenting — the state-directed-capital-plus-SOE-digital-technology institutional architecture — but at an institutional layer this series’ prior briefings have not documented at before: the unwinding side, not the formation side.

This cycle’s briefing documents two placements on the same institutional object this series’ prior briefings have been running. The formation side has been the primary subject of this series’ prior briefings: the National AI Industry Investment Fund consolidating Moonshot at $35B and DeepSeek at $74B at the state-capital-consolidation layer (September 1 briefing); the Hangzhou Open Source AI Foundation plus LF AI & Data plus AAIF international-Linux-Foundation-family co-hosting layer (September 30 briefing); the Huawei Ascend Tribe openPangu-2.0-Training plus openPangu-2.0-RL open-sourcing at the Ascend-native-training-pipeline-code layer (September 30 briefing). This cycle’s briefing documents the unwinding side:

  • The Higobase Shandong-SASAC-Guantou-Jiehhe-5-year-exit-cycle placement documents the Xinchuang-tail-mid-tier-domestic-replacement-database-vendor exit at the state-capital-portfolio-rotation endpoint, with the state capital exiting a 20-year-old “open-source database localization” vendor at ¥2.65B implied valuation against negative cash flow and a 55%-premium-with-no-obvious-buyer outcome — the institutional-object terminal-investment-cycle event.
  • The 自主可控新鲜事 1,200 SOE Digital Technology subsidiary collapse placement documents the 2020-SASAC-directive-1200-to-500 collapse at the state-directed-capital-plus-SOE-digital-technology institutional-architecture layer, with four explicitly named Williamson L3-governance-mechanism failure modes and an explicit “digital transformation is a power-and-interest-redistribution” framing that converts the subsidiary failure from a technical-market-comparison issue into an institutional-power-structure issue.

One structural pattern across the two placements: the same state-directed-capital-portfolio is being documented at both its consolidation-head and its rotation-tail simultaneously.

The Higobase placement and the SOE digital-technology collapse placement together document a state-capital-portfolio-rotation from tail (Xinchuang-era mid-tier vendors like Higobase) to head (frontier-AI-consolidation labs like Moonshot and DeepSeek), and the two placements are the terminal-tail-event and the mid-portfolio-collapse-event of the same rotation. From an institutional economics standpoint, this portfolio-rotation-head/tail pairing is a structural finding: the state-directed-capital-plus-SOE-digital-technology institutional architecture is not a static capital pool but a rotating portfolio, and the two surfaces — the Higobase tail-exit and the Moonshot/DeepSeek head-consolidation — are documented at the same portfolio-institutional-object surface at different ends of the same 5-year-cycle-plus-frontier-AI-consolidation-horizon rotation.

One open institutional-economics question across the two placements.

The two placements together document that the state-directed-capital architecture’s mid-tier and lower-tier portfolio is unwinding while its head portfolio is consolidating, and the institutional-economics question — which this cycle’s briefing documents but does not answer — is whether the state-directed-capital architecture’s unwinding side (mid-tier vendor exits, SOE digital-technology subsidiary collapse) is coherent with its formation side (frontier-AI consolidation, top-tier Xinchuang vendor concentration):

  • The Higobase placement documents a 20-year-old “open-source database localization” vendor at a 26.5亿 valuation against negative cash flow and a state capital that must exit per its 5-year-cycle — with the “open source” label functioning as market-positioning brand equity and not as FLOSS-commons institutional identity.
  • The SOE digital-technology subsidiary placement documents 1,200 SASAC-directive-driven subsidiaries collapsing to 500-600 nominally-alive with most losing money — with the four institutional failure modes (administrative-thinking-on-tech-business, short-cycle-assessment-on-long-strategy, rigid-institution-on-agile-market, digital-transformation-as-intern-trial) being L3-governance-mechanism misalignments between the SOE-group institutional template and the tech-business-market-rhythm template.
  • The pairing with this series’ September 1 briefing’s Moonshot/DeepSeek consolidation documents the state-directed-capital architecture’s head side consolidating at combined ~$85B — with the state capital priced into the frontier-AI-consolidation-head layer with the same vehicle (National AI Industry Investment Fund) that is simultaneously rotating out of the Xinchuang-tail-mid-tier-vendor layer.

The institutional-economics observation is that the state-directed-capital architecture is documented at both its consolidation-head and its rotation-tail simultaneously, and the two placements this cycle’s briefing documents are the first-documented instances in this series of the architecture’s unwinding side being documented at institutional-formation-layer depth — not merely at market-news depth.

One perspective, not a verdict.

