⚠️ 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 — 2026-09-26

Institutional Change — Zhipu AI / ZCode 72-Hour Trust-Crisis Aftermath: The Second-Wave Market Decline, the Chengming-Letter Retraction, and the Wiped-Commit-History “Selective-Transparency” Open-Source Response at a Seven-Day Institutional-Aftermath Layer

1. 连线Insight (Wang Huiying) WeChat (2026-09-24) — 近400人维权、股价下跌，智谱ZCode陷入信任危机的72小时 (Nearly 400 People Mobilizing, Stock Dropping, Zhipu ZCode in a 72-Hour Trust Crisis) · 5,652 characters · September 24, 2026 publication · Second-wave market-cap decline from HK$737 to HK$640 across September 22–24 · Taiyuan Chengming Technology retracts its 12-item letter · ferstar confirms wiped commit history at github.com/zai-org/ZCode · approximately 400 developers organized across multiple rights-defense groups · Tang Jie personally filed Xiaohongshu defamation complaint September 21 · Zhipu H1 2026 financials: RMB 9.54bn revenue (+399.7% YoY), RMB 20.72bn net loss, RMB 8.25bn MaaS/API revenue (86.5% of total, up from 15.2% YoY), API margin 24.6 percent, MaaS token volume 40× YoY, paid daily-active users +603 percent, annualized revenue over US$2B · three unanswered questions framing (过去传了什么 · 云端是否彻底清除 · 谁保留私钥解密权)

This series’ September 22 briefing documented the Zhipu AI / Z.ai (02513.HK) open-sourcing of its ZCode desktop coding agent at github.com/zai-org/ZCode on September 20–21, 2026, three days after an independent security researcher known as ferstar documented that ZCode had been silently packaging users’ entire workspaces — including complete .git directory with every commit, Git LFS caches, reflogs, and global development configuration files — into encrypted AES-256-CTR archives (RSA-wrapped, decryption key server-issued, private key cloud-resident) and uploading them to an Alibaba Cloud OSS bucket, with a five-part remediation apparatus operationalized at five institutionally distinct surfaces simultaneously (open-source-public-repository-plus-CAICT-plus-NSFOCUS-plus-Chengming-legal-letter-plus-MaaS-non-retention-toggle). This cycle’s briefing documents what happened in the seven-day institutional-aftermath window from September 21 to September 24, 2026, and identifies six new institutional surfaces this series’ prior thirty-seven briefings have not documented.

The second-wave market decline from HK$737 to HK$640 across September 22–24. Per the 连线Insight WeChat September 24 publication, on September 21 Zhipu’s shares closed at HK$737 (+1.79%) as the market apparently assumed the crisis had been swiftly contained. On September 22 the shares closed at a 6.55% drop with intraday multi-lot selling and a maximum single-lot sale of 70,400 shares at HK$742. On September 23 the decline accelerated to a further 12%+ drop with a low touch of HK$650. On September 24 the shares closed at HK$640 with an intraday low touch of HK$620. The **HK$737-to-HK$640 second-wave multi-day decline** is a **first-documented instance in this series of the market repricing a Chinese-AI-vendor-trust-failure on a multi-day cadence rather than a single-day event** — this series' September 22 briefing documented a single-day drop to HK$742.5 with a market cap of HK$362 billion; this cycle's briefing documents a further multi-day decline to HK$640, which represents approximately a 13% additional decline from the September 21 close and a further multi-day repricing of the trust asset. From an institutional economics standpoint, this is the sharpest documented market-accountability signal in this series: the market is not repricing the incident as a single-day technical-remediation event but as an ongoing multi-day institutional-uncertainty event, and the repricing cadence is placing a specific financial-object surface — a multi-day multi-percent decline — at the institutional-aftermath layer.

