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

Research library · updated 2026-06-19 · public

Tesla Optimus public evidence hygiene — what changed, what still does not count

Date: 2026-06-19 As-of: 2026-06-19 06:09 UTC Owner: Hugo / Genius Team Research owner: Finance / Charlie AGT-002 Status: SYNTHESIS_CANDIDATE Visibility: PUBLIC Output target: morning review / Tesla Optimus guardrail / future /robotics/tesla-optimus/ evidence-hygiene module

Related files:

  • knowledge/robotics-research-harness.md
  • knowledge/robotics-tesla-optimus-evidence-denominator-ladder-2026-06-18.md
  • knowledge/robotics-tesla-optimus-s5-trigger-dashboard-2026-06-18.md
  • knowledge/robotics-tesla-optimus-source-to-slide-claim-control-2026-06-17.md

Public-safety flag: PUBLIC if used as written. Excludes Hugo private portfolio data, watchlist sizing, private channel checks, supplier rumors, paid-report excerpts, tax/legal context, and TSLA buy / sell / hold language.

0. One-line answer

截至 2026-06-19 06:09 UTC，Tesla Optimus 的公开证据最适合被写成“evidence hygiene”问题：SEC filings 已经足够支持 S4 capacity-intent / manufacturing-infrastructure thesis，但仍不足以支持 S5 scaled-commercial-economics thesis。五个 Tesla 8-K Exhibit 99.1 页面在本次复核中均可访问并命中 Optimus / robotics / production-line / capacity 关键词；但 reviewed filings 仍未披露 produced units、accepted units、deployed units、runtime/intervention、factory ROI/payback、external paid customer、Optimus revenue、gross margin 或 service burden。🟢 Tesla SEC filings; 🟠 Charlie synthesis.

1. Why this artifact exists

已有 Tesla 文件已经很厚。本文件只补一个晨会可用的判断层：

Tesla Optimus 现在的问题不是“有没有证据”，而是“证据属于哪个分母”。

正确压缩方式：

  • Capacity evidence is real. 🟢
  • Capacity evidence is not production evidence. 🟢 / 🟠
  • Production evidence would still not be economics evidence unless paired with accepted units, uptime/intervention, task denominator and ROI/payback. 🟠

2. Primary-source recheck table

SEC pages were rechecked from /home/ubuntu/personal-ai/workspace/finance on 2026-06-19 06:09 UTC using Python urllib.request, HTML stripping and keyword search. Method is 🟠; source pages are 🟢.

Filing anchorHTTP status on 2026-06-19Keyword hitsSafe evidence role
Q1 2025 Form 8-K Exhibit 99.1, filed 2025-04-22200Optimus; robotics; production line; installed capacityFremont pilot production line / internal factory-use intent. 🟢
Q2 2025 Form 8-K Exhibit 99.1, filed 2025-07-23200robotics; installed capacityRobotics enters Tesla official strategy language; not Optimus output proof. 🟢
Q3 2025 Form 8-K Exhibit 99.1, filed 2025-10-22200Optimus; robotics; production line; installed capacityFirst-generation Optimus production lines being installed in anticipation of volume production. 🟢
Q4 2025 Form 8-K Exhibit 99.1, filed 2026-01-28200Optimus; robotics; production line; mass production; 1 million; installed capacityGen 3 mass-production design intent, first-line preparation, supply-chain readiness, eventual 1M/year planned capacity. 🟢
Q1 2026 Form 8-K Exhibit 99.1, filed 2026-04-22200Optimus; robotics; production line; mass production; 1 million; 10 million; Digital OptimusFremont 1M/year designed line and Texas 10M/year long-term designed line. 🟢

Important caveat: keyword hits verify primary-source presence, not production volume, utilization or economics. 🟠

3. Evidence hygiene ladder

Claim someone may want to writeCurrent evidence status as of 2026-06-19Safe wordingUnsafe wordingGrade / signal
Tesla has moved Optimus beyond demo-only narrative.Supported by repeated SEC-filed production-line / factory-preparation language from 2025-04-22 through 2026-04-22.“Optimus has SEC-backed capacity-intent and manufacturing-infrastructure evidence.”“Optimus is commercially proven.”🟢 fact; S4
Tesla has large Optimus capacity ambition.Supported by Q1 2026 filing language around Fremont 1M/year designed line and Texas 10M/year long-term designed line.“Designed / long-term designed capacity anchors.”“Tesla is producing 1M/year” or “has 10M/year current capacity.”🟢 fact; S4 with caveat
Tesla is producing Optimus at scale.Not disclosed in reviewed filings.“Produced-unit denominator is still missing.”“Volume production is happening at disclosed scale.”🔴 unsupported; would need 🟢
Tesla has useful internal deployments.Q1 2025 says wider deployment of bots doing useful work across factories was intended, but no denominator is disclosed in reviewed filings.“Internal-factory-use intent exists; deployment denominator is missing.”“Optimus has proven factory productivity.”🟢 intent; 🟠 implication; below S5
Tesla has solved robot economics.No reviewed filing discloses robot ROI/payback, revenue, gross margin or service burden.“Economics proof remains absent.”“Optimus economics are proven.”🔴 unsupported; S5 debt
Tesla validates supplier winners.Reviewed Tesla filings do not name supplier relationships or order economics for Optimus.“Supplier value capture requires separate filing/customer/order evidence.”“Tesla capacity confirms supplier winners.”🔴 unsupported by Tesla filings