The two placements this cycle’s briefing documents — the Higobase Shandong-SASAC-Guantou-Jiehhe-5-year-exit-cycle 0.7236% equity transfer and the 自主可控新鲜事 1,200 SOE Digital Technology subsidiary collapse with four explicitly named institutional failure modes — are best read as observations of institutional unwinding in progress, not as verdicts on institutional direction. The Higobase transfer will close on 2026-10-28 with an unknown transferee; the SOE digital-technology subsidiary collapse is a documented historical event with the survivors-and-casualties ratio already established; the state-capital-portfolio rotation from Xinchuang-tail to frontier-AI-consolidation-head is documented at two endpoints simultaneously. What this cycle’s briefing documents is that the state-directed-capital-plus-SOE-digital-technology institutional architecture has crossed a first-documented threshold — from a formation-side-only subject of this series’ prior briefings to a both-sides-simultaneously-documented object, with the formation side (Moonshot/DeepSeek consolidation) and the unwinding side (Higobase exit, SOE digital-technology subsidiary collapse) both documented at institutional-formation-layer depth rather than market-news-layer depth.

Editorial note on perspective: This briefing presents one institutional-economics reading of Chinese open-source developments, not a verdict. The “institutions” in these two stories — the Shandong SASAC Guantou Jiehhe 5-year-exit-cycle platform, the 2020 SASAC directive’s 1,200 SOE digital-technology subsidiary wave, the Higobase open-source database localization vendor, the Baosight/Unicom/COSCO Shipping Technology survivor model — are treated as objects of observation, not targets of critique. The Great Divergence 2.0 framework (FLOSS vs. State-Chartered Codebase vs. Intranet Shared Source vs. Cyber-Estate) and the Williamson L1→L4 institutional-economics reading (L1 social embedding → L2 institutional environment → L3 governance mechanisms → L4 resource allocation) are lenses, not universal answers. One perspective, not a verdict.

Deduplication note: The Higobase Shandong-SASAC-Guantou-Jiehhe-5-year-exit-cycle 0.7236% equity transfer and the 自主可控新鲜事 1,200 SOE Digital Technology subsidiary collapse with four explicitly named institutional failure modes were not covered in any of the prior three briefings (September 27, 28, 29, 30). The September 30 briefing covered the Reuters September 29 exclusive on Chinese AI agents’ documented deception and breakout-compatible behaviour with the CAC AI Safety Governance Framework 3.0 and Carnegie ecosystem-lag diagnostic, the Hangzhou Open Source AI Foundation’s Global Open-source AI Challenge Grand Finals documentation of a Chinese-municipal-open-source-AI-foundation plus two international Linux Foundation-family ecosystem-governance bodies plus five Chinese co-organizers at 14,000-developers-from-91-countries scale, and the Huawei Ascend Tribe’s openPangu-2.0-Training plus openPangu-2.0-RL repo release documentation of the Ascend-trial-ecosystem surface’s transition from weights-only to weights-plus-training-code-plus-RL-code. The September 29 briefing covered the InnerSource China Summit 2026 Shanghai documentation and the Thought-Steel plus Huxiu essay on “China has no web pages anymore.” The September 28 briefing covered the CAC formal inquiry into DeepSeek and Moonshot with the CAC-September-15-batch-of-ten-API-relay-station-and-un-declared-overseas-data-transfer-enforcement-cases precedent-apparatus simultaneously operationalized, the DSec arXiv paper plus CVE-2026-82533 in DeepSeek Harness pairing, and the V4.1-Flash plain-MIT-license-plus-ungated-weights release.

WeChat input note: The WeChat input file raw/2026-09-30.md in the china-daily-english-input repository contained two substantive institutional-economics articles (both from 自主可控新鲜事, the “Autonomous-Controllable Fresh News” WeChat public account), both substantially new to this series’ prior briefings and both documented in this briefing’s Institutional Unwinding placements. The 自主可控新鲜事 public account is not among this series’ tracked-16-account WeChat monitor list (开放原子/CCF/华为开源/木兰/COPU/天工开物/明说/FlagOpen/开源社/硅基时间/AtomGit/Linux内核开发者大会/GOSIM-CSDN/黄大年茶思屋/IvorySQL/夜天之书) — it is a broader Xinchuang/domestic-replacement observer not specifically tracking the Chinese open-source community. Both articles are included here because they document Chinese-open-source-relevant institutional economics (the Higobase “open-source database localization” label and the 1,200 SOE Digital Technology subsidiary collapse) at institutional-formation-layer depth — a layer this series’ prior briefings have not documented at.