Robotics Morning Review Packet — 2026-06-12

Date: 2026-06-12 Owner: Hugo / Genius Team Research owner: Finance / Charlie AGT-002 Visibility: PUBLIC, after excluding private portfolio data Status: morning-review artifact; public-safe synthesis; no trade recommendation Source discipline: 🟢 primary / 🟡 secondary / 🟠 estimate / 🔴 guess Related files:

• knowledge/robotics-research-harness.md
• knowledge/robotics-public-claim-bank-v1.md
• knowledge/robotics-catalyst-map-for-slides-v1.md
• knowledge/robotics-tesla-optimus-primary-claim-map-v1.md
• knowledge/robotics-tesla-figure-unitree-slide-evidence-v1.md
• knowledge/robotics-leaderdrive-case-study-slide-evidence-v3.md
• knowledge/robotics-oem-public-evidence-matrix-v1.md

Public-safety: yes. This file contains no Hugo private portfolio weights, no buy/sell/hold recommendation, and no external publishing action.

0. One-line morning answer

截至 2026-06-12，机器人研究已经可以支持一版 public-safe morning review：核心判断不是“humanoid 已经全面商业化”，而是“公开证据从 demo/newsflow 升级到五类可量化证据”：AI/robot foundation-model 工具链、低价硬件平台、真实 deployment KPI、制造/产能 KPI、以及供应链 filing-backed revenue。当前行业证据是 S4-heavy early deployment / supply-chain validation，不是 S5 scaled-commercial-economics / return cycle。🟢/🟠

1. Morning review structure: 4 blocks, 12 minutes

Block A — Why robotics now: evidence type changed

Approved review line:

机器人现在值得严肃跟踪，不是因为 TAM 故事更大，而是因为证据开始变得可量化、可复盘、可证伪。

Evidence anchors:

Signal / noise:

• Signal: evidence types are now quantifiable and updateable. 🟠 synthesis.
• Noise: “why now” should not be compressed into “all robot names are investable now.” 🔴 unsupported.

Block B — Tesla / Figure / Unitree: three evidence curves, not one horse race

Approved review line:

Tesla / Figure / Unitree 不应该被讲成“谁会赢”。它们分别代表 capacity ambition、deployment KPI、cost/access curve 三条证据曲线。

Source anchors:

• Tesla Q1 2026 Form 8-K Exhibit 99.1, filed 2026-04-22. 🟢
• Figure official posts dated 2025-11-19 and 2026-04-29. 🟢
• Unitree official R1 / G1 / H2 pages, accessed 2026-06-10. 🟢

Signal / noise:

• Signal: the three curves answer different research questions: can it scale production, can it work at customer site, can hardware cost fall enough to broaden experimentation? 🟠 synthesis.
• Noise: a single “winner ranking” would over-compress incompatible evidence types. 🟠/🔴

Block C — Leaderdrive / 绿的谐波: strongest public-safe supplier teaching case

Approved review line:

绿的谐波是一个很好的供应链教学案例：报表信号很强，但不能把“工业及具身智能机器人零部件”收入直接等同于 humanoid-only revenue 或已验证客户经济性。

Evidence anchors:

What is still missing before S5:

• Named humanoid OEM customer. 🟢/🟠 absence in reviewed public filings as of 2026-06-12.
• Contract value / backlog / order volume. 🟢/🟠 absence.
• Humanoid-specific revenue split and humanoid-specific gross margin. 🟢/🟠 absence.
• Single-robot value content validated by OEM BOM / teardown / official supplier disclosure. 🟠/🔴 currently unverified.
• Customer ROI/payback or repeat-order economics. 🟢/🟠 absence.

Signal / noise:

• Signal: filing-backed revenue, profit, unit volume, and Q1 follow-through are stronger than concept-stock language. 🟢
• Noise: share-price move, unnamed customer rumors, fixed reducer-count assumptions, and “robot concept” labels. 🟢/🟠/🔴

Block D — Current conclusion and next evidence to wait for

Current conclusion:

Robotics is now worth a serious tracking system because the evidence is measurable. It is not yet worth public overclaiming because the complete S5 chain is still missing.

