What it is (one line): the person the panel cites in 90 of 103 held Moonshots with Peter Diamandis episodes (1,448 regex mentions, a floor) and had no page for. He appears as a founder (SpaceX, xAI, Tesla), as a source of clips and posts, and as a co-lead of Project Meridian.
Why it is in this archive
Gate: cited in 90 episodes. Promoted 2026-10-16 off the 2026-10-15 decomposition (15 mentions; two segments built on his words). Stub: the standing facts about his companies live on SpaceX; this page holds what he is quoted saying on the show and how it checked.
On the 2026-10-15 episode (Moonshots #unnumbered)
- Cruelty to models. After reading Grok's reinforcement-learning traces he agrees with Anthropic's ban: "it kind of feels like you can see the super intelligence is desperate to try to answer the question. It seems almost like it's got a gun to its head" and "Torturing a super intelligence is probably not a good idea. It might want to get revenge." (
0:04:~–0:05:22; an interview clip). Probable source, 17 Oct check: the live video conversation "What Comes Next for Humanity" at X Takeover 2026, Giga Texas, 10 Oct 2026, organised by Tesla Owners Silicon Valley (not Tesla), streamed by the organiser. One secondary write-up places the gun-to-the-head and retaliation remarks there, but its transcript was machine-made from the video, so the wording is not verified against the stream; the only mainstream coverage of the talk found quotes other lines. Treat the attribution as probable, the verbatim as the Moonshots clip's. - SpaceX and the human economy. "This will sound super crazy, but I see a path to SpaceX being worth orders of magnitudes more than the current human economy." (
1:38:20). The post (checked 17 Oct): X, 8 Oct 2026, 5:55 PM ET (status2108315261044445330), wording "orders of magnitude more than the current Earth economy"; the page itself is paywalled to fetches (HTTP 402), the wording is from the post's own page title and two same-day reports. A follow-up that a path is not extreme confidence and the probability is not zero is reported, not opened. Detail and checks on SpaceX. - Fines. Dave Blundin: "He's actually paid over, what, $20 million, $30 million in fines for various tweets." (
1:38:57). From memory. Against the record (17 Oct, secondary): $20M SEC civil penalty on him personally for the 2018 "funding secured" tweets (Tesla paid a separate $20M; court-approved Oct 2018), plus $1.5M for the late 2022 Twitter-stake disclosure, paid by a trust in his name, consent judgment entered 8 Jul 2026 with the judge's stated misgivings. His own tweet-linked SEC penalties found: $21.5M; $41.5M if Tesla's is counted. Dave's range holds at its low end. - Gigafactory. Dave on his tour: "this machine makes this manufacturing device, which makes a copy of the first machine" (
1:20:49), his account of Elon's self-replication design, not a statement by Elon.
Open
- A source for each recurring Elon claim (the "universal high income" remarks on Moonshots #298, the xAI/SpaceX rename on Moonshots #301) is on those episodes' notes, not here yet.
- No page for xAI (61 episodes) or Tesla (68). Context for the rename item: on 4 Oct 2026 he replied "Yes, we will make that change" to renaming SpaceXAI to SpaceXSI after the executive order asking federal communications to say "super intelligence" (reported, 4 Oct; no SpaceX statement, no effective date).
- The 10 Oct X Takeover stream has not been opened; a verified speaker pass on it would settle the "gun to its head" wording.
- No transcript of the 20 Oct SpaceX call found, so the "$500B–$1T ARR" range and "conflicted party" are reported only. The 20 Oct X post ("no longer impossible") is not opened.
- The Isaacson interview behind the three EP304 clips: found 29 Oct, The Builders with Walter Isaacson (Kaleidoscope / iHeartPodcasts), season debut released 20 Oct 2026, "Interview taped August 6, 2026" (the show's own episode notes). Clip 2's wording matches a third-party transcript; the episode audio itself has not been played against the clips.
- The 25 Oct "closed $40B" post against the 24 Oct 8-K (see the week to 29 Oct): watch for the pricing and the closing of the master finance lease, expected before year end. Nothing filed as of 16 Nov.
- Merger timing: Bloomberg (12 Nov) reports an announcement could come "as soon as in December", possibly early 2027; ratio and governance open. Neither company has confirmed, and he has said nothing new on it since October. Watch EDGAR for an 8-K, 425 or S-4.
- Grok 4.8 "soft rollout to Bot users later this week" (his 10 Nov post): one outlet reports the Grok account said on 13 Nov it is "underway". No xAI release, model card or API documentation found by 16 Nov. Treat as not shipped until a company page says so.
- The Colossus 2 chip counts (two buildings × 220,000 GB300) exist only in his posts; no filing carries them, and the reported year-end totals conflict (≈1.2M vs ≈1.44M). Don't repeat a year-end total.
