Robotics morning-review packaging priority matrix — what goes on the main line

Date: 2026-06-17 As-of: 2026-06-17T08:49:26Z Owner: Hugo / Genius Team Agent: Finance / Charlie AGT-002 Status: SYNTHESIS_CANDIDATE Visibility: PUBLIC Target use: morning review / Codex slide-site compression guardrail

Related files:

• knowledge/robotics-research-harness.md
• knowledge/robotics-sharing-backlog.md
• knowledge/robotics-codex-handoff.md
• knowledge/robotics-why-now-s5-evidence-debt-dashboard-2026-06-17.md
• knowledge/robotics-morning-review-evidence-first-executive-brief-2026-06-17.md
• knowledge/robotics-morning-review-claim-check-matrix-tesla-figure-unitree-2026-06-17.md
• knowledge/robotics-leaderdrive-public-safe-slide-10-case-study-2026-06-16.md

Public-safety: PUBLIC if used as written. No Hugo private portfolio data. No trade recommendation. No paid-report excerpts. No private channel checks. No unverified supplier/customer rumors.

0. One-line answer

截至 2026-06-17 早审，Codex/Warrior 最应该打包的不是“哪个公司赢”，而是“哪些公开证据已经硬到能进入主线，哪些只能做脚注”。主线优先级应是：Figure customer-site KPI + BotQ manufacturing KPI、Tesla SEC capacity-intent、Unitree low-cost access、UBTECH/绿的谐波 filing-backed denominator，最后用 NVIDIA/AI stack 解释 why-now 的上游变化。整体仍是 S3/S4 evidence-tracking phase，不是 S5 scaled-commercial-economics proof。🟢/🟠

Recommended headline:

The evidence changed before the economics were proven.

Canonical footer:

Evidence map only. No winner ranking. No trade recommendation. S3/S4 evidence is not S5 economics.

1. Why this artifact exists

已有文件已经覆盖 Tesla Optimus、Figure / Unitree、Leaderdrive / 绿的谐波、why-now catalysts。早晨 review 的剩余风险不是材料不够，而是压缩时把 S4 证据误写成 S5 结论。

本文件只做一件事：给 Codex/Warrior 一个 packaging priority matrix，决定：

1. 哪些证据应该放在主线；
2. 哪些证据可以放侧栏；
3. 哪些证据只能脚注；
4. 哪些表述必须禁止。

2. Packaging priority matrix

3. Recommended morning-review structure

Slide/site section A — opening

Title: Why robotics now: measurable evidence, unpaid economics

Body: The public evidence unit changed. Robotics is moving from viral demos toward runtime, task count, manufacturing yield, price anchors, capacity intent, revenue/unit lines and supplier filing signals. That is enough for S3/S4 research priority. It is not enough for S5 scaled commercial economics.

Source grade: 🟠 Charlie synthesis, grounded in P0/P1 primary anchors above.

Slide/site section B — four main cards

Use only four cards if space is tight:

1. Figure: deployment KPI

   • BMW 90,000+ parts / 1,250+ runtime hours / 30,000+ X3 contribution. 🟢 as of 2025-11-19; live checked in adjacent file 2026-06-17.
   • Missing: ROI, intervention, repeat order, margin. 🟠/🔴 not disclosed.

2. Tesla: capacity intent

   • Fremont 1M/year designed line; Texas 10M/year long-term designed line. 🟢 Q1 2026 8-K filed 2026-04-22; 🟠 local extraction 2026-06-16.
   • Missing: actual output, yield, accepted units, economics. 🟠/🔴 not disclosed.

3. Unitree: cost-access

   • US$4,900–29,900 humanoid price anchors across R1/G1/H2 family. 🟢 official pages checked in adjacent file 2026-06-17.
   • Missing: shipments, uptime, gross margin, support burden, ROI. 🟠/🔴 not disclosed.

4. Filings: denominator evidence

   • UBTECH humanoid revenue/unit disclosures; 绿的谐波 reducer/component revenue, sales volume and gross margin. 🟢 2025 annual reports.
   • Missing: customer ROI, intervention, named humanoid OEM exposure, humanoid-only supplier economics. 🟠/🔴.