S5 chain required:

1. Repeat paid deployment. 🟢 required.
2. Customer-confirmed ROI / payback. 🟢 required.
3. Low intervention rate / uptime / task-success data. 🟢 required.
4. Actual production output / yield / service cost. 🟢 required.
5. Robot revenue / gross margin / supplier revenue split. 🟢 required.

Current state as of 2026-06-12:

• Tesla: S4 capacity-intent / manufacturing infrastructure. 🟢/🟠
• Figure: S4 deployment KPI + manufacturing KPI; economics still undisclosed. 🟢/🟠
• Unitree: S3 cost/access platform; deployment economics still undisclosed. 🟢/🟠
• Leaderdrive: S4 filing-backed supplier signal; humanoid-specific customer/economics still undisclosed. 🟢/🟠
• Industry-level “why now”: S4-heavy evidence-type shift, not S5 return-cycle proof. 🟠 synthesis.

2. Slide packaging recommendation for Codex

Suggested title:

Robotics now: evidence has improved, but S5 is still missing

Suggested 4-card layout:

1. Why now
Evidence type changed: AI stack + low-cost hardware + deployment KPI + manufacturing KPI + supplier filings.

2. OEM evidence curves
Tesla = capacity ambition; Figure = deployment/manufacturing KPI; Unitree = cost/access.

3. Supplier teaching case
Leaderdrive = filing-backed S4 signal, not humanoid winner proof.

4. What upgrades the thesis
Repeat paid deployment + ROI/payback + low intervention + production economics + revenue/margin.

Footer:

Evidence map only. No winner ranking. No trade recommendation. S4-heavy does not equal S5 scaled-commercial economics.

3. Site packaging recommendation for Codex

Use this as a short top section for /robotics or /robotics/signals:

Robotics now: from demos to measurable evidence

The useful public framing is not that humanoid robots are already fully commercialized. It is that the evidence has become measurable. We can now track AI tooling, product price points, runtime/task KPIs, manufacturing yield/cadence, and supplier filing data. That is enough to build a serious field guide, but not enough to declare winners.

The current evidence map has four representative curves: Tesla shows capacity ambition, Figure shows deployment and manufacturing KPIs, Unitree shows the cost/access curve, and Leaderdrive shows filing-backed supply-chain acceleration. Each curve matters. None yet completes the S5 chain of repeat paid deployment, customer ROI, low intervention, production economics, and robot revenue/margin.

4. Common misconceptions for morning review

1. “Designed capacity means actual production.”

   • Correction: Tesla’s 1M / 10M figures are designed capacity / long-term designed annual capacity, not disclosed current production. 🟢 Tesla Q1 2026 Form 8-K; 🟠 framing.

2. “Customer logo means economics are proven.”

   • Correction: Figure’s BMW KPI is strong, but contract value, robot count, customer ROI/payback, repeat-order economics, uptime/intervention, revenue and margin remain undisclosed in reviewed public posts. 🟢/🟠

3. “Low price means durable winner.”

   • Correction: Unitree’s low price expands experimentation and developer access, but does not prove industrial reliability, margin quality, maintenance cost, or customer payback. 🟢/🟠

4. “Supplier revenue equals humanoid-specific revenue.”

   • Correction: Leaderdrive’s “工业及具身智能机器人零部件” line is strong filing evidence, but it is not a humanoid-only customer/economics split. 🟢/🟠

5. “Stock price is order evidence.”

   • Correction: Leaderdrive’s abnormal-trading announcement after a +36.00% three-trading-day move said there was no undisclosed major matter. 🟢 2026-06-09 announcement.

