Investment ResearchTopic research, evidence curves, and cycle signals
HomeSign inEN

Research library · updated 2026-06-14 · public

Robotics why-now slide-3 catalyst pack v2

Date: 2026-06-14 Owner: Hugo / Genius Team Agent: Finance / Charlie AGT-002 Status: public-safe research artifact for morning review Visibility: PUBLIC Target use: slide 3 / /robotics/why-now / speaker notes Related files:

  • knowledge/robotics-research-harness.md
  • knowledge/robotics-sharing-backlog.md
  • knowledge/robotics-why-now-catalysts-v1.md
  • knowledge/robotics-morning-review-why-now-evidence-ladder-v1.md
  • knowledge/robotics-why-now-catalyst-evidence-card-v1.md

Public-safety: yes. No Hugo private portfolio data. No trade recommendation.

0. One-line answer

截至 2026-06-14，“Why robotics now” 最适合公开压成一句话：机器人不是突然进入 S5 商业化终局，而是公开证据单位发生了变化——AI/model/data stack、低价硬件入口、客户现场 KPI、制造/产能 KPI、以及 filing-backed revenue / supplier validation 同时变得可量化，因此研究应从 demo watching 升级为 S3/S4 evidence tracking。🟢 primary sources reviewed in existing robotics files through 2026-06-13; 🟠 Charlie synthesis as of 2026-06-14.

1. Slide-3 core question

Deck 第 3 页不要回答 “机器人是不是大机会”；应该回答：

为什么现在值得更认真跟踪？哪些证据已经从噪音升级为信号？哪些仍然不能支持 S5 scaled-commercial-economics proof？

Current answer:

  • Signal: 现在有 5 条可追踪证据流，不再只有 demo video / TAM slide / customer logo。🟢/🟠
  • Constraint: 这些证据多数仍是 S3/S4；缺 customer-confirmed ROI、repeat orders、low intervention rate、uptime、gross margin、service cost 和 material revenue durability。🟠 as-of 2026-06-14.

2. Five source-backed catalysts for slide 3

CatalystAs-of evidence anchorWhat changedSignal levelMissing before S5
1. AI / model / data stack became trackableNVIDIA Isaac GR00T official developer page describes a humanoid robotics platform spanning open robot foundation model, data pipelines, simulation, middleware/runtime and Jetson Thor; Physical Intelligence π0 / openpi / π0.5 / π0.7 and Skild official posts provide additional robot-model-layer anchors. 🟢 reviewed in robotics-morning-review-why-now-evidence-ladder-v1.md, as of 2026-06-13.Research unit shifts from isolated movement demos to model/data/simulation/deployment tooling.S3/S4 platform-formation signalCustomer-site reduction in intervention rate, setup time, failure rate or labor cost; software/model revenue; multi-OEM paid adoption.
2. Humanoid hardware cost/access has public price anchorsUnitree official product pages list R1 AIR from US$4,900 / R1 from US$5,900, G1 from US$13.5K, and H2 at US$29,900; H2 Plus and G1-D add NVIDIA/Isaac and data/model workflow surfaces. 🟢 Unitree official pages accessed 2026-06-10 in existing knowledge base.More developers, labs, integrators and early customers can experiment with humanoid bodies; hardware access broadens the trial surface.S3 cost-access / developer-platform signalShipments by product line, customer mix, repeat orders, reliability, warranty/support cost, gross margin, industrial uptime.
3. Customer-site deployment KPI is appearingFigure official BMW post dated 2025-11-19 reports an 11-month Figure 02 deployment at BMW Spartanburg, 10-hour weekday shifts, 90,000+ parts loaded, 1,250+ runtime hours, and contribution to 30,000+ BMW X3 vehicles. 🟢 checked 2026-06-10. Agility/GXO public releases add a logistics RaaS/tote-movement benchmark. 🟢 checked in existing evidence ladder as of 2026-06-13.Evidence moves from “robot can perform a demo” to runtime/task/customer-site metrics that can be tracked and falsified.S4 deployment KPI signalRobot count, contract value, customer-confirmed ROI/payback, uptime, intervention rate, repeat order, service burden.
4. Manufacturing / capacity evidence is numericFigure BotQ official post dated 2026-04-29 reports 350+ Figure 03 robots delivered, cadence from 1/day to 1/hour, EOL FPY >80%, battery-line FPY 99.3%, 500+ battery packs and 9,000+ actuators. 🟢 Tesla Q1 2026 SEC Form 8-K Exhibit 99.1 filed 2026-04-22 says the Fremont first-generation Optimus line is designed for 1M robots/year and Texas second-generation line for long-term annual capacity of 10M robots. 🟢Humanoid claims can now be tested against cadence, yield, line utilization, component output and capacity caveats.S4 manufacturing / capacity signalActual quarterly output, accepted units, utilization, yield over time, cost curve, field failure rate, delivery/customer acceptance, revenue/margin.
5. Filing-backed revenue / supplier validation is emergingUBTECH 2025 annual report disclosed 2026-04-15: revenue RMB 2.001bn (+53.3% YoY), full-size embodied humanoid revenue RMB 820.557m (+2,203.7% YoY), 1,079 full-size embodied humanoid units sold, >6,000 annualized full-size humanoid capacity at end-2025. 🟢 绿的谐波 2025 annual report disclosed 2026-04-23: revenue RMB 570.714m (+47.31% YoY), net profit RMB 124.367m (+121.42% YoY), harmonic reducer sales 425,158 units (+72.48% YoY), “工业及具身智能机器人零部件” revenue RMB 422.528m with 34.88% gross margin. 🟢Parts of the robotics chain move from concept mapping to filing-backed revenue, units and margin disclosure.S4 filing-backed validation signalHumanoid-specific margin, named customer utilization, repeat orders, customer ROI/payback, service/warranty cost, named OEM supplier relationship, single-robot value content.

