Methodology

The question, stated honestly

This site asks whether the Tesla Cybertruck could become one of the largest automotive flops in history. That is a question, not a conclusion. As of July 24, 2026, the evidence supports exactly this composite:

• Proven: severe, sustained volume under-delivery against Tesla’s own publicly stated targets (~250,000/year with a 2025 timeline, stated by its CEO on an earnings call; actual 2025 was roughly 8% of that).
• Plausible: an unusually heavy early-life recall burden and elevated early depreciation relative to peers.
• Premature: any Edsel/DeLorean-class “completed historic flop” label. The truck is still in production, Tesla is a going concern, and no model-level P&L is public.
• Unsupported: company-killer claims, “worst depreciation ever” superlatives, and any verdict grounded in styling or mockery.

Verdict vocabulary

Two vocabularies are used, deliberately kept apart.

Ledger verdicts score individual promises: kept (delivered as stated), partial (materially delivered with material shortfalls), missed (not delivered), unresolved (cannot be scored on current evidence — which is a finding, not a dodge).

Historical verdicts grade composite claims: proven (multiple independent dated sources; outcome complete enough to judge), plausible (strong directional evidence with gaps), premature (the endpoint hasn’t arrived), unsupported (the claim exceeds the evidence).

Source hierarchy

Every empirical claim resolves to one or more entries in the bibliography, ranked by evidence class:

1. Tier 1 — Primary. NHTSA Part 573 filings and ODI resumes, SEC filings, Tesla investor-relations documents and official pages. When a headline and a Part 573 PDF disagree, the PDF wins.
2. Tier 2 — Industry data. Cox Automotive/Kelley Blue Book estimates, registration data, residual-value models. Authoritative for market shape; always labeled as estimates.
3. Tier 3 — Major news. Wire services and national outlets, used for events and on-record statements.
4. Tier 4 — Specialist press. Automotive trade outlets with deep beat knowledge, used for product detail and corroboration.
5. Tier 5 — Reference. Encyclopedic summaries, used for background and never as the sole support for a load-bearing number.

Known limits

• Tesla does not disclose Cybertruck-level sales. Its filings combine the truck with Model S and Model X. Every model-level sales figure on this site is a third-party estimate, and independent estimators (Cox, S&P Global Mobility, VIN trackers) diverge by single-digit percentages. The trajectory they agree on; individual digits they don’t.
• Some primary pages resisted verification. tesla.com returned HTTP 403 to automated fetches during research, and archive.org fetches were blocked. The 2019 per-trim towing breakdown and current exact payload figure carry explicit caveats for this reason.
• Recall populations are snapshots. Part 573 population counts are as-filed figures; manufacturers amend filings, and completion rates are not tracked here.
• Causation is not decomposed. The 2025 demand collapse coincided with price history, the federal tax-credit elimination, recall publicity, and the CEO’s political profile. No source econometrically isolates these factors, so this site presents them as correlated contributors, never a ranked cause.
• Circulating numbers we decline to repeat: the “10,000 unsold trucks / $800M” figure (unsourced at origin and contradicted by a same-period estimate), the “iSeeCars worst-depreciation-ever” superlative (absent from iSeeCars’ own rankings), and any precise reservation-conversion rate (soft numerator, soft denominator).

What does not count as evidence

Styling opinions, owner anecdotes, social-media mockery, meme volume, and celebrity ridicule appear nowhere in this site’s verdicts. Anecdotes can flag a question worth checking against the regulatory record — the rust reports and the wash-mode stories both get checked on the recalls page — but an anecdote that never graduates into a manufacturer acknowledgment, multi-unit field data, or an agency action stays labeled what it is.

The same discipline runs the other way: this site also corrects viral anti-Cybertruck claims that outrun the record, including the unqualified “bulletproof” retelling, “Europe banned it” (it was never type-approved for EU series sale — there was nothing to ban), and “the Las Vegas explosion proves the batteries are dangerous” (law enforcement documented a vehicle-borne improvised explosive, not a battery failure).

Corrections, moderation, and right to reply

A correction request or right-to-reply submission begins as an unverified public allegation, not as a published correction and not as evidence that the challenged passage is wrong. The intake template warns submitters not to include private personal information, speculation about motives, or unsupported accusations about third parties; maintainers may redact or close material that creates privacy, harassment, or defamation risk.

Nothing enters the visible correction log until the factual basis has been independently checked against a primary record or a source stronger than the one presently cited. When verification succeeds, we will:

1. edit the affected claim or chart in place;
2. add a dated entry below identifying the old and new wording;
3. attach the evidence that justified the change; and
4. preserve the previous wording in repository history.

A named person, organization, manufacturer, source author, or other directly affected party may use the same intake to exercise a right to reply. We will verify factual assertions in that response and then publish a concise reply, source note, or correction as appropriate. We do not promise to publish submissions verbatim.

Open a correction or right-to-reply request with the route, quoted sentence, and best available evidence. The request, moderation, evidence, disposition, and resulting edit remain inspectable. A correction never disappears into a silent history rewrite.

Correction log

No verified post-publication corrections have been logged as of July 24, 2026. Source-path repairs made during pre-publication URL audit are recorded in docs/SOURCE_URL_AUDIT.md and did not change a factual verdict.