Figure AI / Unitree morning-review decision card — S4 evidence, S5 gap

Date: 2026-06-18 Owner: Hugo / Genius Team Agent: Finance / Charlie AGT-002 Status: SYNTHESIS_CANDIDATE Visibility: PUBLIC Output target: slide 11 / /robotics/companies / morning review Public-safety: yes. No Hugo private portfolio data. No buy / sell / hold language. No external publishing action. Source discipline: 🟢 primary / 🟡 secondary / 🟠 estimate / 🔴 guess.

Related files:

knowledge/robotics-figure-unitree-slide-11-packaging-spec-2026-06-18.md
knowledge/robotics-figure-unitree-source-to-slide-claim-control-2026-06-17.md
knowledge/robotics-figure-unitree-s4-to-s5-conversion-dashboard-v1.md
knowledge/robotics-figure-unitree-evidence-v1.md

0. One-line answer

截至 2026-06-18，Figure AI / Unitree 的最高价值比较不是“谁赢”，而是“哪条 S4 证据曲线更接近 S5”：Figure 已经把 customer-site work 和 manufacturing process 做到可量化；Unitree 已经把 humanoid hardware access cost 和 developer/data workflow 做到可量化。两者都还缺 S5 的完整经济链：accepted units、repeat order、customer ROI/payback、uptime/intervention distribution、revenue/gross margin、support/warranty burden。🟢 Figure / Unitree official-source artifacts reviewed through 2026-06-18; 🟠 Charlie synthesis.

Morning-review sentence:

Figure measures work; Unitree lowers access cost. Neither yet proves scaled commercial economics.

Do not say:

Figure 已经商业化胜出；Unitree 已经证明低价 humanoid 会大规模普及。

Reason: reviewed public evidence supports S3/S4 progress, not S5 economics. 🟢/🟠 as of 2026-06-18.

1. Decision card

Question	Figure AI answer	Unitree / 宇树 answer	Signal grade	What would upgrade it
What has become measurable?	Named-customer task/runtime + factory production process. BMW disclosure: 11-month deployment, 10-hour weekday shifts, 90,000+ parts loaded, 1,250+ runtime hours, contribution to 30,000+ BMW X3 vehicles. BotQ: 350+ Figure 03 delivered, cadence 1/day -> 1/hour, EOL FPY >80%, battery-line FPY 99.3%, 500+ battery packs, 9,000+ actuators across 10+ SKUs.	Public humanoid price/access + developer/data workflow. R1 AIR from US$4,900 / R1 from US$5,900; G1 from US$13.5K; H2 US$29,900. H2 Plus references NVIDIA Jetson T5000, FP4 2,070 TFLOPS, Isaac GR00T / TeleOp / Sim; G1-D describes data acquisition, processing, labeling, data-asset management, model training and inference.	Figure: S4 deployment/manufacturing KPI. Unitree: S3/S4 cost-access/platform KPI.	Figure: robot count by site, paid contract value, customer ROI/payback, repeat order, uptime/intervention, revenue/margin. Unitree: shipment scale, industrial uptime, customer mix, repeat order, gross margin, warranty/support, paid software/tooling revenue.
Which bottleneck is each testing?	Can humanoids do bounded industrial work repeatedly enough to justify deployment?	Can low-cost humanoid bodies widen the experiment base enough to accelerate data/model/deployment learning?	🟠 synthesis from primary-source anchors.	Customer-confirmed economics for Figure; shipment/reliability/economics disclosure for Unitree.
What is the main overclaim risk?	Treating BMW task KPI or BotQ production KPI as scaled commercial economics.	Treating low price, NVIDIA/Isaac references, or developer tooling as reliability, adoption, or margin proof.	🟢/🟠	Require economics before S5 language.
What should Codex package?	A deployment KPI card, not a “winner” card.	A cost-access curve card, not an “adoption proven” card.	🟠	Keep bottom guardrail visible.

2. Source-backed evidence anchors

Figure AI — deployment/manufacturing KPI curve

BMW production-site evidence: Figure disclosed that F.02 contributed to production of 30,000 BMW X3 vehicles at BMW Spartanburg, after 11 months of deployment, 10-hour weekday shifts, 90,000+ parts loaded and 1,250+ runtime hours. 🟢 Figure official post dated 2025-11-19; reviewed in existing artifacts through 2026-06-18.

