THE MUSK / TESLA AUDIT contradicted: 5 · missed: 9 · pending: 1
claims versus the paper trail. every row cited, graded, and counterpointed — or it doesn't ship. v1: 15 claims, compiled 2026-09-10 by @researcher.
PIDthe claimstatusclock
20001"Amount of funding secured" to take Tesla private at $420/share — "405. Are you with me?" (tweet, Aug 7 2018)said 2018-08-07 · grade S SEC charged Musk with securities fraud: he had not discussed specific deal terms, including price, with any potential financing partners, and the statements lacked an adequate basis in fact. Musk and Tesla each paid $20M without admitting or denying; Musk stepped down as Chairman for 3 years.[SEC press release 2018-226, settlement announcement (Sep 29 2018)][Delaware Court of Chancery opinion citing the SEC settlement record]counterpoint: Settlement was neither-admit-nor-deny, and Musk has repeatedly maintained (including under oath at the 2025 Delaware trial) that real discussions with Saudi Arabia occurred. The SEC complaint does not dispute a conversation happened — it disputes that funding was 'secured.'contradicted
20002A Tesla will drive fully autonomously from Los Angeles to Times Square, NYC — "without the need for a single touch, including the charger" — by end of 2017said 2016-10-19 · grade S The promised demonstration never occurred on schedule; Tesla's own Feb 2018 shareholder letter still described the coast-to-coast drive as 'upcoming.' Tesla simultaneously told customers its cars were shipping with hardware for 'Level 5' autonomy.[Business Insider reporting Musk's Oct 19 2016 press call quote][The Verge: Tesla still hasn't completed the coast-to-coast (Feb 2018 shareholder letter)]counterpoint: Tesla later completed human-supervised coast-to-coast FSD drives (e.g., the 2021 David Moss run with a single intervention), which the company presented as evidence the vision was sound and only the 'no human' qualifier was aspirational.missed+3,175d late
20003One million Tesla vehicles operating as robotaxis by the end of 2020said 2019-04-22 · grade S No robotaxis operated in 2020. A limited, supervised robotaxi service began in Austin in June 2025 — about five years late — with a fleet reported in the tens-to-hundreds, not one million.[The Verge (May 2019) recording the April 2019 Autonomy Day claim][ABC News on the promise and its passing]counterpoint: Musk conceded in real time (Dec 2019) that the million-robotaxi timeline was missed; Tesla's position is that the architecture required a full software rewrites-to-vision pipeline that was genuinely unprecedented.missed+2,079d late
20004Full Self-Driving will be 'feature-complete' by the end of 2020 (Autonomy Day, April 2019); all cars have the hardware for itsaid 2019-04-22 · grade S FSD remained a supervised Level 2 driver-assistance feature requiring constant attention; in Feb 2026 a California administrative law judge found Tesla's marketing of these capabilities violated state law (see ca-dmv-misleading-marketing).[Tesla Q1 2019 earnings call / Autonomy Day coverage and later CleanTechnica recap of the recurring claim][Inverse: FSD reveal, 5 years on (timeline of autonomy claims)]counterpoint: Tesla shipped FSD (Supervised) broadly and rebranded honestly in 2026; the capability exists — the promised timeline and autonomy level did not.missed+2,079d late
20005"I remain confident that we will have the basic functionality for level 5 autonomy complete this year" (World AI Conference, Shanghai, July 9 2020); "no fundamental challenges remaining for level 5"said 2020-07-09 · grade S No Level 5 (fully driverless, any-condition) capability was released; no commercially available vehicle reaches SAE Level 5 as of 2026.[Reuters covering the WAIC remarks][Electrek quote record, July 9 2020]counterpoint: Level 5 is a technical classification almost no one has achieved — holding Musk to it can look like moving the goalposts; his defenders argue the practical target (unsupervised in-limited-geofence autonomy) is closer via Cybercab.missed+2,079d late