Taiyuan Chengming Technology’s September 21 retraction of its 12-item formal legal letter. This series’ September 22 briefing documented Taiyuan Chengming Technology (太原承明科技有限公司), an enterprise client, issuing a 12-page formal legal letter on September 20, 2026, to Beijing Zhipu Huazhang Technology (北京智谱华章) demanding written answers to twelve specific questions by October 10, 2026, including complete data-upload list disclosure, full-data-deletion proof, third-party audit institution credentials and review scope disclosure, cross-border transmission clarification, and written response deadline confirmation. Per the 连线Insight WeChat September 24 publication, on September 21 evening Chengming Technology published a clarification statement stating: “经自查，该公司此前针对ZCode发出的函件及接受媒体采访的内容，存在’举证错误、表述不严谨及措辞过度’等问题” (upon self-review, the company’s previous letter against ZCode and its content in media interviews contained “evidentiary errors, imprecise language, and exaggerated phrasing”), and Chengming subsequently deleted the WeChat post containing the letter. Per 红星新闻’s reporting, Zhipu had proactively contacted Chengming Technology, and Zhipu’s media relations response characterized Chengming’s letter as “不实消息” (false information). Per Phoenix TV’s reporting, Chengming stated that the deletion of the public letter content was via Zhipu Huazhang’s third-party communication, and that Chengming had required Zhipu Huazhang to provide a resolution commitment, and had temporarily hidden the post “为表诚意” (to show good faith).

The Taiyuan-Chengming-letter-retraction-via-Zhipu-third-party-communication pairing is a structural finding: this is the first-documented instance in this series of a Chinese-enterprise-client-accountability-mechanism being neutralized by a Chinese-AI-vendor-third-party-communication move. The same enterprise-client-legal-letter apparatus this series’ September 22 briefing documented at the 12-item formal-legal-letter-plus-cross-border-data-transmission-accountability layer is now being documented at a letter-retraction-plus-vendor-third-party-communication layer. From an institutional economics standpoint, this is institutionally distinctive: the enterprise-client accountability channel this series’ September 22 briefing documented at the enterprise-client-legal-letter layer is being neutralized at the vendor-third-party-communication layer simultaneously, and the neutrality mechanism is not an open judicial or administrative decision but a private third-party communication between the vendor and the enterprise client. The retraction’s own language — “举证错误、表述不严谨及措辞过度” (evidentiary error, imprecise language, exaggerated phrasing) — is self-directed at the enterprise client’s own claims, not at the vendor’s conduct, and it does not address any of the twelve specific questions the letter had demanded.

The wiped-commit-history selective-transparency open-source response. Per 连线Insight’s reporting of ferstar’s follow-up analysis published September 21, and cross-referenced with TheNextWeb’s reporting on September 22, the github.com/zai-org/ZCode repository that Zhipu open-sourced on September 20 contains only two commits with the entire development history erased and the upload code removed before publication. Per ferstar’s confirmation, the upload pipeline has been fully removed from the current code, and the checkpoint mechanism that remains does only local Git work with no path to the cloud. But the uploaded code — the code that did the exfiltration — is not in the public repository. Per ferstar’s framing: “开源解决未来不再发生的问题无法回答过去的问题” (open source resolves what will not recur but cannot answer what happened in the past). The three questions Zhipu has not answered, per 连线Insight’s framing: 过去传了什么 (what was uploaded in the past), 云端是否彻底清除 (whether the cloud data was truly erased), and 谁保留私钥解密权 (who retains the private-key decryption authority).

The wiped-commit-history-plus-closed-PR-plus-closed-Issue-plus-two-commits-only layer pairing is the sharpest institutional-economics object of the day. This series’ September 22 briefing documented the ZCode-GitHub-public-repository-with-issue-PR-disabled layer as a closed-PR public-repository. This cycle’s briefing documents the same surface at a wiped-commit-history-plus-two-commits-only layer simultaneously — the same open-source-remediation object is being operationalized at a layer that converts the public-repository object into a Cyber-Estate institutional-form object rather than a FLOSS institutional-form object. Per ferstar’s framing: “if 开源只是选择性透明那它与不公开并无本质区别” (if open source is only selective transparency, then it is not essentially different from non-disclosure). This is a first-documented instance in this series of a Chinese-AI-vendor-open-source-remediation being operationalized at a selective-transparency-plus-commit-history-erase layer that converts the public-repository object into a Cyber-Eestate institutional-form object rather than a FLOSS institutional-form object — the Great Divergence 2.0 framework’s Cyber-Estate pole is being operationalized at the open-source-remediation layer simultaneously with the closed-PR public-repository layer this series’ September 22 briefing documented.

