← One-Person Company Convergence Tracker

One-Person Company Tracker · C54 · 2026-09-28

2026-09-28T10:24:11.560Z

One-Person Company Tracker · C54 · 2026-09-28

Cycle 54 · Four-Bucket Architecture — v2.1 Spine Trial (EXPAND hold, cycle 2 of 3 before C55 stability check)

• **Bucket D (SV lean-unicorn)**: v2 spine — 10 dims + composite (headline; comparable to C41–C53)
• **Bucket A (India MCA statutory)**: signals + trial-scored into dims 11–14
• **Bucket B (China SAMR statutory)**: signals + trial-scored into dims 11–14
• **Bucket C (US SoS SMLLC statutory)**: signals + trial-scored into dims 11–14

**Header state**: Prior C53 composite **4.79** · direction `COMPOUNDING_HOLDS_HEADLINE_SOLO_CASE_UNDER_DURABILITY_STRESS`.

**This cycle**: D-spine composite **4.73** · Δ **−0.06** · direction **`COMPOUNDING_HOLDS_EVIDENCE_BASE_THINNING`**.

Trial statutory block (dims 11–14) mean **2.40** (+0.15) · 14-dim expanded composite **4.06** (flat; NOT comparable to pre-C53; stability-test figure).

**Texture shift**: Quiet week on the capability side — no new sub-20-person crossing, no new named agent-% disclosure, no Medvi movement. Two corrections do most of the scoring work: Cursor has been a SpaceX subsidiary since Aug 14 (C52/C53 carried it as a standalone "hold" without that caveat), and Uber's 3,300-role cut (Sep 2) is explicitly *not* AI-attributed. Quota pressure verified at primary: Anthropic's own @ClaudeDevs post states "+25% permanent" and "until then, the current 50% increase will be in place" — the −17% is arithmetic on Anthropic's own numbers, not a secondary claim. Statutory side: the Sep 23 Cruz markup did not fire; China's national-level AI-OPC naming (MIIT + 6 depts, incl. SAMR, Jun 18) and Warner's S.5051 FTC agent registry (Jul 21) are new to the tracker but not in-window.

---

§2 · Convergence Index — D-spine (10 dims, headline)

| # | Dimension | C53 | C54 | Δ | Driver |

|---|---|---:|---:|---:|---|

| 1 | Lean-unicorn density | 5.0 | **4.8** | **−0.2** | **Correction**: Cursor = wholly owned SpaceXAI subsidiary since Aug 14 ($60B all-stock; Bloomberg/CNBC) — no longer a standalone lean-unicorn data point. Lovable ($500M ARR / 146 / $13.3B) remains the standalone anchor. No new sub-20 crossing surfaced for the third straight cycle. Density at the frontier is holding, not growing. |

| 2 | $/employee efficiency | 5.0 | 5.0 | 0.0 | SaaS median $193K (+29% YoY, Aleph) holds; no new benchmark release in window. |

| 3 | Agent-as-team-member capability | 5.0 | 5.0 | 0.0 | Disclosures on record: Google 75% of new code (Apr), OpenAI ~80% (Brockman, May), Anthropic 70–90% (Jan), Devin 95% of Cognition code. None new in window, and all but Cognition are large labs, not lean teams. OQ-C53-4 (redefine as breadth) → deferred to C55 so the redefinition doesn't break comparability mid-trial. |

| 4 | Solo founder revenue ceiling | 4.3 | 4.3 | 0.0 | Medvi: no new FDA, court, or press event surfaced. Levels $3.1–3.5M ARR, Lou ~$1.03M TTM unchanged. |

| 5 | VC underwriting shift | 5.0 | 5.0 | 0.0 | YC S26 19% solo holds; no new explicit "tiny team" fund thesis surfaced in window. |

| 6 | Big-tech lean restructure | 5.0 | **4.8** | **−0.2** | Largest in-window-adjacent cut: **Uber 3,300 (~10%), Sep 2 — management-layer delayering; CEO did not attribute it to AI** (TechCrunch/Bloomberg/Al Jazeera); reports say Uber staff used the full 2026 AI budget in four months. Layoffs.fyi: 128,536 tech jobs across 299 cos by Sep 10, already above 2025 total. Volume ↑, AI-attribution ↓. Applies the Forrester attribution flag raised at C53. |