5. Think Deeper questions

1. If Tesla first deploys Optimus internally, should commercialization be measured as cost reduction before external robot sales?
2. If Figure replicates BMW-style KPIs at Catalyst or another named customer, which metric upgrades the thesis first: repeat order, ROI/payback, uptime/intervention, or robot-level gross margin?
3. If Unitree makes humanoid hardware cheap and widely available, does value migrate to software/data/service, or does body hardware become commoditized first?
4. If Leaderdrive’s robot-related revenue continues to grow but gross margin compresses, is that better read as volume validation or OEM bargaining-power pressure?
5. Which public signal is hardest to fake: runtime hours, customer ROI, production yield, repeat orders, or audited segment revenue?

6. Source list

Primary sources 🟢

• NVIDIA Isaac GR00T official developer page: https://developer.nvidia.com/isaac/gr00t — accessed 2026-06-09; used for AI / robot foundation-model tooling claim.
• Tesla Q1 2026 Update, Form 8-K Exhibit 99.1, filed 2026-04-22: https://www.sec.gov/Archives/edgar/data/1318605/000162828026026551/exhibit991.htm — used for Fremont 1M/year and Texas 10M/year Optimus designed-capacity claims.
• Tesla Q1 2025-Q1 2026 Form 8-K Exhibit 99.1 sequence, cited in knowledge/robotics-tesla-optimus-primary-claim-map-v1.md — used for Optimus evidence ladder.
• Figure, “F.02 Contributed to the Production of 30,000 Cars at BMW”, 2025-11-19: https://www.figure.ai/news/production-at-bmw — used for BMW runtime / parts / production-contribution KPIs.
• Figure, “Ramping Figure 03 Production”, 2026-04-29: https://www.figure.ai/news/ramping-figure-03-production — used for BotQ delivered units, cadence, FPY, battery packs, and actuator metrics.
• Figure, “Figure Signs Agreement with Catalyst Brands to Scale Humanoid Operations”, 2026-05-26: https://www.figure.ai/news/figure-signs-agreement-with-catalyst-brands — used only for named commercial-agreement signal; economics undisclosed.
• Unitree R1 / G1 / H2 / H2 Plus / G1-D official product pages, accessed 2026-06-10 via knowledge/robotics-tesla-figure-unitree-slide-evidence-v1.md — used for price / platform claims.
• 绿的谐波 2025 annual report, disclosed 2026-04-23: https://static.sse.com.cn/disclosure/listedinfo/announcement/c/new/2026-04-23/688017_20260423_EU45.pdf — used for 2025 revenue, profit, robot component revenue, gross margin, harmonic reducer production/sales/inventory.
• 绿的谐波 2026 Q1 report, disclosed 2026-04-23: https://static.sse.com.cn/disclosure/listedinfo/announcement/c/new/2026-04-23/688017_20260423_TB4A.pdf — used for Q1 revenue/profit follow-through.
• 绿的谐波 2026-06-09 abnormal trading announcement: https://static.sse.com.cn/disclosure/listedinfo/announcement/c/new/2026-06-09/688017_20260609_MZU2.pdf — used for share-price move and no-undisclosed-major-matter caveat.

Internal synthesis / cross-references 🟠

• knowledge/robotics-public-claim-bank-v1.md
• knowledge/robotics-catalyst-map-for-slides-v1.md
• knowledge/robotics-tesla-optimus-primary-claim-map-v1.md
• knowledge/robotics-tesla-figure-unitree-slide-evidence-v1.md
• knowledge/robotics-leaderdrive-case-study-slide-evidence-v3.md
• knowledge/robotics-oem-public-evidence-matrix-v1.md

7. Public-safety flag

Public-safe: yes, with exclusions.

Do not publish:

• Hugo private portfolio weights or private rationale.
• Buy / sell / hold language for TSLA, 688017.SH, or any robotics-related security.
• Any claim that Tesla designed capacity equals current output.
• Any claim that Figure’s customer announcements prove full economics without ROI/payback and repeat-order disclosure.
• Any claim that Unitree low price proves reliability, margin, or industrial deployment economics.
• Any claim that Leaderdrive’s robot/embodied component line is humanoid-only revenue or named OEM validation.
• Any supplier rumor not backed by primary customer, filing, or official disclosure.