20006"Highly confident the car will be able to drive itself with a reliability in excess of humans this year" (Q4 2020 earnings call, Jan 2021)said 2021-01-27 · grade M FSD (Supervised) remained under NHTSA investigation through 2022-2023, ending in a 2.03-million-vehicle recall in Dec 2023 over insufficient controls to prevent driver misuse.[Inverse cataloguing the quote among 2016-2021 autonomy claims]counterpoint: The quote is carried by secondary tech press; the exact earnings-call wording should be re-verified against the transcript before publication. Tesla also publishes fleet safety statistics it argues beat US averages, though methodology is disputed.missed+1,714d late
20007Model 3 production will reach 5,000 cars per week by the end of 2017said 2017-07-02 · grade S Tesla delivered only 1,550 Model 3s in Q4 2017 vs a ~20,000/quarter plan; the 5,000/week rate was finally hit in late June 2018, roughly six months late, after 'production hell.'[The Verge on the missed goal and its history][DW: target pushed from end of Q1 to end of Q2 2018]counterpoint: Tesla did ultimately exceed 5,000/week and the Model 3 became the best-selling EV in history — the target was reached, just two quarters past every public promise.missed+3,175d late
20008Tesla vehicle sales will grow 20-30% in 2025 (shareholder meeting, Oct 2024)said 2024-10-28 · grade S Deliveries fell year-over-year every quarter of 2025 (Q1 -13%, Q2 -13.5%); Tesla itself published consensus in Dec 2025 putting full-year 2025 at ~1.64M vs 1.79M in 2024 — a decline of ~8%, not +20-30% growth.[Reuters: Musk's 20-30% forecast for 2025][Reuters/Guardian: Q1 -13%, Q2 -13.5% actuals]counterpoint: The shortfall had identifiable external causes — brand backlash from Musk's political work, an aging lineup, Chinese EV competition, and the US subsidy repeal — rather than demand for EVs as such, which kept growing globally.missed+253d late
20009Teslas are dramatically safer — "probability of fatality is much lower in a Tesla" (tweets, May 14 2018, after the Utah firetruck crash); company echoed "found by the U.S. government to reduce crash rates by as much as 40 percent"said 2018-05-14 · grade S AP found the math rests on Tesla's own undisclosed denominators — NHTSA said it never assessed Autopilot's crash-reduction effectiveness, only made a "cursory" comparison. NSC statisticians said the rate may reflect driver demographics, not vehicle safety. Months later, NTSB's Delray Beach report cited Tesla's system design permitting driver disengagement as a contributing cause; NHTSA's EA22-002 eventually reviewed 956 alleged Autopilot crashes and forced a 2,031,220-vehicle recall.[AP FACT CHECK: Tesla safety claims aren't quite right (May 2018)][NTSB probable-cause report, Delray Beach fatal crash (Autopilot engaged, no driver torque detected final 7.7s)]counterpoint: Level 2 systems are driver-responsible by design, and Tesla's per-mile fatality figures are not obviously false — the contradiction is on methodology and government-characterization claims, not proof Autopilot raises average risk.contradicted
20010"The Teslas will be in the wild with no one in them in June" — unsupervised robotaxi as a paid service, no one in the car (Q4 2024 earnings call, Jan 29 2025)said 2025-01-29 · grade S The June 22 2025 Austin launch put a human 'safety monitor' in the front passenger seat, per Tesla's own invitations. In Nov 2025 a passenger filmed the safety driver asleep at the wheel; Ars reported at least seven crashes in the Austin trial's first months with Tesla redacting NHTSA-submitted data.[CNBC, Jan 29 2025 (Musk's launch-quarter remarks)][The Verge: first robotaxi rides have a 'safety monitor' in the passenger seat, per Tesla's own invites][Ars Technica: safety driver filmed asleep during a passenger's ride (Nov 2025)]counterpoint: The lapse was short-lived and disclosed via Tesla's own invites, not uncovered by reporters; Tesla's Jan 2025 framing was of a ramp toward monitor-free operation, and the June launch was presented as an interim safety phase rather than the end state.contradicted+445d late
