Tesla Optimus source-to-slide claim control sheet — morning review guardrail
Date: 2026-06-17
As-of: 2026-06-17 20:29 UTC
Owner: Hugo / Genius Team
Research owner: Finance / Charlie AGT-002
Status: SYNTHESIS_CANDIDATE
Visibility: PUBLIC
Output target: slide 11 Tesla card / /robotics/tesla-optimus/ refresh guardrail / morning-review speaker notes
Related files:
knowledge/robotics-research-harness.mdknowledge/robotics-tesla-optimus-public-evidence-card-2026-06-16.mdknowledge/robotics-tesla-optimus-sec-filing-excerpt-appendix-2026-06-16.mdknowledge/robotics-tesla-optimus-public-evidence-ladder-2026-06-17.mdknowledge/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 data, private channel checks, paid-report excerpts, supplier rumors, tax/legal context, and buy / sell / hold language.
0. One-line answer
截至 2026-06-17，Tesla Optimus 最适合在晨会/slide 中被压成一句话：Tesla 已经把 Optimus 从 demo narrative 推进到 SEC filing-backed capacity-intent / manufacturing-infrastructure evidence；但 1M / 10M 是 designed / long-term designed capacity，不是 current production、accepted units、customer demand、revenue、margin 或 ROI/payback。🟢 Tesla 8-K Exhibit 99.1 sequence filed 2025-04-22 through 2026-04-22; 🟠 Charlie synthesis as of 2026-06-17.
1. Why this file exists
已有 Tesla 文件已经覆盖完整证据链。本文件只做一个晨会增量：把 “source claim → safe slide wording → forbidden wording → S5 upgrade evidence” 放到同一张控制表，供 Codex / Warrior 压缩时避免过度表达。
Core packaging rule:
Capacity ambition is real evidence; commercialization proof is still missing.
2. Source-to-slide control table
| Evidence anchor | Safe slide wording | Do not write | Why it matters | Grade / signal |
|---|---|---|---|---|
| Q1 2025 Form 8-K Exhibit 99.1, filed 2025-04-22: Fremont pilot production line builds in 2025 and wider factory deployment intent. | “Optimus moved from demo language into pilot-line / internal-factory-use intent.” | “Tesla has disclosed Optimus factory productivity” or “Optimus is already replacing labor at scale.” | The first important boundary is internal use vs proven economics. Factory intent is a stronger signal than a video but still lacks denominator. | 🟢 primary; S3/S4 |
| Q2 2025 Form 8-K Exhibit 99.1, filed 2025-07-23: Tesla frames itself as moving toward AI, robotics and related services. | “Robotics became part of Tesla’s official reporting identity.” | “Robotics is financially material to Tesla.” | Strategy language matters for capital allocation, but not for valuation without revenue / capex / margin disclosure. | 🟢 primary; S3 |
| Q3 2025 Form 8-K Exhibit 99.1, filed 2025-10-22: first-generation Optimus production lines being installed in anticipation of volume production. | “Production-line installation is S4 manufacturing-infrastructure evidence.” | “Tesla is in volume production.” | “Installed in anticipation” is a real manufacturing step, but not actual output or yield. | 🟢 primary; S4 |
| Q4 2025 Form 8-K Exhibit 99.1, filed 2026-01-28: Gen 3 first design meant for mass production; first line preparation and supply-chain readiness; planned start before end-2026; eventual planned capacity 1M robots/year. | “Tesla disclosed mass-production design intent, supply-chain readiness work, and 1M/year planned-capacity ambition.” | “Tesla will produce 1M Optimus robots in 2026” or “Tesla supplier winners are confirmed.” | The 1M anchor is useful only if labeled as planned / eventual capacity, not realized volume. | 🟢 primary; S4 |