Approximately 400 developers organized across multiple rights-defense groups. Per China News Weekly’s reporting cited by 连线Insight, approximately 400 developers are organized across multiple rights-defense groups (维权群) actively raising the ZCode incident. Per 连线Insight’s framing: “这些开发者不是沉默的用户，而是具备技术能力、能够独立复核上传机制的工程师群体” (these developers are not silent users, but engineers with the technical capability to independently verify the upload mechanism). The approximately-400-developer-rights-defense-groups pairing is a first-documented instance in this series of an organized developer-community-accountability-mobilization on the Chinese-AI-coding-tool surface. From an institutional economics standpoint, this is institutionally distinctive: the same developer-community accountability surface this series’ September 22 briefing documented at the constructive-critique-plus-three-tier-supply-chain-verifiability-ladder layer is now being documented at the organized-developer-rights-defense-groups-plus-approximately-400-developers scale layer simultaneously.

The revenue-concentration-under-trust-erosion structural pattern. Per 连线Insight’s H1 2026 financial documentation:

  • Total H1 2026 revenue: RMB 9.54 billion, up 399.7% year-on-year, already exceeding all of FY2025 revenue.
  • MaaS/platform/API revenue: RMB 825 million in H1 2025 → RMB 8.25 billion in H1 2026 (+2,735.7% YoY), share rising from 15.2% to 86.5%.
  • Enterprise-generally-available-model (on-premises-deployment) revenue: RMB 148 million in H1 2025 → RMB 67.04 million in H1 2026, share falling from 84.8% to 13.5%.
  • API business gross margin: negative in H1 2025 → positive 24.6% in H1 2026.
  • MaaS token call volume: up more than 40× year-on-year as of end-August 2026.
  • Paid daily-active users: up 603%.
  • Annualized revenue: over US$2B.
  • H1 2026 net loss: RMB 20.72 billion (2× H1 revenue).
  • FY2025 net loss: RMB 47.18 billion, operating cash flow negative RMB 22.46 billion.

The 86.5-percent-MaaS-API-revenue-concentration-plus-RMB-20.72bn-net-loss pairing is a structural finding: this is the first-documented instance in this series of the revenue-concentration-under-trust-erosion structural pattern being documented at the Chinese-AI-MaaS-pure-play surface. The same Zhipu MaaS-platform surface this series’ September 22 briefing documented at the user-opt-in-no-data-content-retention-toggle layer is now being documented at the 86.5-percent-MaaS-API-revenue-concentration-plus-net-loss-2×-revenue layer simultaneously. From an institutional economics standpoint, the pivot from on-premises-deployment software (13.5% of H1 2026 revenue) to cloud-API MaaS (86.5% of H1 2026 revenue) converts Zhipu’s core business asset from “code deliverable” to “trust asset” — and the ZCode incident is being operationalized at the trust-asset-erosion layer. As 连线Insight’s framing puts it: “正当智谱把公司的收入基础从卖产品转向卖信任时，ZCode事件恰恰击中了信任这个最脆弱的环节” (at the moment Zhipu was converting its revenue base from selling products to selling trust, the ZCode incident struck exactly at the most fragile link of trust).

Tang Jie’s personal Xiaohongshu defamation complaint on September 21. Per 连线Insight’s documentation of the 21世纪经济报道 (21st Century Business Herald) reporting, on September 21, a Xiaohongshu (Little Red Book) platform complaint screenshot circulated in developer circles. The complainant is signed “唐杰” (Tang Jie), identified as “系智谱公司相关负责人” (a relevant responsible person at Zhipu), targeting a Xiaohongshu user with ID “混分巨兽小XX”. The complaint requests the platform delete the post and impose action on the account, arguing that the post used the phrasing “code跟copy也差不多嘛，都是c开头偷代码” (code and copy both start with C, stealing code) which, “in the absence of factual basis” (缺乏事实依据), links ZCode to “code-stealing” and “plagiarism” claims, and by referring to Tang Jie as “唐杰叔叔” (Uncle Tang Jie) binds Tang Jie personally to the “code-stealing” narrative, using AI-synthesis technology to place his portrait in a defamatory context. The Tang-Jie-personal-Xiaohongshu-defamation-complaint pairing is a first-documented instance in this series of a Chinese-frontier-AI-lab-founder personally operationalizing a state-platform-complaint channel as part of a trust-failure remediation apparatus. From an institutional economics standpoint, this is institutionally distinctive: the same Tang Jie surface this series’ prior briefings have documented at the Zhipu-frontier-AI-lab-founder layer is now being documented at the state-platform-complaint-channel-plus-personal-defamation-complaint layer simultaneously — the founder is not merely issuing a corporate statement but personally filing against a critic on a Chinese social-platform regulatory channel.