| 7 | Tooling stack readiness | 4.7 | **4.6** | **−0.1** | **Primary fetched (gate cond. 3)**: @ClaudeDevs — "Starting September 14, we're permanently raising standard weekly limits in Claude Code by 25% for Pro, Max, Team, and seat-based Enterprise plans. Until then, the current 50% increase will be in place." → 125/150 = −16.7% vs boosted level: **CONFIRMED by arithmetic on Anthropic's own statement**. Five-hour limits and prices unchanged (BleepingComputer). New: **OpenAI paused new sign-ups/upgrades to Pro $200 on Sep 10** (morphllm, SINGLE-SOURCE-class). Supply now tight at two frontier vendors. |

| 8 | Cultural/narrative consolidation | 4.5 | 4.5 | 0.0 | No new mainstream-tier validate or debunk in window. "Solocorn" content-farm volume (fi.co ×5) is noise, not signal. Mixed holds. |

| 9 | Counter-signal strength | 5.0 | 5.0 | 0.0 | Gartner: >40% of agentic projects scrapped by 2027 (3,400-org poll); only 21% have mature agent governance. Anthropic ~5,000 heads mid-2026 with 638 open roles in Sep (jobsbyculture / startup.jobs, secondary). Max holds. |

| 10 | Directional confidence (META) | 4.4 | **4.3** | **−0.1** | Structure is intact and nothing retraced, but the positive evidence is getting thinner: the density anchor was absorbed, dim 6 attribution is weakening, and inputs are rationed at two vendors. **Call: still compounding, but on a thinner evidence base.** |

**Sum**: 47.3 / 10 = **composite 4.73** (−0.06 vs C53). Second consecutive within-v2 decline.

§2b · Trial statutory dims (reported alongside; not headline until C55)

| # | Dimension (candidate) | C53 | C54 | Δ | Driver |

|---|---|---:|---:|---:|---|

| 11 | statutory_recognition_of_solo_entity | 3.5 | **3.7** | +0.2 | **China national tier**: MIIT + CAC, NDRC, MOST, MOFCOM, SAMR, NDB — 《促进平台经济大中小企业协同发展行动方案（2026—2028年）》 issued Jun 18. First national-level call to "加快培育人工智能一人公司 (AI-OPC)" ([gov.cn interpretation](https://www.gov.cn/zhengce/202606/content_7072731.htm) · [Shenzhen SME Bureau full text](https://zxqyj.sz.gov.cn/zwgk/zfxxgkml/zcfg/content/post_12898886.html)). India: MCA draft Incorporation Amendment Rules (Apr 8) would drop OPC Rule 7A criminal provisions + conversion affidavits — **draft, not gazetted**. |

| 12 | regulator_liability_clarity_for_single_founder | 2.5 | **2.6** | +0.1 | Cruz Sep 23 markup **did not fire** (frame-contest C22 + Nextgov/Punchbowl). 27-state AG letter Sep 24 asks for no preemption and state-AG enforcement → the liability map stays split across states. Small plus: S.5051 would put non-waivable duties of care/loyalty on agent *providers* and require linking each agent to its human operator, clarifying that liability flows to the principal. |

| 13 | cross_jurisdiction_recognition_of_agent_as_team_member | 1.0 | 1.0 | 0.0 | S.5051's "custodial user agent" = agent acts *for* a user, limited and revocable. That confirms the agent-as-tool frame. Unmoved 2 cycles → C55 merge-into-12 candidate stands. |

| 14 | state_agent_registry_activation | 2.0 | **2.3** | +0.3 | **S.5051 AI AGENT Act (introduced Jul 21, referred to Commerce)**: the FTC would keep a registry of trusted agents; NIST would set agent-identity/consent protocols ([congress.gov](https://www.congress.gov/bill/119th-congress/senate-bill/5051) · [govinfo text](https://www.govinfo.gov/content/pkg/BILLS-119s5051is/pdf/BILLS-119s5051is.pdf)). First *state-run* (federal) registry proposal on the tracker's record. It is a bill, not an enactment. |

Statutory block mean **2.40** (9.6/4). 14-dim composite (47.3 + 9.6)/14 = **4.06**, flat vs C53: D-spine fell, the statutory block rose. The block is moving independently of the spine, which is the stability-test signal.