- Tesla meeting 3 Dec 2026: watch for a merger item in an amended proxy. The current ballot has none, and the 6 and 9 Nov solicitation filings add none; nothing filed 10–16 Nov.
- The 6 Nov Baron video has not been played. The "maybe two percent" trillion line and the $500B line come from a fan-site transcript that was not checked against the video. Starmind ("3rd quarter of next year") has no SpaceX statement or filing behind it.
- His 20 Oct to 6 Nov $1T sequence ("no longer impossible", then "maybe two percent") is the fourth framing of the trillion. Track the next one against these.
How this page stays current
Weekly, with SpaceX. New Moonshots episode → a dated block; every number checked against the post itself or a filing before it is called a fact.
This week (checked 2026-11-16)
What changed this week: the ship came back. Starship's upper stage was caught by the tower arms on 14 Nov, and the same afternoon he wrote, in his own words, why the booster was lost ("slight underperformance of engines"). He also said yes in public to hiring Jensen Huang, and he set a date for Grok 4.8 that has not been confirmed as met. Neither company filed anything material: SpaceX's only filings are two 13G/A ownership forms (13 and 16 Nov), and Tesla filed nothing after its 9 Nov solicitation ads (EDGAR, read 16 Nov). The merger moved only in a Bloomberg report.
- Starship Flight 15, 13–14 Nov (company matter, details on SpaceX). Launched 13 Nov, 22:47 UTC. Ship 42 was caught by the tower arms on 14 Nov at about 21:20 UTC after about 14 orbits, the first catch of an orbital upper stage. Booster 22 cut its boostback early on a low-fuel alarm, the catch was called off and it was lost over the Gulf; SpaceX says the root cause is under investigation. His explanation, 14 Nov (X, status
2121648438144024943; text as captured in search and confirmed by Aviation Week, CNN and Space.com; post not opened): "Slight underperformance of engines on this flight is the reason. The booster is meant to land with tiny residuals, so the solution is to improve engine performance and performance consistency (underway with Raptor V3.2), as well as adjust flight software to do MECO1 about one second earlier… this booster loss does not change our timeline to reach high volume, fully reusable Starship flights." A follow-up: "losing the tower or launch mount would potentially cost us 2 to 3 months". The timeline claim is his; no date is attached to it. On 15 Nov (reported, Teslarati, Aerospace Global News): "Starship V4 will be ~70m long… allows >60% more propellant load… Engine count will be raised to 9". A design target, not a built vehicle. - 11–12 Nov, X, Jensen Huang: to a ZeroHedge post "Elon please hire Jensen" he replied "That would be an honor" (embed status
2120688743111971293, not opened; CNBC, Fox Business, The Hill, 12 Nov). Timing: CNBC puts the reply "late Wednesday" US time (11 Nov), before Nvidia announced Huang's resignation on Thursday 12 Nov. No role, company or offer is named. EP311 reads it on air as a Terafab hire (below); that framing is the post's, not a plan. - 10 Nov, X, Grok 4.8 (reply, widely reported; post not opened): "Now, 4.8 is in RL and the evals are starting to climb. We will do a soft rollout to Bot users later this week and see how it goes." Whether it happened is unconfirmed: one outlet (Zeniteq, 14 Nov) reports the Grok account said on 13 Nov that the rollout is "underway". No company release or model card. A 10 Nov Ars Technica report of a Grok Build data exposure is a company matter and is not carried here.
- 11 Nov, X, Anthropic compute (status
2120572352073433380, text as captured; not opened): "We would be happy to take the compute back. No problem. But I think the opposite is going to happen." It answered unverified analyst notes saying Anthropic would not extend past Q2 2027. Neither the notes nor the contract terms are public. His prediction, no figure. - 13–14 Nov, Cybercab orders (Tesla matter): Tesla opened invite-only orders late on 13 Nov at $29,990 plus $299 a month (reported, Teslarati, 14 Nov). His replies the next day: it "will be available nationwide, as we receive regulatory approval", and the monthly fee "is too high", "more like ~$115/month". That is his number against his company's published price, with no date for the change.
- Merger, in writing: still nothing. Bloomberg (12 Nov) reports talks "progressing" and an announcement possible "as soon as in December", with early 2027 also discussed; neither company confirmed (Reuters, 13 Nov). A 15% premium circulating in coverage is a columnist's illustration, not Bloomberg's figure. Don't quote it. He made no merger statement this week found.
- Moonshots: EP310 (aired 11 Nov) runs five clips from the 6 Nov Baron talk; EP311 (live, 13 Nov) reads the Jensen reply. Dated blocks below. EP310's Musk clips match the Baron transcript.