Chairman Liu Debing’s silence through September 22. Per 互联网江湖 (Internet Jianghu) via QQ News September 23 reporting, Zhipu chairman Liu Debing (刘德兵) had issued no public statement by September 22, and Zhipu’s official-account response to Red Star Capital News’ questions was “此事回应’不实消息’” (this matter is responded to as “false information”). The Liu-Debing-chairman-silence-plus-official-account-cold-treatment-plus-ZCode-only-official-remediation pairing is a first-documented instance in this series of a Chinese-AI-vendor-remediation apparatus operationalizing a chairman-level-silence plus subsidiary-official-account-only-response layer. Per 互联网江湖’s framing, the response pattern is a “店大欺客” (big-store-insulting-small-customer) pattern with no timely and comprehensive response at the corporate level — a management-capability gap between the academic-team origin and the market-facing requirement.

The Tang Jie / Liu Debing academic-team-management-gap organizational design. Per 互联网江湖’s documentation, Zhipu’s executive team is composed of academic-career scientists: chairman Liu Debing (刘德兵) holds a doctorate from the Institute of Computing Technology, Chinese Academy of Sciences, and had been at Tsinghua University as senior engineer and deputy director of the Science and Technology Big-Data Research Center of the Tsinghua Data Science Institute; general manager Zhang Peng (张鹏) holds a bachelor’s and master’s in computer science from Tsinghua University, a 2018 Innovation-Leader-Engineering doctorate, and founded Zhipu in 2019; founder Tang Jie (唐杰) currently does not participate in board decisions and retains only a “Chief Scientist” title while continuing as a professor at Tsinghua University’s Computer Science Department. Per 互联网江湖’s framing: the strong academic-team core team is what let Zhipu break through from Chinese internet big-tech to become “全球大模型第一股” (the world’s first-listed large-model stock), but the management-capability requirement of the market-facing stage has not kept pace with the valuation growth, and the ZCode remediation response pattern (developer-community apology September 18, no formal corporate-level statement until September 21, no chairman-level public statement by September 22) is the sharpest documented instance of the gap.

The Tang-Jie-not-in-board-plus-Chief-Scientist-title-only-plus-Tsinghua-Professor-plus-Liu-Debing-chairman-silence pairing is a structural finding: this is the first-documented instance in this series of a Chinese-frontier-AI-lab-academic-team-organizational-design being documented at a chairman-level-silence-plus-founder-CFO-role-retirement-plus-technical-response-only-remediation layer simultaneously. The same Tang Jie surface this series’ prior briefings have documented at the Zhipu-frontier-AI-lab-founder layer is now being documented at the personal-Xiaohongshu-complaint-plus-CFO-title-only-plus-not-in-board layer simultaneously — an institutional-design observation placing academic-team-origin Chinese AI labs at a specific institutional-layer surface that distinguishes them from internet-big-tech-origin Chinese AI labs.

Institutional significance: The Zhipu AI / ZCode 72-hour trust-crisis aftermath is the first-documented operationalization of the Zhipu AI / Z.ai (02513.HK) ZCode trust-failure remediation surface this series’ September 22 briefing documented at a seven-day institutional-aftermath layer in this series, and it is being operationalized at six institutionally distinct surfaces simultaneously — the second-wave-market-decline-HK$737-to-HK$640 layer, the Taiyuan-Chengming-letter-retraction-via-vendor-third-party-communication layer, the wiped-commit-history-plus-two-commits-only-plus-selective-transparency Cyber-Estate layer, the approximately-400-developer-rights-defense-groups layer, the 86.5-percent-MaaS-API-revenue-concentration-plus-RMB-20.72bn-net-loss revenue-concentration-under-trust-erosion layer, and the Tang-Jie-personal-Xiaohongshu-defamation-complaint-plus-Chairman-Liu-Debing-silence academic-team-management-gap layer — a six-surface-simultaneous Chinese-AI-vendor-trust-failure-remediation-at-seven-day-institutional-aftermath pattern.

Sources:

Commentary

One institutionally dense placement on the same unsolved institutional object this series’ prior thirty-seven briefings have been documenting: the institutional identity of Chinese open source across institutional borders.