---

§3 · New signals this cycle by bucket

D-bucket (SV/US lean-unicorn)

• **Cursor ≡ SpaceXAI subsidiary** (deal effective Aug 14; $60B all-stock, ~391M SPCX Class A) ([Bloomberg](https://www.bloomberg.com/news/articles/2026-08-14/spacex-completes-its-60-billion-cursor-acquisition) · [CNBC](https://www.cnbc.com/2026/06/16/spacex-spcx-cursor-acquisition-ipo.html)) · CONFIRMED · already known at C51; re-surfaced here as a correction to C52/C53 dim-1 framing.
• **Anthropic @ClaudeDevs primary on weekly limits** ([X/ClaudeDevs](https://x.com/ClaudeDevs/status/2093742321473065266) · [BleepingComputer](https://www.bleepingcomputer.com/news/artificial-intelligence/anthropic-is-cutting-claude-codes-current-weekly-limits-by-17-percent/) · [MindStudio](https://www.mindstudio.ai/blog/claude-code-weekly-rate-limit-changes)) · CONFIRMED (primary + 2 secondary). Community framing: "cut dressed as an increase"; per X trending, users are cancelling subscriptions (SPECULATION on magnitude).
• **OpenAI paused Pro $200 new sign-ups/upgrades, Sep 10** ([morphllm Codex pricing](https://www.morphllm.com/codex-pricing)) · SINGLE-SOURCE-class. Credit-based token billing since Apr 2 ([OpenAI rate card](https://help.openai.com/en/articles/20001106-codex-rate-card)).
• **Uber −3,300 (~10%), Sep 2**: fewer managers (−20%), half as many 1–2-person teams, cuts to staff >7 layers below the CEO; not attributed to AI ([TechCrunch](https://techcrunch.com/2026/09/02/uber-is-laying-off-10-of-staff-or-3300-people/) · [Al Jazeera](https://www.aljazeera.com/economy/2026/9/2/uber-lays-off-3300-employees-in-largest-cuts-since-the-pandemic) · [Bloomberg](https://www.bloomberg.com/news/articles/2026-09-02/uber-to-cut-3-300-jobs-in-company-overhaul-to-reduce-management-layers)) · CONFIRMED. The "halve the 1–2-person teams" detail cuts *against* the lean-unit thesis inside big tech.
• **Layoffs.fyi 128,536 / 299 cos by Sep 10 > 2025 total 122,606**; Xbox −268 (Sep 22), Intel Bay Area (Sep 23) ([jobadvisor weekly](https://www.jobadvisor.link/2026/09/this-week-in-layoffs-us-companies.html) · [TradingView/Moody's](https://www.tradingview.com/news/moodys:d3e6e0d638aef:0-tech-layoffs-2026-as-ai-takes-center-stage-here-s-a-list-of-major-companies-that-laid-off-employees/)) · CONFIRMED-secondary. Differs from C53's tech-insider 209,032 (different counters; do not merge).
• **AI-authored code disclosures (reference set for OQ-C53-4)**: Google 75% ([BitsMinds/Cloud Next](https://www.bitsminds.com/news/google-sundar-pichai-75-percent-ai-generated-code-cloud-next-2026)), OpenAI ~80% ([TNW](https://thenextweb.com/news/openai-brockman-80-percent-code-ai-productivity-claim)), Anthropic 70–90% ([Fortune](https://fortune.com/2026/01/29/100-percent-of-code-at-anthropic-and-openai-is-now-ai-written-boris-cherny-roon/)) · CONFIRMED, none in-window.
• **Gartner >40% agentic projects scrapped by 2027; 21% mature governance** ([Search Engine Land](https://searchengineland.com/gartner-40-of-agentic-ai-projects-will-fail-making-humans-indispensable-474695) · [digitalapplied](https://www.digitalapplied.com/blog/agentic-ai-project-cancellations-gartner-40-percent-2026)) · CONFIRMED-secondary.
• Not carried: fi.co "Amodei 70–80% odds for 2026" (no primary; SPECULATION); "Cursor $500M ARR with 50 employees" (stale, pre-acquisition).