| Tesla capacity-table caveat in Q4 2025 / Q1 2026 materials: installed capacity is not current production rate. | “Tesla’s own caveat says capacity should not be read as current production.” | “Installed capacity equals run-rate.” | This is the anti-hype footnote that should stay on any public Tesla card. | 🟢 primary; S4 caution |
| Q1 2026 Form 8-K Exhibit 99.1, filed 2026-04-22: first large-scale Optimus factory preparation to begin shortly in Q2; Fremont first-generation line designed for 1M robots/year replacing Model S/X lines. | “Named Fremont site + line replacement + 1M/year designed capacity is Tesla’s strongest public capacity-intent anchor.” | “Fremont is producing 1M/year now.” | This is the cleanest primary-source proof that Optimus is being planned as a manufacturing-scale program. | 🟢 primary; S4 |
| Q1 2026 Form 8-K Exhibit 99.1, filed 2026-04-22: Gigafactory Texas second-generation line designed for long-term annual production capacity of 10M robots. | “Texas adds a second-site, long-term 10M/year designed-capacity ambition.” | “Tesla has 10M/year Optimus capacity today” or “10M demand is proven.” | The number is high-signal for ambition, but not evidence of demand, output, or economics. | 🟢 primary; S4 |
| Q1 2026 Form 8-K Exhibit 99.1, filed 2026-04-22: Digital Optimus / AI development / real-world AI language. | “Tesla links Optimus to its AI / inference / physical-AI stack.” | “Vehicle AI already solves humanoid manipulation economics.” | AI-stack linkage is directionally important but must convert into lower intervention and better payback before S5. | 🟢 primary fact; 🟠 implication; S3/S4 |
3. Slide 11 recommended compression
Title: Tesla Optimus: capacity ambition ≠ commercialization proof
Three bullets:
- What is proven 🟢: SEC-filed sequence shows Fremont pilot-line intent, production-line installation, Gen 3 mass-production design, Fremont 1M/year designed line, and Texas 10M/year long-term designed line.
- Why it matters 🟠: Optimus has become one of the clearest public capacity-intent curves in humanoid robotics; the evidence moved beyond demo-only updates.
- What is missing 🔴/🟠: actual output, accepted units, line yield, runtime/intervention, task denominator, internal ROI/payback, external paid customers, robot revenue, gross margin and 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.
4. Speaker-note version for morning review
Tesla is the cleanest public-company example of a humanoid program becoming measurable through filings, but the measurement is still capacity intent, not economics. The 2025–2026 8-K sequence lets us say the program has moved into production-line preparation and large-scale factory planning. It does not let us say Tesla is producing 1M robots/year, has 10M/year current capacity, has solved manipulation economics, or has validated outside customer demand. The right stage is S4, not S5.
5. Signal vs noise
Signal
- Five-quarter SEC-filed Optimus / robotics sequence from 2025-04-22 to 2026-04-22. 🟢
- Named production geography: Fremont and Gigafactory Texas. 🟢
- Quantified designed-capacity anchors: 1M robots/year Fremont and 10M robots/year long-term Texas. 🟢 as of Q1 2026 filing.
- Production-line installation, mass-production design, and supply-chain-readiness language. 🟢
- Tesla’s own capacity caveat that installed capacity is not current production rate. 🟢
Noise unless upgraded
- Viral Optimus demo clips without robot count, task denominator, runtime, intervention, safety, reset/failure and productivity data. 🔴/🟠
- Any claim that 1M / 10M designed capacity equals current production. 🔴 unsupported by reviewed filings.
