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

Research library · updated 2026-06-17 · public

Tesla Optimus public evidence ladder — SEC-backed claims vs packaging boundaries

Date: 2026-06-17 As-of: 2026-06-17 10:29 UTC Owner: Hugo / Genius Team Research owner: Finance / Charlie AGT-002 Status: SYNTHESIS_CANDIDATE Visibility: PUBLIC Output target: morning-review artifact / slide 11 Tesla card / /robotics/tesla-optimus/ packaging guardrail

Related files:

  • knowledge/robotics-research-harness.md
  • knowledge/robotics-tesla-optimus-public-evidence-card-2026-06-16.md
  • knowledge/robotics-tesla-optimus-sec-filing-excerpt-appendix-2026-06-16.md
  • knowledge/robotics-tesla-optimus-s4-to-s5-checklist-v1.md
  • knowledge/robotics-morning-review-claim-check-matrix-tesla-figure-unitree-2026-06-17.md

Public-safety flag: PUBLIC if used as written. Excludes Hugo private portfolio information, private channel checks, supplier rumors, paid-report excerpts, and buy / sell / hold language.

0. One-line answer

截至 2026-06-17，Tesla Optimus 的公开强证据不是“已商业化”，而是“SEC filing-backed manufacturing intent has become measurable”：从 2025-04-22 到 2026-04-22 的 Tesla 8-K Exhibit 99.1 序列显示 Optimus 从 Fremont pilot production line、production-line installation、Gen 3 mass-production-intended design，推进到 Fremont 1M robots/year designed line 与 Gigafactory Texas 10M robots/year long-term designed line；但公开 filings 仍未披露 actual output、accepted/deployed units、line yield、runtime/intervention、factory ROI/payback、external paid customers、robot revenue、gross margin 或 service burden。🟢 Tesla primary filings; 🟠 Charlie claim-boundary synthesis as of 2026-06-17.

1. Why this artifact exists

已有 Tesla Optimus 文件足够厚。本文件只做一个增量：把 Tesla 的公开证据压成一条可包装、可核验、不会误导的 evidence ladder。

Codex / Warrior 压缩时要保留三句话：

  1. Tesla 的强项是 filing-backed capacity intent / manufacturing infrastructure。🟢
  2. 1M / 10M 是 designed / long-term designed capacity，不是 current production。🟢
  3. Tesla 仍缺 S5 scaled-commercial-economics proof。🟠