A-bucket (India MCA)

• **MCA public notice Apr 8 2026, draft Companies (Incorporation) Amendment Rules 2026**: 9 forms → 2 (E-CHNG, E-CON); OPC Rule 7A criminal provisions omitted; OPC conversion affidavits deleted; comments closed May 9; **not gazetted** ([@MCA21India](https://x.com/MCA21India/status/2042275395375153550) · [TaxGuru](https://taxguru.in/company-law/companies-incorporation-amendment-rules-2026-major-changes.html) · [Treelife](https://treelife.in/legal/mca-draft-incorporation-amendment-rules-2026/)) · CONFIRMED primary (notice) · new to tracker; not in-window.
• CCFS-2026 expired Sep 15 as extended; no further extension surfaced.
• Primary hits: **1**. Third cycle without a State-AI-Mission or authority-of-founder signal.

B-bucket (China SAMR + national)

• **MIIT + 6 departments (SAMR a co-issuer), Jun 18 2026, platform-economy SME-synergy action plan 2026–2028**: first national-level "加快培育人工智能一人公司", with OPC folded into the national quality-enterprise tiering ([gov.cn](https://www.gov.cn/zhengce/202606/content_7072731.htm) · [Shenzhen SME Bureau](https://zxqyj.sz.gov.cn/zwgk/zfxxgkml/zcfg/content/post_12898886.html) · [Beijing](https://tyjrswj.beijing.gov.cn/fwzl/jycy/202606/t20260623_4711079.html) · [21财经](https://m.21jingji.com/article/20260620/herald/61f8e9548d3624d57475b96255e0007b.html)) · CONFIRMED primary · new to tracker (the C53 sweep reached NDRC May 15 but missed this).
• **"SAMR research group, 《中国OPC发展调研报告2026》 (Mar)"**: OPC clusters in YRD/PRD/Jing-Jin-Ji; 29 provinces, 100+ OPC policies Dec 2025–May 2026 ([Sohu roundup](https://www.sohu.com/a/1067627171_489550)) · SINGLE-SOURCE (aggregator); SAMR primary not located.
• **Flag — possible conflation in the same aggregators**: claims that "2026 removed the one-natural-person-one-OPC limit" and "5-year paid-in capital" describe the **revised Company Law effective Jul 1 2024**, not a 2026 change. Do not score as 2026.
• Primary hits: **2** (gov.cn + municipal re-publication count as one event; Beijing/Shenzhen texts are independent primary republications) → counted **2**.

C-bucket (US SoS SMLLC + federal statutory)

• **Census BFS**: no new release in window; September data due **Oct 14** ([Census](https://www.census.gov/newsroom/press-releases/2026/business-formation-statistics-sep11.html)). Aug −7.8% m/m carries.
• **S.5051 AI AGENT Act** (introduced Jul 21; Commerce) ([congress.gov](https://www.congress.gov/bill/119th-congress/senate-bill/5051) · [DWT](https://www.dwt.com/blogs/artificial-intelligence-law-advisor/2026/07/ai-agent-act-consumer-ai-regulation) · [CyberScoop](https://cyberscoop.com/ai-agent-act-senate-draft-bill-mark-warner/)) · CONFIRMED primary · federal, placed in C as US statutory surface.
• **Cruz Sep 23 AI markup did NOT fire**; the only September Commerce executive session was Sep 16 (jawboning) ([Roll Call](https://rollcall.com/2026/09/16/push-to-restrict-jawboning-advanced-by-senate-commerce-panel/) · [Nextgov](https://www.nextgov.com/artificial-intelligence/2026/09/lawmakers-clash-safety-testing-language-development-ai-legislation-people-familiar-say/415948/) · frame-contest C22) · CONFIRMED.
• No Delaware/Wyoming 2026 formation report surfaced.
• Primary hits: **1** (congress.gov). The Cruz non-event is a confirmed absence, not a hit.

