Elon Musk

Where the scored claims come from

Sources searched

• Tesla self-driving and robotaxi predictions (curated, dated list)

  The best-documented slice of Musk's forecasting record: his repeated year-by-year promises that Tesla cars would drive themselves. A public encyclopedia list gathers 31 separate dated predictions from September 2013 through January 2026, each tied to when and where he said it — a shareholder meeting, an interview, an earnings call, an Autonomy Day, a post — and what actually happened. This is effectively a complete catalogue of his headline autonomy calls, and it is the closest thing Musk has to an enumerable archive. Independent scoreboards of the same promises corroborate it. Every row is worth extracting, and each is traced back to the original source where he made the claim rather than relying on the list's summary.

  Total known

  31

  We examined

  31

  Time span

  2013-09-01..2026-01-31

  How we sampled

  The curated list was read in full — all 31 dated autonomy and robotaxi predictions from 2013 to 2026. Because the list is already a hand-built inventory of forecast statements rather than a raw archive, no title filter was needed to find the calls; instead every entry was checked for the same forecast signals used elsewhere (a promised action such as full autonomy, coast-to-coast drive, robotaxi launch, or drivers sleeping at the wheel, plus a deadline such as 'next year', 'by 2020', 'end of this year', or a named date), and all of them qualify. The main task is not discovery but provenance: matching each summarized claim to Musk's own words in the cited earnings call, keynote, interview, or post before it can count.

• Cross-venture promise trackers and a published prediction dataset

  Several independent efforts have catalogued Musk's promises across all his ventures — one public tracker holds around 94 predictions with a dated deadline and an outcome for each on its own page; a second scoreboard focuses on Tesla and SpaceX; a research write-up published a spreadsheet of 43 of his dated predictions across roughly nine areas, each with a verdict, and reported that he was right about one time in six when he attached a deadline. These are used as a map of where his falsifiable calls live and as a cross-check on each other, not as the final word: every promising entry is traced back to the original moment Musk said it. One of the trackers blocked automated access on so far and will need to be reached another way.

  Total known

  >=94 in the largest single tracker; ~131 across the trackers and dataset before de-duplication (a floor, with heavy overlap)

  We examined

  43

  Time span

  2011-01-01..2026-06-30

  How we sampled

  The trackers and the published dataset were surveyed and the 43-entry research spreadsheet was mined most closely because it already pairs each dated call with a verdict. Entries were kept when they contained a specific action plus a deadline that has passed or is near, spanning cars, rockets, robots, brain implants, the everything-app, crypto and the economy. Overlap between the sources is heavy, so the combined count is a de-duplication-pending floor rather than a sum. Each kept call is traced to Musk's own words at the source; summaries that only paraphrase him are held aside until his exact wording is confirmed. One tracker could not be fetched automatically and is noted as a gap to revisit.

• Mars, Starship and space-milestone timelines

  Musk's Mars and Starship targets, drawn largely from his keynote presentations at the International Astronautical Congress and follow-up interviews, and gathered into a dated timeline: a 2016 plan for an uncrewed landing around 2024 with tickets falling toward $100,000 and a self-sustaining city by 2050; a 2017 keynote pointing at crewed landings by 2024; a 2022 interview shifting crewed flights to 'no earlier than 2029'; a September 2024 target of uncrewed ships in the 2026 launch window; and a February 2026 announcement deprioritizing Mars for five to seven years in favor of the Moon. The keynote transcripts and videos are the primary record; the timeline is the map to them.

  Total known

  unknown (a floor; the headline dated Mars/Starship milestones are captured, but not every restated target across a decade of talks and posts)

  We examined

  6

  Time span

  2011-01-01..2026-02-28

  How we sampled

  The headline dated space milestones were collected from a curated program history and cross-checked against news coverage of the original keynotes: uncrewed landing date, crewed landing date, ticket cost, colony/city timeline, and the successive delays. Each is anchored to the keynote presentation or interview where Musk stated it, so his own words can be quoted rather than a summary. Restatements of the same target across many talks were not each logged separately; the earliest clear dated version and the major revisions were prioritized.

• Open-search sweep of his other dated calls (robots, brain implants, AI, crypto, the economy, energy)

  Everything outside cars and rockets was reached by targeted searching rather than a single list: the Optimus humanoid robot (announced 2021 with production promised 'next year', targets that then slipped year after year); Neuralink (a December 2022 claim of a human brain implant 'in about six months', achieved roughly thirteen months later); AI-capability claims (AI smarter than any human by the end of 2026); AI-risk warnings (a 2017 post calling an AI race the most likely cause of a third world war, a 2018 'immortal dictator' warning); the X everything-app and payments promised by end of 2024; crypto calls on Bitcoin and Dogecoin; his repeated warnings that the US is 'headed for bankruptcy'; the DOGE government-savings goal that fell from $2 trillion to a small fraction; and the Solar Roof promise to out-compete ordinary roofs. These are his most-quoted calls in each area, not an exhaustive inventory.

  Total known

  unknown (open search located the most-quoted dated calls per area; not an exhaustive inventory)

  We examined

  12

  Time span

  2011-01-01..2026-06-30

  How we sampled

  For each area, searches combined Musk's name with a forecast action and a timing word — for example a robot, brain implant, AI capability, app, coin, or savings figure paired with 'by', a year, 'next year', 'soon', or a named deadline — to surface his most-cited dated claims and their outcomes. Where a reputable outlet quotes him directly (for instance the Neuralink 'about six months' line or the AI-and-world-war post) his own words are used; where only a summary exists, the call is stored with its context and held until his exact wording is traced to the original. Because there is no single archive to page through in these areas, coverage is the well-known calls rather than a complete census.

Open web search

Known gaps

This record now scores across three domains — self-driving/robotaxi (catalogued nearly in full), Mars/space milestones, and a set of dated macro-economic calls (a 2022 recession-timing forecast, the DOGE “$2 trillion” savings claim, and the X payments-by-2024 promise) — which lifted the resolved-scorable base past the 20-prediction threshold, so the score is no longer provisional on sample size (METHODOLOGY §5). It is still NOT a complete census. The recall audit puts the fraction of his total datable calls captured here at roughly a minority (~0.35): his own posting history on X, where many of his most specific calls originate, cannot be enumerated from beginning to end, so that channel is limited to posts others have surfaced and quoted. Outside self-driving, the space, robot, brain-implant, AI, energy and economic areas are captured at the level of their well-known instances, not exhaustively, and a broad tracker suggests many genuine dated calls still sit outside this record. His crypto activity is largely endorsement and hype rather than dated, falsifiable forecasts, so it contributes no scored calls. Several calls remain in non-blocking quarantine pending a word-for-word primary verbatim (e.g. the 5,000-Model-3/week and Red-Dragon-2018 claims, which survive only as corporate statements or reporter paraphrase) or pending a genuinely ambiguous resolution (the “appreciating asset” claim against the 2021-22 used-car bubble). And several famous forecasts — a self-sustaining Mars city, AI surpassing all humans, US bankruptcy — carry deadlines that have not yet arrived and cannot contribute to a resolved score.

What we didn’t score & why

Implied call — no explicit horizon stated

2

Out of scope for scoring

1

Outcome not yet known

5

Set aside for human review

5