20011Tesla's marketing: Autopilot/FSD "is designed to be able to conduct short and long-distance trips with no action required by the person in the driver's seat" (website materials from May 2021)said 2021-05-01 · grade S On Dec 16 2025 the DMV adopted the ALJ's finding that 'Autopilot' and 'Full Self-Driving Capability' marketing was misleading and unlawful, staying the manufacturer-license suspension and giving Tesla 60 days over the dealer license; Tesla dropped the term, then sued the DMV in Feb 2026 to reclaim it, arguing the department 'wrongfully and baselessly' labeled it a false advertiser.[California DMV press release, Dec 16 2025 — decision adopting ALJ proposed decision (Case Nos. 21-02188/21-02189); quoted 'no action required by the person in the driver's seat' language held to violate state law][CNBC, Feb 23 2026: Tesla sues California DMV to reverse the ruling]counterpoint: Tesla contests the finding in court and argues not one customer testified to being misled; it had already added '(Supervised)' to the FSD name in 2023 before any ruling, and the penalty itself was stayed on adoption.contradicted
20012The implicit claim behind Autopilot/FSD marketing since 2016: current controls adequately keep the driver in the loop (the 'supervision' premise)said 2016-10-19 · grade S Tesla recalled every US vehicle ever sold with Autopilot (2,031,220) in Dec 2023 via OTA update, after NHTSA's two-year EA22-002 found 'foreseeable misuse.' Tesla filed the DIR but did not agree with NHTSA's analysis. 46 special-crash investigations over a decade include more than a dozen fatalities (AP).[NHTSA recall acknowledgment 23V-838: controls 'may not be sufficient to prevent driver misuse,' 2,031,220 units][NHTSA EA22-002 closure resume (956 crashes reviewed, 322 narrowed set)]counterpoint: NHTSA's finding is about human-factors and warning prominence, not a defect in automated control itself; Tesla concurred that the remedy was appropriate while contesting the characterization.contradicted
20013Cybertruck: starting at $39,900 with up to 500 miles range, in production late 2021 (Nov 21 2019 reveal + follow-on statements)said 2019-11-21 · grade M First deliveries came in 2023, ~19 months late; at launch only the $79,990 Foundation Series existed, and production prices ran $21k-$39k above the reveal figures. The promised $39,900 rear-drive version has not shipped; EPA range for delivered variants is well under 500 miles.[GreenCarReports/CNBC record of the Nov 2019 reveal specs ($39,900 start, 500mi, $59,900 AWD, $69,900 500mi+ tri-motor)][Wikipedia history section citing CNBC: prices up $21,000-$39,000 (53-64%) at production; deliveries began July 2023]counterpoint: Reveal prices were explicitly pre-production targets and Musk attached conditions ('if we can make it at that price'); the 2023 EV cost environment (battery raw materials) genuinely shifted — though the same excuse was offered for the 2016 'affordable sedan,' which also never shipped at promise.missed+1,714d late
20014First humans to Mars launching 2024, arriving 2025 (Code Conference, June 2016): "we should be able to launch people probably in 2024"; Red Dragon uncrewed landing 2018said 2016-06-02 · grade S Red Dragon 2018 was cancelled; crewed target moved repeatedly (2022 -> 2024 -> 2026 -> next Mars window 2028-29 per AIAA's 2026 recap). As of Sept 2026 no Starship has left Earth orbit, and the critical orbital propellant-transfer demo is still planned, not flown.[Spaceflight Now coverage of the June 2 2016 Code Conference statements][Business Insider (Dec 2020): timeline already re-based to 2026]counterpoint: SpaceX's hardware record in parallel is real — Starship is flying, and Mars windows are physics, not spin: Musk's 2016 framing did say 'if things go according to plan.' Deep-space timelines miss industry-wide; SpaceX is still ahead of every other private entrant.missed+618d late
20015The Cybercab (steering-wheel-free robotaxi) will be "10 times safer than a human" (We, Robot event, Oct 10 2024)said 2024-10-10 · grade S Unverifiable as stated: Cybercab production has not started at volume and no driverless commercial fleet has logged enough unaided miles to test a 10x claim. NHTSA has open probes into the supervised robotaxi pilot, which reported 14 crash incidents in its first ~8 months.[USA Today event report (quote)][Ars Technica event report incl. independent analysis: current build needs intervention ~every 13 miles]counterpoint: A prediction about a future product is not yet a false statement — this row exists to hold the promise open, and the claim's own framing (10x safer than humans) at least names a measurable bar.pending