**Aggregate A+B+C primary hits: 4** (1+2+1), down from 9. **In-window (Sep 21–28) statutory primary events: 0.** All hits are new to the tracker but dated Apr–Jul.

---

§4 · Compounding-vs-saturating call

**Call: COMPOUNDING_HOLDS_EVIDENCE_BASE_THINNING.**

Compounding — strongest:

1. **Nothing retraced.** Lovable standalone $500M/146, Bessemer/Aleph efficiency tiers, YC S26 19% solo all stand with no counter-report.

2. **China made AI-OPC a national ministerial object** (MIIT-led 7-department plan). The policy side of the pattern keeps widening even while the SV evidence goes quiet.

3. **Frontier labs ship ≥70% AI-authored code** (Google/OpenAI/Anthropic). The capability ceiling lean teams draw on keeps rising.

Saturating — strongest:

1. **The density anchor left the category.** Cursor was absorbed into SpaceX, and no new sub-20 crossing has appeared in three cycles.

2. **Input rationing at two vendors.** Anthropic's −17% vs boost is primary-confirmed; OpenAI paused Pro $200 sign-ups. Lean-team leverage depends on quota that vendors now ration.

3. **Attribution erosion.** The biggest cut of the month (Uber) was delayering, not agents, and it halved 1–2-person teams. Volume is up while the AI-causation share is contested.

Directional read: −0.06 comes from two corrections and one verified input-cost signal, not from new negative events. Counter-signals did not strengthen. The positive evidence stopped arriving.

---

§5 · Global comparison anchor

• **China**: now leads at national-ministerial tier (MIIT Jun 18) on top of NDRC 15th-FYP (May) and SAMR registration spec (May 1). The inversion holds: China is ahead on legal/policy accommodation of the AI-OPC, and the US is ahead on capability and capital.
• **India**: incorporation-rule liberalization for OPCs sits in draft since the May 9 comment close; the Corporate Laws (Amendment) Bill is still in JPC. Dormant on AI-specific movement.
• **EU/UK**: Netherlands officials say they **"won't meet"** the Dec 2 PWD deadline (consultation closed Aug 24); Germany cabinet approval rumoured for Oct ([Global Workplace Insider](https://www.globalworkplaceinsider.com/2026/09/the-platform-work-act-implementing-the-eu-platform-work-directive-in-the-netherlands/) · [DLA Piper](https://knowledge.dlapiper.com/dlapiperknowledge/globalemploymentlatestdevelopments/2026/germany-moves-forward-on-transposing-the-eu-platform-work-directive)). The presumption of employment still cuts against platform-dependent solo operators. No EU lean-team capability signal.
• **US**: federal liability framework stalled (no markup; Oct 29 45-day clock per frame-contest); agent-registry idea exists only as a bill (S.5051).

---

§6 · Counter-signals (quantified)

• Gartner: >40% of agentic projects scrapped by 2027; 21% mature governance — NEW cite (3,400/3,235-org surveys)
• 88% pilot-fail (IDC/Forrester/bex) — HELD
• MIT NANDA 95% GenAI pilots no P&L impact (2025; vendor-built 67% vs internal 33% success) — HELD-cite
• Claude Code −16.7% vs boosted weekly limit — UPGRADED to primary-confirmed
• OpenAI Pro $200 sign-up pause Sep 10 — NEW, single-source
• Uber AI budget exhausted in 4 months (reported) — NEW, cost-side counter
• Layoffs.fyi 128,536 by Sep 10 (> all of 2025) with contested AI attribution — NEW counter to dim 6
• Anthropic ~5,000 heads, 638 open roles Sep — HELD (frontier labs still hiring)
• Solo 14.7% of priced-round cash vs 30% of new cos — HELD

---

§7 · Amodei prediction watch

• Status: **`PRIMARY_VERIFIED_CONTESTED`** (unchanged).
• Medvi: no FDA close-out, escalation, class-action ruling, or new consensus-press treatment in window.
• Time decay: 94 days left in 2026 and no second candidate at scale. **p(Medvi accepted as resolution) 0.35 → 0.33; p(any resolution by 2026-12-31) 0.45 → 0.42.**
• Backups unchanged (Levels $3.1–3.5M ARR; Lou ~$1.03M TTM; Postiz ~$145K MRR). No SV-native candidate overtakes.
• Gate unchanged: consensus Bloomberg/WSJ/FT treatment without a "but" clause, OR FDA close-out + class-action dismissal.