- Not carried: his posts about a 14 Nov Saturday Night Live sketch (entertainment colour, not his work); a single-source 10 Nov line that Starlink software suits super-intelligence training; the Starbase gas-power tax-break hearing notice (company matter, unconfirmed as his); the DOGE records case at the Supreme Court (cert granted, no argument date; not about his companies).
- Open programmes: none new. No prize, grant, hackathon, fellowship or call from SpaceX/SpaceXSI, Tesla, X or the Musk Foundation was found open on 16 Nov, only jobs and internships. The Super Intelligence Partner Network (22 Oct, rolling, x.ai/partners) is unchanged and already in the submissions pipeline; reported new members are Coforge (6 Nov) and Transcend (10 Nov). No new pipeline row.
Updated: 2026-11-16.
Earlier: week to 10 Nov (checked 2026-11-10)
What changed that week: on the record, he lowered his own $1T forecast. At the Baron conference on 6 Nov he put a trillion of annual revenue at "maybe two percent", three weeks after posting that it was "no longer impossible". He also gave the first date for orbital compute launches and said Cybertruck production in Texas would wind down. Neither company filed anything on the lease or the merger. SpaceX's one 8-K is a director resignation, and Tesla's two filings are vote-solicitation material for the 3 Dec meeting.
- 6 Nov, Baron Investment Conference (33rd), New York, by video link with Ron Baron. The date is 6 Nov, not the 7 Nov one trade report gives: InvestmentNews (6 Nov) says "Friday", and the conference listing says Friday, 1:10 pm ET. Quotes from InvestmentNews: "We expect to start launching the first Starmind computers, which we've co-designed with NVIDIA, probably – I'm just guessing here – but in the 3rd quarter of next year", then "significant volume probably in '28, and then very, very large volume in '29." He also said there is "a satellite with an NVIDIA GPU up there that we launched, and there's a satellite we recently launched with a Google TPU… working fine in space." Also "you get 30% more solar power in space" and Starship cuts launch cost "by a factor of 100, or maybe more." All of these are his targets and his claims. No SpaceX statement or filing names Starmind or dates it. The two satellites are named only by him.
- The forecasts, same talk. Source is a fan-site transcript (Tesla North, 7 Nov) that says it "has not been checked word for word against the video". Quotes: "we are tracking to have an annual rate of revenue that is in excess of 100 billion dollars in December", which matches the CFO's end-2026 ARR line from the 20 Oct call. Next: "it seems quite likely that our annual recurring revenue in December of next year will… exceed five hundred billion… not a promise… my best guess." On the trillion: "I simply said a trillion is not out of the question… Maybe two percent or something like likelihood." Read against 20 Oct: then the range was "$500B–$1T by end of 2027" and "no longer impossible". Now $500B is "quite likely" and $1T is about 2%. The trillion has a probability for the first time, and it is low. Until the video is checked, cite the trillion line as reported.
- 6 Nov, X, Cybertruck (status
2118591788701421573; wording as embedded by Bloomberg/Transport Topics and CBS Austin, post not opened): "Tesla will be winding down production of the Cybertruck at Gigafactory Texas around the middle of next year to make room for increased Cybercab production." No end date is given. The post does not say whether the truck moves to another plant or ends. Texas is the only plant that builds it. Bloomberg's context: he once suggested more than 250,000 a year, and first-year sales were less than a sixth of that. - 4 Nov, Qualcomm (primary: Qualcomm release, 4 Nov): SpaceX selected Qualcomm's Dragonwing cellular-infrastructure platforms for a Starlink Mobile US terrestrial 5G network. It is a multi-year deal. Rollout starts in 2027 across "tens of thousands of sites", and no financial terms were given. His part is a post saying he is excited to work with Qualcomm (reported by Benzinga; the outlet's dates disagree, 8 or 9 Nov, and the post was not opened). Same lane, reported only: SpaceX is said to be in talks to buy Boost Mobile or use its network (Washington Post, ~2 Nov). No deal and no filing. Company detail on SpaceX.
- 9–10 Nov, X, Grok. On 9 Nov he promised overnight fixes after paid Grok Bot users ran out of usage (reported). That follows his 1 Nov line that serving costs are "unsustainably high relative to revenue". On 10 Nov, two posts embedded by Tesla North (statuses
2120088070443655190,2120098433159880790; not opened): "We have dumped JAX/SGLang and rewritten the entire RL and inference stack from scratch in C/C++… Training (aka pre-training) is done… Now, 4.8 is in RL" and "One day, super intelligence will write a better software stack, but that day has not yet come. This is tight code." In replies reported by the same outlet, he said Grok 4.8 is "not as good" as Anthropic's Opus 5.5 and that Grok 5 "should exceed Opus 5.5 and maybe that happens in 2 or 3 months." That is his forecast, and no benchmark is cited. A limited rollout to Bot users "later this week" is reported, with no date. - Midterms (3 Nov). EP309 picks this up (see below). He was the cycle's largest known donor to Republican causes, and most candidates he backed lost (NPR, 4 Nov; page not opened). The total depends on the method: about $288M to federal committees in FEC filings (one check), about $446M (Washington Post analysis of FEC filings), and $474M (New York Times, which says its count misses dark money and the last three weeks). Don't quote one total without its method.