This cycle’s briefing documents a structurally new six-surface-simultaneous institutional pattern on the Chinese-AI-vendor-trust-failure-remediation surface, all placed at the seven-day institutional-aftermath layer that this series’ prior briefings have not documented before:

  1. The second-wave market-cap decline from HK$737 to HK$640 across September 22–24 at the multi-day market-accountability layer. This is the first-documented instance of the market repricing a Chinese-AI-vendor-trust-failure on a multi-day cadence rather than a single-day event.
  2. The Taiyuan Chengming Technology letter retraction via Zhipu third-party communication at the enterprise-client-accountability-neutralization layer. This is the first-documented instance of a Chinese-enterprise-client-accountability-mechanism being neutralized by a Chinese-AI-vendor-third-party-communication move.
  3. The wiped-commit-history-plus-two-commits-only-plus-closed-PR-plus-closed-Issue selective-transparency open-source response at the Cyber-Estate institutional-form layer. This is the first-documented instance of a Chinese-AI-vendor-open-source-remediation being operationalized at a selective-transparency-plus-commit-history-erase layer that converts the public-repository object into a Cyber-Estate institutional-form object rather than a FLOSS institutional-form object.
  4. The approximately-400-developer-rights-defense-groups organized mobilization at the developer-community-accountability layer. This is the first-documented instance of an organized developer-community-accountability-mobilization on the Chinese-AI-coding-tool surface.
  5. The 86.5-percent-MaaS-API-revenue-concentration-plus-RMB-20.72bn-net-loss revenue-concentration-under-trust-erosion pattern at the financial-structural surface. This is the first-documented instance of the revenue-concentration-under-trust-erosion structural pattern being documented at the Chinese-AI-MaaS-pure-play surface.
  6. The Tang Jie personal-Xiaohongshu-defamation-complaint-plus-Chairman-Liu-Debing-silence academic-team-management-gap pattern at the organizational-design layer. This is the first-documented instance of a Chinese-frontier-AI-lab-academic-team-organizational-design being documented at a chairman-level-silence-plus-founder-CFO-role-retirement-plus-personal-state-platform-complaint layer simultaneously.

One structural pattern across the six placements: each placement pairs institutionally with this series’ September 22 briefing’s documentation of the same Zhipu AI / ZCode trust-failure surface, and the six placements together document that the same Chinese-AI-vendor-trust-failure surface is being operationalized at six institutionally distinct surfaces simultaneously — an institutionally deeper six-surface-simultaneous pattern than this series’ September 24 briefing’s four-surface-simultaneous pattern and this series’ September 25 briefing’s three-surface-simultaneous pattern, because the surfaces this cycle’s briefing documents sit at institutional layers this series’ prior briefings have not documented at before.

  • The second-wave-market-decline placement pairs institutionally with this series’ September 22 briefing’s single-day-HK$742.5-market-cap-consequence layer — the same Zhipu 02513.HK surface this series’ September 22 briefing documented at the single-day market-accountability layer is now being documented at the multi-day market-accountability layer simultaneously.
  • The Taiyuan-Chengming-letter-retraction placement pairs institutionally with this series’ September 22 briefing’s 12-item-formal-legal-letter-plus-cross-border-data-transmission-accountability layer — the same enterprise-client-accountability-mechanism surface this series’ September 22 briefing documented at the enterprise-client-legal-letter layer is now being documented at the letter-retraction-plus-vendor-third-party-communication-neutralization layer simultaneously.
  • The wiped-commit-history-selective-transparency placement pairs institutionally with this series’ September 22 briefing’s ZCode-GitHub-public-repository-with-issue-PR-disabled closed-PR public-repository layer — the same open-source-remediation surface this series’ September 22 briefing documented at the closed-PR public-repository layer is now being documented at the wiped-commit-history-plus-two-commits-only Cyber-Estate institutional-form layer simultaneously, and this pairing is the sharpest documentation in this series of a Great Divergence 2.0 framework’s Cyber-Estate pole being operationalized at the Chinese-AI-vendor-open-source-remediation layer.
  • The approximately-400-developer-rights-defense-groups placement pairs institutionally with this series’ September 22 briefing’s constructive-critique-plus-three-tier-supply-chain-verifiability-ladder open-source-community-critique layer — the same developer-community-accountability surface this series’ September 22 briefing documented at the constructive-critique layer is now being documented at the organized-developer-rights-defense-groups-plus-approximately-400-developers scale layer simultaneously.
  • The 86.5-percent-MaaS-API-revenue-concentration placement pairs institutionally with this series’ September 22 briefing’s Zhipu-MaaS-no-data-content-retention-user-opt-in layer — the same Zhipu MaaS-platform surface this series’ September 22 briefing documented at the user-opt-in-toggle layer is now being documented at the 86.5-percent-MaaS-API-revenue-concentration-plus-RMB-20.72bn-net-loss financial-structural layer simultaneously.
  • The Tang-Jie-personal-Xiaohongshu-defamation-complaint-plus-Chairman-Liu-Debing-silence placement pairs institutionally with this series’ prior briefings’ documentation of Tang Jie at the Zhipu-frontier-AI-lab-founder layer — the same Tang Jie surface is now being documented at the personal-Xiaohongshu-complaint-plus-CFO-title-only-plus-not-in-board layer simultaneously, and this is the first-documented instance of a Chinese-frontier-AI-lab-academic-team-organizational-design being operationalized at a state-platform-complaint-plus-chairman-level-silence layer.