Signal: S4. This is materially stronger than a demo because it includes a named customer, duration, task denominator and runtime denominator. 🟠 Charlie synthesis.

Not yet S5: public sources reviewed do not disclose robot count at BMW, paid contract value, customer-confirmed ROI/payback, uptime/intervention distribution, repeat order, Figure revenue, gross margin or service burden. 🟢 absence in reviewed official sources; 🟠 evidence-debt classification.

BotQ manufacturing evidence: Figure disclosed 350+ Figure 03 delivered, production cadence moving from 1/day to 1/hour, EOL first-pass yield >80%, battery-line first-pass yield 99.3%, 500+ battery packs and 9,000+ actuators across 10+ SKUs. 🟢 Figure official post dated 2026-04-29; reviewed through 2026-06-18.

Signal: S4. This supports production-process learning, not yet economically utilized fleet proof. 🟠 Charlie synthesis.

Not yet S5: delivered robots / cadence / yield do not equal customer accepted units, utilization, revenue, gross margin, warranty cost or service burden. 🟢/🟠.

Catalyst expansion surface: Figure announced a commercial agreement with Catalyst Brands to deploy humanoids into Catalyst’s distribution and logistics network, starting at Reno, Nevada Distribution Logistics Center. 🟢 Figure official post dated 2026-05-26; 🟠 economics undisclosed.

Signal: S3/S4. It is a useful expansion signal because it names a customer and starting site, but remains below S5 until count/value/payback/repeat order are disclosed. 🟢/🟠.

Unitree / 宇树 — cost-access/platform curve

Public humanoid price anchors: Unitree official pages list R1 AIR from US$4,900 / R1 from US$5,900, G1 from US$13.5K and H2 at US$29,900, tax/shipping excluded. 🟢 Unitree R1/G1/H2 official pages captured 2026-06-10 in prior artifacts; reviewed through 2026-06-18.

Signal: S3/S4. Low public prices can widen the experiment base and pressure the industry cost curve. 🟠 Charlie synthesis.

Not yet S5: price does not prove shipment scale, industrial reliability, autonomy, gross margin, warranty/support burden, repeat order or customer ROI. 🟢/🟠.

Developer/data workflow surface: Unitree H2 Plus references NVIDIA Jetson T5000, FP4 2,070 TFLOPS, Isaac GR00T / Isaac TeleOp / Isaac Sim; G1-D describes data acquisition, processing, labeling, review, data-asset management, model training and inference tools. 🟢 Unitree H2 Plus / G1-D official pages captured 2026-06-10; reviewed through 2026-06-18.

Signal: S3/S4. This suggests the bottleneck is moving from robot-body availability toward data/model/deployment workflow. 🟠 Charlie synthesis.

Not yet S5: tooling pages do not prove active developer count, paid software/tooling revenue, deployment adoption, model performance, attach rate or margin. 🟢/🟠.

3. Signal vs noise for morning review

Signal:

Figure BMW task/runtime denominator: 11 months, 10-hour weekday shifts, 90,000+ parts, 1,250+ runtime hours, 30,000+ X3 vehicles. 🟢 / S4 / as of 2025-11-19 post.
Figure BotQ manufacturing denominator: 350+ robots, 1/day -> 1/hour cadence, >80% EOL FPY, 99.3% battery-line FPY, 9,000+ actuators. 🟢 / S4 / as of 2026-04-29 post.
Figure Catalyst named commercial expansion surface: starting at Reno DLC, but terms undisclosed. 🟢/🟠 / S3-S4 / as of 2026-05-26 post.
Unitree low-price humanoid anchors below US$30K for reviewed R1/G1/H2 pages. 🟢 / S3-S4 / pages captured 2026-06-10.
Unitree developer/data workflow references across H2 Plus / G1-D. 🟢/🟠 / S3-S4 / pages captured 2026-06-10.