2. Public evidence ladder

RungDated source anchorWhat is safe to sayWhat must not be inferredGrade / signal
1. Pilot-line intentQ1 2025 Form 8-K Exhibit 99.1, filed 2025-04-22, says Tesla remained on track for Optimus builds on the Fremont pilot production line in 2025 and wider deployment of bots doing useful work across factories.Tesla had moved Optimus beyond pure demo language into pilot-production and internal-factory-use intent.No unit count, accepted builds, task KPI, uptime, intervention, productivity, safety record, cost saving or payback.🟢 primary; S3/S4
2. Strategic identityQ2 2025 Form 8-K Exhibit 99.1, filed 2025-07-23, frames Tesla as transitioning toward AI, robotics and related services.Robotics became part of Tesla's official quarterly strategic language.Does not prove Optimus output, customer demand, revenue, margin or financial materiality.🟢 primary; S3
3. Production-line installationQ3 2025 Form 8-K Exhibit 99.1, filed 2025-10-22, says first-generation production lines for Optimus were being installed in anticipation of volume production.This is stronger than a viral demo: production-line installation appears in SEC-filed materials.“In anticipation” is not actual volume production, accepted units, utilization, sell-through or economics.🟢 primary; S4
4. Gen 3 mass-production designQ4 2025 Form 8-K Exhibit 99.1, filed 2026-01-28, says Gen 3 is the first Optimus design meant for mass production; preparations were underway for the first production line, including supply-chain readiness; start of production planned before end-2026; eventual planned capacity 1M robots/year.Tesla disclosed mass-production design intent, supply-chain-readiness work, a planned start before end-2026, and a 1M/year planned-capacity anchor.Planned start and planned capacity are not actual output, realized yield, unit cost, accepted deployment, supplier revenue or customer economics.🟢 primary; S4
5. Tesla's own anti-hype caveatQ4 2025 / Q1 2026 capacity tables include Robotics / Optimus construction language and the caveat that installed capacity is not current production rate.Tesla itself gives the packaging rule: capacity / construction should not be read as current production.No current run-rate, utilization, output or economic proof.🟢 primary; S4 caution
6. Fremont 1M designed lineQ1 2026 Form 8-K Exhibit 99.1, filed 2026-04-22, says preparations for the first large-scale Optimus factory would begin shortly in Q2; the first-generation line designed for 1M robots/year would replace Model S/X lines in Fremont.This is the strongest public Tesla Optimus capacity-intent anchor: named site, line replacement, 1M/year designed capacity.Does not prove the line is completed, utilized, yielding accepted robots or profitable. 1M/year is designed capacity, not current production.🟢 primary; S4
7. Texas 10M long-term lineQ1 2026 Form 8-K Exhibit 99.1, filed 2026-04-22, says Tesla was preparing Gigafactory Texas for a second-generation line designed for long-term annual production capacity of 10M robots.Texas adds second-site, long-term scale ambition and the largest disclosed humanoid capacity-intent number in this research set. 🟠 comparative synthesis.Not a dated production forecast, not current capacity, not customer demand, not revenue, not margin.🟢 primary; S4
8. Digital Optimus / AI stackQ1 2026 Form 8-K Exhibit 99.1, filed 2026-04-22, links Digital Optimus to AI development, real-world AI, vehicles and humanoid robots.Tesla is connecting Optimus to AI / inference / physical-AI infrastructure, not just a standalone hardware program.Does not prove vehicle AI transfers cleanly to manipulation, lowers intervention, creates software revenue, or improves robot unit economics.🟢 primary fact; 🟠 implication; S3/S4

3. Signal vs noise

Signal

  • Repeated Optimus / robotics references across five SEC-filed quarterly updates from 2025-04-22 to 2026-04-22. 🟢
  • Named manufacturing sites / lines: Fremont pilot line, Fremont first-generation line, Gigafactory Texas second-generation line. 🟢
  • Quantified designed capacity: Fremont 1M robots/year and Texas long-term 10M robots/year. 🟢 as of Q1 2026 filing.
  • Production-line installation and supply-chain-readiness language, which is stronger than demo-only evidence. 🟢
  • Tesla's own caveat that installed capacity is not current production rate. 🟢

Noise unless upgraded

  • “Tesla is producing 1M Optimus robots/year” or “Tesla has 10M/year Optimus capacity today.” 🔴 Unsupported by reviewed filings.
  • Viral Optimus videos without robot count, runtime, intervention, reset/failure, safety and task denominator. 🔴/🟠
  • Supplier exposure inferred from Tesla capacity ambition without named customer, order value, filing revenue, teardown, or official relationship. 🔴
  • Treating internal Tesla-factory intent as external paid-customer proof. 🟠
  • Treating Digital Optimus / AI-stack language as humanoid manipulation economics proof. 🟠

4. S4-to-S5 proof debt

Tesla Optimus should remain S4 unless multiple S5 gates are disclosed by primary or high-quality customer-side sources:

S5 gateEvidence requiredCurrent status as of 2026-06-17Grade needed
Actual outputQuarterly Optimus produced units, accepted units, completed build count or deployed-unit count.Missing; filings disclose designed capacity / line preparation, not output.🟢 filing / official update
Line yield and bottlenecksFPY, EOL yield, uptime, bottleneck station, scrap/rework, component constraint, cost curve.Missing; Tesla caveats mention production-rate dependency but not solved metrics.🟢 filing / official update
Internal deployment denominatorRobot count by Tesla site, task, shift pattern and operating duration.Missing; Q1 2025 referenced wider factory deployment intent without denominator.🟢 Tesla official; 🟡 credible factory reporting as support
Runtime / interventionFleet runtime hours, uptime %, intervention rate, reset frequency, safety incidents.Missing in reviewed filings.🟢 Tesla/customer source
Factory ROI / paybackThroughput improvement, labor-hour substitution, cost saving, payback months, productivity uplift.Missing in reviewed filings.🟢/🟡
External commercial proofNamed customer, robot count, contract value, delivery schedule, repeat order, revenue recognition.Missing; no external paid Optimus customer disclosed in reviewed filings.🟢 customer / filing
Financial materialityRobot revenue, gross margin, warranty/service/support burden, capex or segment disclosure.Missing; no robot revenue / GM disclosed in reviewed filings.🟢 filing / audited disclosure

5. Packaging-ready wording

Suggested card title:

Tesla Optimus: capacity ambition ≠ commercialization proof

Three-card compression:

  1. Proven by filings 🟢

    • Fremont pilot-line builds in 2025.
    • Production lines installed in anticipation of volume production.
    • Gen 3 is the first design meant for mass production.
    • Fremont first-generation line designed for 1M robots/year.
    • Texas second-generation line designed for long-term 10M robots/year.

  2. Why it matters 🟠

    • Optimus has become one of the clearest public capacity-intent curves in humanoid robotics.
    • Tesla is linking Optimus to manufacturing, AI / inference infrastructure, and long-term physical-AI strategy.
    • The filing sequence is stronger than demos, but weaker than operating economics.

  3. Missing before S5 🔴 / 🟠

    • Actual output / accepted units.
    • Line yield / uptime / bottlenecks.
    • Deployed robot count by site and task.
    • Runtime / intervention / reset / safety data.
    • Factory ROI / payback.
    • External paid customers.
    • Robot revenue / gross margin / service burden.

Required footer:

Evidence map only. No TSLA trade recommendation. Designed capacity is not current production. Internal factory intent is not external paid-customer economics.

6. Common misconceptions

Misconception 1: “1M / 10M means Tesla already has mass production.”

Correction: the disclosed language is designed / long-term designed capacity, and Tesla's own capacity-table caveat says installed capacity is not current production rate. 🟢

Misconception 2: “Optimus in Tesla factories equals commercialization.”

Correction: internal deployment can be the first economic proof only if Tesla discloses robot count, task denominator, runtime, intervention, safety, productivity improvement and payback. 🟠

Misconception 3: “Tesla capacity validates all humanoid suppliers.”

Correction: supplier value capture requires confirmed relationship, order/revenue, value content, ASP, gross margin, architecture fit and replacement risk. Tesla may vertically integrate or redesign modules. 🟠

Misconception 4: “Digital Optimus means humanoid autonomy is solved.”

Correction: AI-stack language must translate into lower intervention, higher task success, shorter deployment time, safer operation, and better payback before it becomes S5 economics. 🟠

7. Think Deeper questions

  1. If Optimus first scales inside Tesla factories, should the first S5 proof be factory productivity/payback rather than robot revenue?
  2. Which disclosure would change the thesis more: 10,000 produced units, 1,000 accepted internal deployments, or 1,000,000 autonomous runtime hours with low intervention?
  3. If Tesla discloses output before yield / intervention, should the evidence upgrade to production proof but remain below economics proof?
  4. If Tesla keeps suppliers unnamed, which public supplier claims deserve attention and which remain rumor?
  5. Is Optimus closer to an EV-style manufacturing ramp, a telecom-style capacity buildout, or a robotics-specific internal-productivity cycle?

8. Source list

9. Public-safe / internal-only boundary

Public-safe:

  • Evidence ladder, source grades, dated anchors, S4/S5 classification, proof-debt dashboard, misconceptions, Think Deeper questions, and footer wording above.

Keep internal-only / exclude:

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