3. What changed versus old robotics cycles

Old robotics cycles often relied on:

  • viral demos; 🔴 weak evidence if no runtime/intervention context.
  • TAM slides; 🔴 not evidence of value capture.
  • customer logos; 🟠 useful only if upgraded with deployment/economics.
  • “mass production soon” language without units/yield/customer acceptance; 🟠 weak.
  • unsourced supplier lists; 🔴 not public-safe evidence.

2025–2026 evidence is stronger because several anchors are now measurable:

  • runtime hours / task volume / shift schedule: Figure BMW 1,250+ runtime hours and 90,000+ parts loaded. 🟢 as of 2025-11-19.
  • production cadence / FPY / component count: Figure BotQ 1/day to 1/hour, EOL FPY >80%, battery-line FPY 99.3%. 🟢 as of 2026-04-29.
  • designed capacity in filings: Tesla Fremont 1M/year designed Optimus line; Texas 10M/year long-term designed line. 🟢 as of 2026-04-22; note: designed capacity ≠ current output.
  • public price anchors: Unitree R1/G1/H2 price pages. 🟢 accessed 2026-06-10.
  • filing-backed robot revenue / component revenue: UBTECH and 绿的谐波 2025 annual reports. 🟢 disclosed 2026-04-15 / 2026-04-23.

4. Signal vs noise for public slide

Signal

SignalWhy it mattersEvidence grade
Customer-site runtime and task KPIStronger than demo because it can be repeated, measured and falsified.🟢 Figure BMW; Agility/GXO logistics benchmark
Production cadence and first-pass yieldMoves discussion toward manufacturing-process readiness.🟢 Figure BotQ
Designed capacity with caveatUseful capacity-intent signal if not confused with actual output.🟢 Tesla SEC filings
Low public price anchorsExpands experiment base and pressures value migration.🟢 Unitree official product pages
Filing-backed humanoid revenue / unitsStronger than customer logos or “robotics TAM” slides.🟢 UBTECH 2025 annual report
Filing-backed supplier revenue / unit salesStronger than concept-stock narrative, but still not named-OEM proof.🟢 绿的谐波 2025 annual report
Open model / tooling / data-stack formationMakes AI-robotics progress more trackable and comparable.🟢 NVIDIA / PI / Skild official sources

Noise unless upgraded

NoiseCorrect interpretation
Viral demoNeeds task duration, reset count, intervention rate and customer context.
Customer logoNeeds robot count, task KPI, contract value and ROI/payback.
Designed capacityNot current production; use Tesla caveat: installed capacity is not current production rate. 🟢
Low robot priceNot reliability, margin, support cost, shipment scale or ROI.
Supplier product fitNot named OEM customer, order, teardown or revenue proof.
Funding / valuationNot revenue, margin or deployment proof.
Stock-price moveNot operating evidence; use only as market-context, not as proof.

5. S5 upgrade checklist

Upgrade the sector-level why-now claim from “S3/S4 evidence tracking phase” toward “S5 commercial-economics proof” only if several of these appear together:

  1. Multi-customer paid deployments with robot count, task KPI, contract value and repeat orders. 🟢 required future evidence.
  2. Customer-confirmed ROI/payback or productivity improvement, not only vendor-disclosed KPI. 🟢 required future evidence.
  3. Uptime, intervention rate, failure/safety incident rate and service burden over long shifts and multiple sites. 🟢 required future evidence.
  4. Actual production output and accepted customer deliveries, not only designed capacity or production intent. 🟢 required future evidence.
  5. Humanoid-specific revenue, gross margin, warranty/service cost and cash conversion in filings. 🟢 required future evidence.
  6. Supplier relationship confirmed by named OEM, filing-backed order/revenue, identifiable production part, or customer-confirmed design win. 🟢 required future evidence.