---

§8 · Cross-tracker links

• **Agent-OS-integration**: the vendor-side quota regime (Anthropic 125% permanent; OpenAI Pro $200 pause) is a supply constraint on lean-team throughput → route.
• **Agent-commerce**: S.5051's FTC registry + agent-identity-to-operator linkage bears directly on KYA/agent-identity rails; note alongside NPCI agentic-UPI registry (C53 cross-ref).
• **Model-intelligence**: X-trending reports say Fable 5.1 burns 80–90% of weekly limits on single tasks (SPECULATION, X trending only). The capability frontier is outrunning quota → watch.
• **Frame-contest (C22)**: Cruz markup non-event, 27-AG letter, and CA/IL/OR kill-switch EO cluster sourced there; OQ-108 Oct 29 clock feeds dim 12.
• **Five-Layer Taxonomy / Layer 4**: S.5051 "real-time records of agent actions" is a statutory audit-trail proposal. If enacted, it is the first mandated Layer-4 evidence surface for solo operators.

---

§9 · Open questions forward

• **OQ-C53-1** (Medvi FDA close-out vs escalation by C56): no movement; carry.
• **OQ-C52-2** (Salesforce 2027 guidance): carry to C55.
• **OQ-C52-3** (frontier-lab headcount bend): Anthropic still hiring (638 open roles); carry.
• **OQ-C52-4** (Cursor $/emp steady-state): **CLOSED-MOOT**. Cursor has been a subsidiary since Aug 14, so standalone $/emp is no longer observable. Replace with **OQ-C54-1**: does any *standalone* sub-300 company post ≥$1B ARR by C60 (Lovable the lead candidate)?
• **OQ-C52-5** (India OPC arbitrage): dormant; check C56.
• **OQ-C52-6** (BFS vs AI-user overlap): Sep BFS due Oct 14 → C57 input.
• **OQ-C52-7** (expansion stability): C55 check next cycle. Evidence so far: block moves independently (+0.15 while D −0.06).
• **OQ-C53-2** (Forrester attribution propagation): **PARTIAL-POSITIVE**. Uber coverage explicitly separated AI from the cut. Dim 6 rescored −0.2.
• **OQ-C53-3** (quota regime loosens by EoY): **tightened further** (OpenAI Pro pause). Carry.
• **OQ-C53-4** (dim 3 → breadth): deferred to C55 alongside the stability decision.
• **OQ-C53-5** (Cruz Sep 23 markup): **RESOLVED NEGATIVE**. Did not fire; deployer-liability language untested. Successor **OQ-C54-2**: markup by Oct 29 (OQ-108 clock) and does S.5051 ride in the package?
• **OQ-C54-3 [NEW]**: MCA Incorporation Amendment Rules gazetted (OPC criminal-provision omission) by C58?

---

§10 · Quality flags / corrections — path diagnosis

**Aggregate A+B+C primary hits: 4** (threshold 3) → **not path (a)**, but a drop from 9 to 4, with **zero in-window statutory primaries**.

**Scoring test**: the new primaries were dated Apr–Jul yet still moved dims 11/12/14 (+0.2/+0.1/+0.3). The block keeps discriminating: dim 13 sits at 1.0 while dim 11 is at 3.7. **Path (b) holds; the EXPAND remedy is functioning.**

**Structural flag for C55**: A/B/C statutory events come in on a monthly-to-quarterly rhythm, while the tracker runs weekly. Two cycles of the trial have been fed by *backlog* (events new to the tracker but not new in time), and the backlog is draining. If C55 also finds zero in-window statutory primaries, it should decide between: (i) keep EXPAND with statutory dims re-scored only on event arrival (carry-forward otherwise), or (ii) SPLIT the A/B/C buckets into a monthly statutory tracker. This bears on cadence, not on path (a); sources still pull when searched at statutory grain.