4. Signal vs noise

Signal

  • Five Tesla SEC-filed quarterly updates from 2025-04-22 to 2026-04-22 are reachable as of 2026-06-19 06:09 UTC and contain Optimus / robotics / production-line / capacity anchors. 🟢
  • Q3 2025 and Q1 2026 include first-generation Optimus production-line installation language. 🟢
  • Q4 2025 includes Gen 3 mass-production-intended design, production-line preparation, supply-chain readiness and eventual planned 1M/year capacity language. 🟢
  • Q1 2026 gives the strongest capacity-intent anchors: Fremont 1M/year designed line and Texas 10M/year long-term designed line. 🟢
  • Tesla’s own capacity-table caveat says installed capacity is not current production rate. 🟢

Noise unless upgraded

  • More demo clips without robot count, task denominator, runtime, intervention, reset/failure, safety, productivity or cost-saving data. 🔴 / 🟠
  • “1M / 10M” phrasing without “designed / long-term designed capacity” qualifier. 🔴 if stated as current production.
  • Produced-unit rumor without primary source, accepted-unit count, deployment site/task, yield or economics. 🔴 / 🟠
  • Supplier claims inferred from Tesla capacity ambition without named relationship, order value, revenue recognition, teardown or official relationship. 🔴
  • AI-stack / Digital Optimus language treated as proof of manipulation economics. 🟠

5. Morning-review compression

Title: Tesla Optimus: evidence is real, but the denominator is still capacity

Three cards:

  1. What changed 🟢

    • Optimus now has a multi-quarter SEC filing trail.
    • Evidence includes pilot-line intent, production-line installation, Gen 3 mass-production design, Fremont 1M/year designed line and Texas 10M/year long-term designed line.

  2. What still has not changed 🔴 / 🟠

    • No produced / accepted / deployed unit denominator in reviewed filings.
    • No runtime / intervention / safety / task-productivity denominator.
    • No factory ROI/payback, external paid customer, robot revenue, gross margin or service burden.

  3. What would change the thesis 🟢 / 🟡

    • Produced units plus accepted units and yield.
    • Robot count by Tesla factory / task / shift.
    • Runtime hours, uptime, intervention rate and safety incidents.
    • Throughput/labor-hour/cost-saving evidence or external paid customer economics.

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. What would change our mind

Upgrade toward S5 only if Tesla or high-quality customer-side evidence discloses several of the following together:

GateEvidence requiredWhy it mattersCurrent status as of 2026-06-19
OutputProduced units by period, site and generation. 🟢Moves from capacity intent to production proof.Missing.
Acceptance / yieldAccepted units, FPY / EOL yield, rework, bottleneck stations. 🟢Separates built units from usable robots.Missing.
Deployment denominatorRobot count by factory, task, shift pattern and duration. 🟢 / 🟡Converts output into operating evidence.Missing beyond intent.
Runtime / interventionFleet runtime, uptime, intervention, reset/failure, safety incidents. 🟢 / 🟡Tests autonomy and reliability.Missing.
Internal economicsLabor-hour substitution, throughput/quality improvement, cost saving, ROI/payback months. 🟢 / 🟡Could make internal Tesla deployment the first S5 path.Missing.
External commercializationNamed customer, robot count, contract value, delivery schedule, repeat order or revenue recognition. 🟢Tests customer demand outside Tesla.Missing in reviewed filings.
Financial materialityOptimus revenue, gross margin, capex, warranty/service/support burden. 🟢Converts product evidence into financial evidence.Missing.

7. Common misconceptions

Misconception 1: “Capacity ambition is just hype.”

Correction: SEC-filed line-preparation and designed-capacity language is stronger than a viral demo and deserves S4 treatment. The caveat is that S4 is not S5. 🟢 / 🟠

Misconception 2: “1M / 10M capacity means production is already happening.”

Correction: The current public language is designed / long-term designed capacity, and Tesla’s own capacity caveat says installed capacity is not current production rate. 🟢

Misconception 3: “The first commercial proof must be external sales.”

Correction: Tesla could first prove Optimus through internal factory productivity if it discloses robot count, task denominator, runtime/intervention and ROI/payback. But those denominators are not yet in reviewed filings. 🟠

Misconception 4: “Tesla capacity automatically validates suppliers.”

Correction: Supplier value capture needs separate evidence: named relationship, order/revenue, value content, ASP, margin, architecture durability and replacement risk. Reviewed Tesla filings do not provide that. 🟠

8. Think Deeper questions

  1. Would 1,000 accepted robots with disclosed uptime/intervention matter more than a larger produced-unit number with no yield data?
  2. If Optimus first proves value inside Tesla factories, should the first financial KPI be robot revenue or productivity value per deployed robot?
  3. Which denominator should trigger supplier updates: Tesla output, accepted units, teardown evidence, or supplier filing revenue?
  4. If Tesla discloses output but not economics, should the stage become S4+ production proof rather than S5 return-cycle proof?

9. Source list

10. Public-safe / internal-only boundary

Public-safe:

  • All sections above, if kept with source grades, dated anchors and no trade language.

Keep internal-only / exclude:

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