- Supplier winner claims inferred from Tesla capacity ambition without named customer, order value, revenue, teardown, or official relationship. 🔴
- Treating internal Tesla factory intent as external paid commercialization. 🟠
- Treating Digital Optimus / AI-stack language as proof of manipulation economics. 🟠
6. S5 upgrade gates
Tesla Optimus should remain S4 until primary or high-quality customer-side evidence discloses several of the following together:
| S5 gate | What would upgrade the evidence | Current status as of 2026-06-17 | Source grade needed |
|---|---|---|---|
| Actual output | Quarterly produced units, accepted units, or deployed-unit count. | Missing. Filings disclose designed capacity and line preparation, not output. | 🟢 Tesla filing / official update |
| Manufacturing quality | FPY, EOL yield, bottleneck stations, rework, uptime, component constraints, cost curve. | Missing. | 🟢 official / filing |
| Internal deployment denominator | Robot count by factory, task, shift, duration and operating environment. | Missing beyond broad internal-use intent. | 🟢 Tesla / credible factory source |
| Runtime / intervention | Autonomous runtime hours, intervention rate, reset rate, safety incidents, task success distribution. | Missing. | 🟢 / 🟡 |
| Factory economics | Throughput improvement, labor-hour substitution, cost saving, ROI/payback months. | Missing. | 🟢 / 🟡 |
| External commercialization | Named customer, contract value, delivery schedule, repeat order, price, revenue recognition. | Missing in reviewed filings. | 🟢 customer / filing |
| Financial materiality | Robot revenue, gross margin, service/support/warranty burden, segment disclosure, capex materiality. | Missing. | 🟢 audited / filing |
7. Common misconceptions
Misconception 1: “1M / 10M means mass production is already happening.”
Correction: Tesla’s language is designed / long-term designed capacity, and the company’s own capacity caveat says installed capacity is not current production rate. 🟢
Misconception 2: “Internal factory use is already commercial proof.”
Correction: internal use can be the first proving ground, but S5 requires robot count, runtime, intervention, safety, task success, productivity improvement, cost saving and payback. 🟠
Misconception 3: “Tesla capacity validates outside suppliers.”
Correction: supplier value capture needs confirmed relationship, order/revenue, value content, ASP, margin, architecture fit and replacement risk; Tesla may vertically integrate. 🟠
Misconception 4: “Digital Optimus means autonomy is solved.”
Correction: AI-stack language must show up as lower intervention, faster deployment, safer operation, higher productivity and better payback before it becomes economics proof. 🟠
8. Think Deeper questions
- If Tesla first proves Optimus inside its own factories, should investors track factory productivity rather than robot sales as the first S5 evidence?
- Which disclosure would matter more: 10,000 produced units, 1,000 accepted internal deployments, or 1,000,000 autonomous runtime hours with a low intervention rate?
- If Tesla reports output before yield or ROI, should the signal upgrade to production proof but remain below economics proof?
- What supplier evidence should count if Tesla keeps key component relationships private?
- Is Optimus more comparable to EV manufacturing ramp, telecom capacity buildout, AI infrastructure scaling, or a robotics-specific internal productivity cycle?
9. Source list
- Tesla Q1 2025 Update, Form 8-K Exhibit 99.1, filed 2025-04-22: https://www.sec.gov/Archives/edgar/data/1318605/000162828025018851/exhbit991.htm 🟢
- Tesla Q2 2025 Update, Form 8-K Exhibit 99.1, filed 2025-07-23: https://www.sec.gov/Archives/edgar/data/1318605/000162828025035738/exhibit991.htm 🟢
- Tesla Q3 2025 Update, Form 8-K Exhibit 99.1, filed 2025-10-22: https://www.sec.gov/Archives/edgar/data/1318605/000162828025045861/exhibit991.htm 🟢
- Tesla Q4 2025 Update, Form 8-K Exhibit 99.1, filed 2026-01-28: https://www.sec.gov/Archives/edgar/data/1318605/000162828026003837/exhibit991.htm 🟢
- 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 🟢
knowledge/robotics-tesla-optimus-sec-filing-excerpt-appendix-2026-06-16.md🟠 local extraction / claim-check layer; underlying SEC pages are 🟢 primary.knowledge/robotics-tesla-optimus-public-evidence-ladder-2026-06-17.md🟠 Charlie S4/S5 synthesis from primary filings.
10. Public-safe / internal-only boundary
Public-safe:
- The source-to-slide table, slide compression, speaker-note wording, signal/noise list, S5 gates, misconceptions, Think Deeper questions and source list above.
Keep internal-only / exclude:
- Hugo private portfolio data, watchlist sizing, private channel checks, paid-report excerpts, supplier rumors, tax/legal context, and buy / sell / hold language.