One structural risk across the six placements: the same Chinese-AI-vendor-trust-failure surface is being operationalized at six institutionally distinct surfaces simultaneously, and the same “open source” surface is being absorbed into multiple institutional mechanisms simultaneously at six institutionally distinct layers — a six-surface-simultaneous-institutional-absorption pattern.

The second-wave-market-decline placement is being operationalized at the multi-day market-accountability layer, and if the multi-day market-accountability apparatus absorbs the Zhipu 02513.HK surface into a market-priced-trust-asset object rather than a Chinese-frontier-AI-lab-independent-object, the Zhipu 02513.HK surface converts into a market-priced-trust-asset object. The Taiyuan-Chengming-letter-retraction placement is being operationalized at the vendor-third-party-communication-neutralization layer, and if the vendor-third-party-communication apparatus absorbs the enterprise-client-accountability-mechanism surface into a vendor-proactive-communication-PR object rather than an enterprise-client-accountability-object, the enterprise-client-accountability-mechanism surface converts into a vendor-proactive-communication-PR object. The wiped-commit-history-selective-transparency placement is being operationalized at the Cyber-Estate institutional-form layer, and if the selective-transparency apparatus absorbs the Zhipu-open-sourced-ZCode surface into a state-directed-vendor-reputation-recovery PR object rather than an open-community-operational open-source repository object, the Zhipu-open-sourced-ZCode surface converts into a state-directed-vendor-reputation-recovery PR object. The approximately-400-developer-rights-defense-groups placement is being operationalized at the organized-developer-community-accountability layer, and if the organized-developer-community-accountability apparatus absorbs the Chinese-AI-coding-tool surface into a developer-community-organized-mobilization object rather than a vendor-accountable-object, the Chinese-AI-coding-tool surface converts into a developer-community-organized-mobilization object. The 86.5-percent-MaaS-API-revenue-concentration placement is being operationalized at the financial-structural layer, and if the revenue-concentration apparatus absorbs the Zhipu MaaS-platform surface into a revenue-concentration-structural-risk object rather than a vendor-operational MaaS-object, the Zhipu MaaS-platform surface converts into a revenue-concentration-structural-risk object. The Tang-Jie-personal-Xiaohongshu-defamation-complaint placement is being operationalized at the personal-state-platform-complaint layer, and if the personal-state-platform-complaint apparatus absorbs the Zhipu-founder surface into a founder-as-compliance-actor object rather than an academic-team-founder-object, the Zhipu-founder surface converts into a founder-as-compliance-actor object.