Slide/site section C — proof-debt strip

S5 proof debt: accepted robots by quarter; uptime/intervention distribution; repeat orders; customer ROI/payback; revenue quality; gross margin after service burden; warranty/liability/risk allocation.

4. Signal vs noise

Signal

• Customer-site runtime / task-count KPI with named industrial context. Figure BMW: 1,250+ runtime hours and 90,000+ parts. 🟢
• Manufacturing KPI with robot count / yield / battery / actuator anchors. Figure BotQ: 350+ robots, >80% EOL FPY, 99.3% battery-line FPY, 9,000+ actuators. 🟢
• SEC-filed capacity intent. Tesla Q1 2026 8-K: 1M/year Fremont designed line and 10M/year long-term Texas designed capacity. 🟢/🟠
• Official low-cost hardware price anchors. Unitree: US$4,900–29,900 humanoid page anchors. 🟢
• Filing-backed revenue/unit/margin denominators. UBTECH and 绿的谐波 2025 annual reports. 🟢

Noise unless upgraded

• Viral videos without runtime, reset/intervention rules, task denominator, safety/maintenance burden, or customer-site context. 🔴/🟠
• Capacity target treated as actual production. 🔴 if written as fact; 🟠 if framed as interpretation risk.
• Customer logo treated as ROI/payback. 🔴/🟠
• Low price treated as reliability, gross margin, autonomy, or customer ROI proof. 🔴/🟠
• Supplier/BOM multiplication without named OEM, ASP, volume, margin and architecture-fit evidence. 🔴/🟠
• Funding rounds treated as deployment economics. 🟡 for funding fact; 🟠 for economics inference.

5. Must-not-write list for Codex / Warrior

Do not write:

1. “Tesla is producing 1M Optimus robots per year” or “Tesla already has 10M/year Optimus capacity.”

   • Correct: Tesla disclosed designed / long-term designed capacity anchors for planned/installed lines; reviewed filings do not disclose actual output. 🟢/🟠

2. “Figure has proven humanoid economics.”

   • Correct: Figure has strong S4 deployment and manufacturing KPI; reviewed public sources still do not disclose ROI/payback, intervention rate, repeat order, revenue, gross margin or service burden. 🟢/🟠

3. “Unitree wins because it is cheap.”

   • Correct: Unitree has the clearest public low-cost access signal; low price does not prove reliability, autonomy, margin, shipment scale or customer economics. 🟢/🟠

4. “绿的谐波 is confirmed as Tesla/Figure/Unitree supplier.”

   • Correct: reviewed public filings support strong robot-component / reducer evidence but not named humanoid OEM customer exposure. 🟢/🟠

5. “NVIDIA robotics stack is already financially material.”

   • Correct: NVIDIA official materials support robotics stack formation; reviewed public sources here do not establish robotics revenue materiality or paid humanoid software economics. 🟢/🟠

6. “Robotics is S5 because the evidence is now abundant.”

   • Correct: abundance of S3/S4 evidence changes research priority; S5 still requires repeatable economics and financial materiality. 🟠

6. What would change the morning-review conclusion

Upgrade toward S5 if at least three of these become primary/customer-side/filing-backed:

1. Accepted robots by customer/site and by quarter. 🟢
2. Uptime / intervention-rate distribution under production conditions. 🟢
3. Repeat orders or multi-site expansion with comparable workflow economics. 🟢
4. Customer ROI/payback or enough disclosed inputs for a transparent estimate. 🟢/🟠
5. Robot revenue, gross margin, service/support cost, warranty reserve and cash conversion. 🟢
6. Supplier order/revenue exposure with named OEM or humanoid-specific filing split. 🟢
7. Clear evidence that AI/model/data stack lowers deployment time, failure rate, or support burden in customer sites. 🟢/🟡

Downgrade if:

• KPI disclosures stay one-off for 12–24 months. 🟠
• Capacity expands without field utilization disclosure. 🟠
• Low-cost hardware creates high service burden and weak repeat adoption. 🟠
• Supplier revenue improves cyclically but named humanoid exposure remains absent. 🟠
• Customer deployments remain logos/agreements without robot count, ROI or repeat order. 🟠

7. Public-site paragraph

The reason to revisit robotics in 2026 is not that the TAM slide got larger. It is that the evidence unit changed. Public sources now show customer-site runtime and task counts, manufacturing yield and robot-count anchors, SEC-filed capacity intent, low-cost humanoid price points, and filing-backed revenue/unit/margin denominators. That is not commercialization proof. It is a better research clock: the next question is which evidence chain reaches repeat orders, low intervention, customer ROI, revenue quality and durable margin first.