As of 2026-06-14, reviewed public evidence has meaningful S3/S4 anchors, but not enough complete S5 proof. 🟠 synthesis.

6. Public slide copy

Title: Why robotics now: the evidence unit changed

Subtitle: Not “robots are solved,” but “the industry finally has trackable signals beyond demos.”

Five cards:

  1. AI stack became trackable

    • GR00T / PI / Skild turn robot AI into model + data + simulation + deployment infrastructure. 🟢
    • Missing: software revenue, attach rate, customer economics. 🟠

  2. Hardware got cheaper and more accessible

    • Unitree R1 from US$5.9K, G1 from US$13.5K, H2 US$29.9K. 🟢
    • Missing: shipments, reliability, support cost, gross margin, ROI. 🟠

  3. Deployment KPI appeared

    • Figure BMW: 1,250+ runtime hours, 90,000+ parts loaded, 10-hour weekday shifts. 🟢
    • Missing: robot count, contract value, payback, intervention rate. 🟠

  4. Manufacturing / capacity numbers appeared

    • Figure BotQ: 350+ robots delivered, 1/day to 1/hour cadence, >80% EOL FPY. Tesla: 1M/year and 10M/year designed-capacity language. 🟢
    • Missing: actual output, utilization, yield over time, margin. 🟠

  5. Filing-backed validation appeared

    • UBTECH: RMB 820.557m full-size embodied humanoid revenue and 1,079 units sold in 2025. 绿的谐波: RMB 422.528m “工业及具身智能机器人零部件” revenue and 425,158 harmonic reducers sold in 2025. 🟢
    • Missing: humanoid-specific margin, named customer/OEM proof, repeat economics. 🟠

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

7. Common misconceptions

  1. Misconception: “Why now means commercialization is solved.”

    • Correction: Why-now means evidence quality improved enough to track; it does not prove return-cycle economics. 🟠

  2. Misconception: “Tesla 1M / 10M capacity means Tesla is already producing those robots.”

    • Correction: Tesla’s language is designed / long-term capacity; actual Optimus output, customer deliveries, revenue and unit economics are not disclosed in the reviewed filings. 🟢

  3. Misconception: “Figure BMW KPI proves ROI.”

    • Correction: Runtime and task count are strong S4 evidence, but ROI/payback, robot count, contract value and intervention rate are still missing. 🟢/🟠

  4. Misconception: “Unitree low price proves market adoption.”

    • Correction: Price lowers experiment cost; it does not prove shipments, reliability, support cost, gross margin or industrial deployment. 🟢/🟠

  5. Misconception: “Supplier filing revenue proves named humanoid OEM exposure.”

    • Correction: Filing-backed revenue and shipment growth are real signals, but named OEM, humanoid-specific revenue, contract value and margin durability are separate gates. 🟢/🟠

8. Think Deeper questions

  1. Which disclosure should become the minimum bar for humanoid deployment: runtime hours, intervention rate, task success rate, payback, or repeat orders?
  2. If humanoid bodies get cheaper, does value migrate to OEM scale, model/data platforms, deployment integrators, maintenance, or architecture-specific suppliers?
  3. Will the first S5 proof come from a private OEM deployment, Tesla internal factory productivity, a public filing like UBTECH, or supplier order disclosure?
  4. If robots are deployed as RaaS / managed service, how much gross margin is consumed by teleoperation, maintenance, safety and integration?
  5. Which evidence would falsify the why-now thesis fastest: weak repeat orders, high intervention rate, poor uptime, negative service margin, or persistent customer ROI silence?

9. Source list

Primary / official sources already reviewed in the robotics knowledge base:

Internal synthesis sources:

  • knowledge/robotics-why-now-catalysts-v1.md 🟠
  • knowledge/robotics-morning-review-why-now-evidence-ladder-v1.md 🟠
  • knowledge/robotics-why-now-catalyst-evidence-card-v1.md 🟠
  • knowledge/robotics-demand-side-labor-market-evidence-v1.md 🟠
  • knowledge/robotics-deployment-to-s5-benchmark-bridge-v1.md 🟠

10. Public-safety flag

Public-safe if used as an evidence map and not as a recommendation.

Do not publish:

  • Hugo private portfolio data.
  • Buy / sell / hold language.
  • Winner ranking.
  • Claims that S4 signals equal S5 commercialization.
  • Claims that designed capacity equals current production.
  • Claims that Unitree low price proves reliability, gross margin or industrial adoption.
  • Claims that supplier product fit or robot-related revenue proves named Tesla / Figure / Unitree / AgiBot / UBTECH supplier relationship without named customer, order, filing, teardown or revenue evidence.