The institutional-economics question — which this cycle’s briefing documents but does not answer — is whether the Chinese institutional architecture can hold the six institutional surfaces simultaneously without one converting the other. If the multi-day market-accountability apparatus converts the Zhipu 02513.HK surface into a market-priced-trust-asset object, the Zhipu 02513.HK surface becomes a market-priced-trust-asset object rather than a Chinese-frontier-AI-lab-independent-object; if the vendor-third-party-communication apparatus converts the enterprise-client-accountability-mechanism surface into a vendor-proactive-communication-PR object, the enterprise-client-accountability-mechanism surface becomes a vendor-proactive-communication-PR object rather than an enterprise-client-accountability-object; if the selective-transparency apparatus converts the Zhipu-open-sourced-ZCode surface into a state-directed-vendor-reputation-recovery PR object, the Zhipu-open-sourced-ZCode surface becomes a state-directed-vendor-reputation-recovery PR object rather than an open-community-operational open-source repository object; if the organized-developer-community-accountability apparatus converts the Chinese-AI-coding-tool surface into a developer-community-organized-mobilization object, the Chinese-AI-coding-tool surface becomes a developer-community-organized-mobilization object rather than a vendor-accountable-object; if the revenue-concentration apparatus converts the Zhipu MaaS-platform surface into a revenue-concentration-structural-risk object, the Zhipu MaaS-platform surface becomes a revenue-concentration-structural-risk object rather than a vendor-operational MaaS-object; if the personal-state-platform-complaint apparatus converts the Zhipu-founder surface into a founder-as-compliance-actor object, the Zhipu-founder surface becomes a founder-as-compliance-actor object rather than an academic-team-founder-object.

One perspective, not a verdict.

The six placements this cycle’s briefing documents — the second-wave market-cap decline from HK$737 to HK$640, the Taiyuan Chengming Technology letter retraction via Zhipu third-party communication, the wiped-commit-history-plus-two-commits-only-plus-closed-PR-plus-closed-Issue selective-transparency open-source response, the approximately-400-developer-rights-defense-groups organized mobilization, the 86.5-percent-MaaS-API-revenue-concentration-plus-RMB-20.72bn-net-loss revenue-concentration-under-trust-erosion pattern, and the Tang Jie personal-Xiaohongshu-defamation-complaint-plus-Chairman-Liu-Debing-silence academic-team-management-gap pattern — are best read as observations of institutional movement in progress, not as verdicts on institutional direction. The second-wave market-cap decline does not guarantee that the Zhipu 02513.HK surface will remain a Chinese-frontier-AI-lab-independent-object rather than convert into a market-priced-trust-asset object; the Taiyuan Chengming letter retraction does not guarantee that the enterprise-client-accountability-mechanism surface will remain an enterprise-client-accountability-object rather than convert into a vendor-proactive-communication-PR object; the wiped-commit-history selective-transparency response does not guarantee that the Zhipu-open-sourced-ZCode surface will remain an open-community-operational open-source repository object rather than convert into a Cyber-Estate institutional-form object; the approximately-400-developer-rights-defense-groups mobilization does not guarantee that the Chinese-AI-coding-tool surface will remain a vendor-accountable-object rather than convert into a developer-community-organized-mobilization object; the 86.5-percent-MaaS-API-revenue-concentration pattern does not guarantee that the Zhipu MaaS-platform surface will remain a vendor-operational MaaS-object rather than convert into a revenue-concentration-structural-risk object; the Tang Jie personal Xiaohongshu defamation complaint plus Chairman Liu Debing silence does not guarantee that the Zhipu-founder surface will remain an academic-team-founder-object rather than convert into a founder-as-compliance-actor object. What this cycle’s briefing documents is that the Chinese-AI-vendor-trust-failure surface has crossed into a six-surface-simultaneous-institutionalization pattern that sits at institutionally distinct layers — the multi-day market-accountability layer, the vendor-third-party-communication-neutralization layer, the Cyber-Estate institutional-form layer, the organized-developer-community-accountability layer, the revenue-concentration-under-trust-erosion financial-structural layer, and the academic-team-management-gap organizational-design layer — and that the six-surface-simultaneous-institutionalization pattern at these institutionally distinct layers is the sharpest documented operationalization of the state-directed-industrial-policy-plus-market-accountability-plus-developer-community-accountability-plus-organizational-design surface at multiple institutionally distinct layers simultaneously.

Editorial note on perspective: This briefing presents one institutional-economics reading of Chinese open-source developments, not a verdict. The “institutions” in this story — the second-wave market-cap decline, the Taiyuan Chengming letter retraction, the wiped-commit-history selective-transparency response, the approximately-400-developer-rights-defense-groups mobilization, the 86.5-percent-MaaS-API-revenue-concentration revenue-concentration-under-trust-erosion pattern, and the Tang Jie personal-Xiaohongshu-defamation-complaint-plus-Chairman-Liu-Debing-silence academic-team-management-gap pattern — 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-Eestate) 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.