Footer: Evidence map only. No winner ranking. No trade recommendation. S3/S4 evidence is not S5 economics.

8. Common misconception corrections

1. “More evidence means the sector is ready.”

   • Correction: more measurable evidence means the research problem improved; readiness still needs S5 economics. 🟠

2. “The company with the biggest capacity number leads.”

   • Correction: capacity intent is not output, yield, utilization or economics. 🟢/🟠

3. “The company with the clearest customer KPI has already won.”

   • Correction: customer KPI is strong S4, but not ROI/payback, repeat order or margin proof. 🟢/🟠

4. “The cheapest robot wins the platform.”

   • Correction: cheap hardware may widen experimentation and compress margins at the same time; value may migrate elsewhere. 🟢/🟠

5. “Supplier filing growth equals humanoid design-win proof.”

   • Correction: supplier filing growth is stronger than rumor, but named OEM / volume / ASP / margin durability must still be verified. 🟢/🟠

9. Think Deeper questions

1. Which first S5 bridge matters most for public-market timing: accepted robot count, intervention-rate decline, repeat order, humanoid gross margin, or customer ROI/payback?
2. If Tesla proves internal-factory ROI before external robot sales, should the market value Optimus as a cost-saving platform or a product business?
3. If Figure has the cleanest deployment KPI but remains private, which public signals can capture the learning without relying on supplier rumors?
4. If Unitree compresses humanoid ASP, does value migrate from OEM hardware toward model/data, compute, integrators, services, components, or customer productivity?
5. If supplier filings improve before named OEM disclosure, should the evidence stay S4 or be discounted toward S3?

10. Source list

• Figure, “F.02 Contributed to the Production of 30,000 Cars at BMW”, 2025-11-19, captured/live-checked in robotics-morning-review-claim-check-matrix-tesla-figure-unitree-2026-06-17.md. 🟢
• Figure, “Ramping Figure 03 Production”, 2026-04-29, captured/live-checked in adjacent claim-check artifact; exact 1/hour cadence wording should be manually reopened before direct quotation. 🟢/🟠
• Figure, “Figure Signs Agreement with Catalyst Brands to Scale Humanoid Operations”, 2026-05-26, captured/live-checked in adjacent claim-check artifact. 🟢 for announcement; 🟠 economics undisclosed.
• Tesla Q1 2026 Update, Form 8-K Exhibit 99.1, filed 2026-04-22, via robotics-tesla-optimus-sec-filing-excerpt-appendix-2026-06-16.md; live SEC request returned HTTP 403 in this environment on 2026-06-17. 🟢 original filing; 🟠 extraction/access note.
• Unitree R1 / G1 / H2 / H2 Plus / G1-D official product pages, live-checked in adjacent claim-check artifact on 2026-06-17. 🟢
• UBTECH Robotics Corp Ltd Annual Report 2025, disclosed 2026-04-15, captured in prior why-now / company-map artifacts. 🟢
• 绿的谐波 2025 annual report, disclosed 2026-04-23, captured in Leaderdrive / why-now artifacts. 🟢
• NVIDIA Isaac GR00T / robotics official and technical materials captured through 2026-06-16 in prior why-now artifacts. 🟢 for official/technical claims; 🟠 commercialization implication.
• IFR World Robotics 2025 service-robot material, cited in prior morning-review executive brief. 🟢/🟡 as captured; 🟠 humanoid implication.
• Charlie cross-source packaging-priority synthesis, 2026-06-17. 🟠

11. Public-safe / internal-only boundary

Public-safe:

• Evidence matrix, source grades, dated anchors, S3/S4/S5 classification, packaging priority, caveats and public-site paragraph above.

Keep internal-only / exclude:

• Hugo private portfolio data, watchlist sizing, purchase price, trade rationale, private channel checks, paid-report excerpts, tax/legal context, buy/sell/hold language, and any supplier-customer inference not backed by primary or high-quality customer-side evidence.