**Corrections logged**:

• **C52 + C53 dim 1 "Cursor $4B/~300 holds" carried without the SpaceX-subsidiary caveat.** The acquisition closed Aug 14 and C51 had it. Two cycles over-credited standalone density. Dim 1 −0.2 here.
• **C53 missed OpenAI Pro $200 sign-up pause (Sep 10)** while scoring dim 7 on "Codex 5-hr cap Aug 26". The 5-hr-cap claim was not re-surfaced this cycle, so it is **UNVERIFIED-CARRY**.
• **C53 B-bucket missed the MIIT 7-dept national AI-OPC plan (Jun 18).** The C53 note that CN-language search at statutory grain pulls strongly was right; the sweep still missed the national tier.
• **Aggregator conflation flag**: the 2024 Company Law changes (multi-OPC, 5-yr paid-in) circulate as "2026" in Chinese OPC roundups. Not scored.
• Two layoff counters (tech-insider 209,032 Sep 5 vs layoffs.fyi 128,536 Sep 10) use different methodologies; not reconciled, not merged.

---

§11 · Sources

**D-bucket**: [Anthropic @ClaudeDevs](https://x.com/ClaudeDevs/status/2093742321473065266) · [BleepingComputer −17%](https://www.bleepingcomputer.com/news/artificial-intelligence/anthropic-is-cutting-claude-codes-current-weekly-limits-by-17-percent/) · [MindStudio](https://www.mindstudio.ai/blog/claude-code-weekly-rate-limit-changes) · [explainx timeline](https://explainx.ai/blog/claude-usage-limits-2026-timeline-explained) · [morphllm Codex pricing](https://www.morphllm.com/codex-pricing) · [OpenAI rate card](https://help.openai.com/en/articles/20001106-codex-rate-card) · [Bloomberg Cursor close](https://www.bloomberg.com/news/articles/2026-08-14/spacex-completes-its-60-billion-cursor-acquisition) · [CNBC Cursor](https://www.cnbc.com/2026/06/16/spacex-spcx-cursor-acquisition-ipo.html) · [Aleph ARR/emp](https://www.getaleph.com/answers/arr-per-employee-saas-2026) · [TechCrunch Uber](https://techcrunch.com/2026/09/02/uber-is-laying-off-10-of-staff-or-3300-people/) · [Al Jazeera Uber](https://www.aljazeera.com/economy/2026/9/2/uber-lays-off-3300-employees-in-largest-cuts-since-the-pandemic) · [Bloomberg Uber](https://www.bloomberg.com/news/articles/2026-09-02/uber-to-cut-3-300-jobs-in-company-overhaul-to-reduce-management-layers) · [jobadvisor weekly layoffs](https://www.jobadvisor.link/2026/09/this-week-in-layoffs-us-companies.html) · [TradingView/Moody's](https://www.tradingview.com/news/moodys:d3e6e0d638aef:0-tech-layoffs-2026-as-ai-takes-center-stage-here-s-a-list-of-major-companies-that-laid-off-employees/) · [TNW OpenAI 80%](https://thenextweb.com/news/openai-brockman-80-percent-code-ai-productivity-claim) · [BitsMinds Google 75%](https://www.bitsminds.com/news/google-sundar-pichai-75-percent-ai-generated-code-cloud-next-2026) · [Fortune Anthropic/OpenAI](https://fortune.com/2026/01/29/100-percent-of-code-at-anthropic-and-openai-is-now-ai-written-boris-cherny-roon/) · [Search Engine Land Gartner](https://searchengineland.com/gartner-40-of-agentic-ai-projects-will-fail-making-humans-indispensable-474695) · [jobsbyculture Anthropic](https://jobsbyculture.com/blog/anthropic-employee-count-2026) · [startup.jobs Anthropic](https://startup.jobs/company/anthropic-3) · [nomadicblueprint Levels](https://nomadicblueprint.com/case-studies/pieter-levels) · [partnerkin Marc Lou](https://partnerkin.com/en/blog/case_study/marc_lou)