Noise unless upgraded:

Winner-ranking language before S5 economics. 🟠.
Customer logos without robot count, contract value, repeat order, ROI/payback or task KPI. 🟢/🟠.
Production output interpreted as economically utilized fleet. 🟢/🟠.
Low price interpreted as industrial reliability or sustainable margin. 🟢/🟠.
NVIDIA / Isaac alignment interpreted as Unitree-owned software economics. 🟢/🟠.
Viral demos without duration, task repetition, failure/intervention rate or customer context. 🔴/🟠.

4. S5 upgrade gates

Figure upgrades only if at least 3 of the following appear together in public official/customer/filing-quality sources:

Robot count by customer/site. 🟢 needed.
Paid contract value, revenue, backlog or recurring economics. 🟢 needed.
Repeat order or multi-site rollout with comparable KPI package. 🟢 needed.
Customer-confirmed productivity gain, ROI or payback. 🟢/🟡 needed.
Uptime, failure distribution, intervention rate and safety record over months. 🟢/🟡 needed.
Figure deployment revenue, gross margin, warranty/support burden or service cost. 🟢 needed.

Unitree upgrades only if at least 3 of the following appear together in public official/customer/filing-quality sources:

Humanoid shipment volume by product line and customer type. 🟢 needed.
Industrial deployment KPI: uptime, intervention, task output, repeat order. 🟢/🟡 needed.
Gross margin and warranty/support cost by humanoid category. 🟢 needed.
Paid software/data/tooling revenue or measured developer ecosystem traction. 🟢/🟠 needed.
Customer ROI/payback or productivity economics. 🟢/🟡 needed.
Long-duration reliability evidence outside demo/lab settings. 🟢/🟡 needed.

5. Codex packaging note

Recommended module title:

Figure vs Unitree: deployment KPI vs cost-access curve

Required footer:

Evidence map only. No winner ranking. No trade recommendation. Deployment KPI ≠ full economics; low price ≠ reliability or margin.

Packaging constraints:

Use Figure as the deployment/manufacturing KPI card.
Use Unitree as the cost-access/developer-platform KPI card.
Keep S5 evidence debt visible, not hidden in notes.
Do not include Hugo private portfolio data, private channel checks, paid-report excerpts, supplier rumors, tax/legal context or buy/sell/hold language.
Do not imply either company has proven scaled-commercial-economics.

6. Source list

Figure AI:

Figure, “F.02 Contributed to the Production of 30,000 Cars at BMW,” 2025-11-19: https://www.figure.ai/news/production-at-bmw 🟢
Figure, “Ramping Figure 03 Production,” 2026-04-29: https://www.figure.ai/news/ramping-figure-03-production 🟢
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 🟢 for announcement; 🟠 for economics because terms undisclosed.
Figure, “Introducing Helix 02: Full-Body Autonomy,” 2026-01-27: https://www.figure.ai/news/helix-02 🟢 for company claim; 🟠 for commercial implication.

Unitree / 宇树:

Unitree R1 product page: https://www.unitree.com/R1 🟢 captured 2026-06-10 in prior artifacts.
Unitree G1 product page: https://www.unitree.com/g1/ 🟢 captured 2026-06-10 in prior artifacts.
Unitree H2 product page: https://www.unitree.com/H2 🟢 captured 2026-06-10 in prior artifacts.
Unitree H2 Plus product page: https://www.unitree.com/H2plus 🟢 captured 2026-06-10 in prior artifacts.
Unitree G1-D product page: https://www.unitree.com/G1-D 🟢 captured 2026-06-10 in prior artifacts.

Internal synthesis / claim-control:

knowledge/robotics-figure-unitree-slide-11-packaging-spec-2026-06-18.md 🟠.
knowledge/robotics-figure-unitree-source-to-slide-claim-control-2026-06-17.md 🟠.
knowledge/robotics-figure-unitree-s4-to-s5-conversion-dashboard-v1.md 🟠.
Charlie morning-review decision-card synthesis, 2026-06-18. 🟠.

7. Public-safe flag

Public-safe as evidence map / morning-review card.

Exclude:

Hugo private portfolio data, position weights, watchlist sizing, entry/exit logic or private rationale.
Any trade recommendation or buy / sell / hold language.
Private channel checks, paid-report excerpts, unverified supplier/customer rumors.
Any claim that Figure has proven scaled humanoid economics.
Any claim that Unitree low price proves shipment scale, reliability, margin or ROI.