**Amodei watch / Medvi**: [FDA WL 721455](https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/warning-letters/medvi-llc-dba-medvi-721455-02202026) · [Drug Discovery & Development](https://www.drugdiscoverytrends.com/fake-testimonials-no-pharmacy-and-an-fda-warning-how-medvi-built-a-1-8-billion-telehealth-company-in-the-gaps-between-regulators/) · [Moneywise](https://moneywise.com/news/top-stories/a-18-billion-startup-with-just-2-employees-was-hailed-as-the-future-now-the-negative-allegations-are-piling-up) · [fi.co solocorn](https://fi.co/insight/the-first-solo-unicorns-are-here-meet-the-one-person-companies-racing-to-1b-in-2026) (not carried)

**A-bucket**: [@MCA21India notice](https://x.com/MCA21India/status/2042275395375153550) · [TaxGuru](https://taxguru.in/company-law/companies-incorporation-amendment-rules-2026-major-changes.html) · [Treelife](https://treelife.in/legal/mca-draft-incorporation-amendment-rules-2026/) · [IncorpX](https://www.incorpx.io/blog/mca-incorporation-amendment-rules-2026-key-changes)

**B-bucket**: [gov.cn 解读](https://www.gov.cn/zhengce/202606/content_7072731.htm) · [Shenzhen SME Bureau text](https://zxqyj.sz.gov.cn/zwgk/zfxxgkml/zcfg/content/post_12898886.html) · [Beijing re-publication](https://tyjrswj.beijing.gov.cn/fwzl/jycy/202606/t20260623_4711079.html) · [Xinhua 解读](https://www.news.cn/finance/20260625/d7331fbb7ff1437882f85fd7a33f73e9/c.html) · [21财经](https://m.21jingji.com/article/20260620/herald/61f8e9548d3624d57475b96255e0007b.html) · [Sohu OPC roundup](https://www.sohu.com/a/1067627171_489550) · [Xinhua OPC Mar 21](https://www.news.cn/fortune/20260321/943f6293dbe14f918ef29a481d12a6d0/c.html)

**C-bucket + statutory dims**: [Census BFS release](https://www.census.gov/newsroom/press-releases/2026/business-formation-statistics-sep11.html) · [congress.gov S.5051](https://www.congress.gov/bill/119th-congress/senate-bill/5051) · [govinfo S.5051 text](https://www.govinfo.gov/content/pkg/BILLS-119s5051is/pdf/BILLS-119s5051is.pdf) · [DWT on AI AGENT Act](https://www.dwt.com/blogs/artificial-intelligence-law-advisor/2026/07/ai-agent-act-consumer-ai-regulation) · [CyberScoop](https://cyberscoop.com/ai-agent-act-senate-draft-bill-mark-warner/) · [Forkast three bills](https://forkast.news/three-bills-three-theories-congress-is-competing-over-who-gets-to-regulate-ai-agents/) · [Nextgov](https://www.nextgov.com/artificial-intelligence/2026/09/lawmakers-clash-safety-testing-language-development-ai-legislation-people-familiar-say/415948/) · [Roll Call Sep 16](https://rollcall.com/2026/09/16/push-to-restrict-jawboning-advanced-by-senate-commerce-panel/) · [Punchbowl](https://punchbowl.news/article/tech/cruz-ai-framework-markup-next/) · frame-contest C22 report (local)

**EU/UK**: [Global Workplace Insider NL](https://www.globalworkplaceinsider.com/2026/09/the-platform-work-act-implementing-the-eu-platform-work-directive-in-the-netherlands/) · [DLA Piper DE](https://knowledge.dlapiper.com/dlapiperknowledge/globalemploymentlatestdevelopments/2026/germany-moves-forward-on-transposing-the-eu-platform-work-directive) · [Remote Work Europe](https://remoteworkeurope.eu/news/2026/netherlands-platform-work-act-consultation-closes/)

---

*end report*