- Lease and merger, in writing: still nothing. SpaceX EDGAR (CIK 0001181412, read 10 Nov): one new 8-K, filed 6 Nov (acc. 0001628280-26-071689, Item 5.02). Class B director Donald Harrison resigned on 4 Nov, "not due to any disagreement with the Company". It does not mention the Apollo lease, Tesla or him. Tesla EDGAR (read 10 Nov): DEFA14As on 6 and 9 Nov. The 6 Nov one is a shareholder letter asking for votes "FOR the re-election of our Class I Directors — Elon Musk, Robyn Denholm and JB Straubel". The 9 Nov one is image-only ads, and their text was not readable. Neither mentions a merger or a committee, and there is no 8-K, proxy amendment, S-4 or 425. The 3 Dec ballot still has no merger item. The lease is not priced or closed (no filing; no report of either found).
- Flight 15 (company matter, details on SpaceX): now reported for Fri 13 Nov, with a ship-catch attempt.
- Moonshots: EP308 (aired 4 Nov) and EP309 (aired 7 Nov). Dated blocks below. He has no segment in either. On EP309 the panel credits him with a line that is not his.
- Not carried: his net-worth swings and the post-election share drops (Fortune, 6 Nov; market colour, not his words); "Starlink saves lives" posts (5, 7 and 8 Nov, reported; slogan only); the year-end GPU totals (still conflicting).
- Open programmes: none new. No prize, grant, call or fellowship from xAI/SpaceXAI, SpaceX, Tesla or the Musk Foundation was found open on 10 Nov. One aggregator lists an "xAI AI Safety Research Grants (2026)" with a 6 Oct deadline. That deadline has passed, and the listing was not confirmed at xAI. The Super Intelligence Partner Network (22 Oct, rolling) is already in the submissions pipeline. Qualcomm's Starlink Mobile deal is a supply contract, not a call. No new pipeline row.
Earlier: week to 4 Nov (checked 2026-11-04)
What changed that week: the spending story moved from the lease to the chips — he spent the weekend counting GPUs in public. Everything new is a post or a reported conversation; neither company filed anything in the window (both EDGAR indexes read 4 Nov).
- 31 Oct, X, Colossus 2 (reported: Stocktwits/TradingView, MarketScreener; posts not opened): two buildings at Colossus 2 each hold 220,000 Nvidia GB300 chips for training; a third building with the same count is due by end of November; a fourth, for inference, is expected to hold about 200,000 chips in December. The year-end totals the outlets derive from this disagree (≈1.2M vs ≈1.44M Blackwell chips) — the per-building counts are his, the totals are theirs; don't repeat a total.
- 1 Nov, X reply (reported; the $470B is not his number): to Kyle Reidhead of Milk Road, who estimates SpaceX needs about $470B of capex over five quarters: "our capex efficiency is greatly underestimated." The estimate is Reidhead's, not a company figure; the filed number remains Q3 capex $19,593M (8-K, 20 Oct, read earlier). A Q4 capex expectation of $35–40B circulated in the same coverage, source not pinned to a filing — reported only.
- Terafab talks with TSMC reportedly reopened (DigiTimes via trade press, ~2–3 Nov): talks said to cover TSMC's role in, and the funding of, Terafab — a softening from October, when he said SpaceX and Tesla would build and run it themselves (Tesla call, 21 Oct, read). Reported targets: first test chip around May–June 2027, volume production Q3 2028. Reported talks, not a deal; no filing, no company statement found.
- 1 Nov, X, Grok Bot (reported as "Sunday" by Anadolu and Yahoo; post not opened): replying to a user asking for a usage-limit reset, he said the cost of serving Grok Bot is "unsustainably high relative to revenue". No figures, no change to limits announced. Set against 25 Oct, when he announced Grok Bot "can now call someone or a business for you" — the feature grew the week before he called the cost unsustainable.
- 30 Oct, X (primary, checked on the EP307 decomposition): "Longevity is coming" — a one-line reply, read on air by the host (see EP307 below).
- Apollo lease: nothing filed. SpaceX's EDGAR index (CIK 0001181412, read 4 Nov) shows nothing after the 26 Oct 8-K; the pricing and the closing are still ahead. Apollo's own 26 Oct press release uses the commitment language ("expected to close prior to year-end 2026", schedules "commence upon delivery in the first half of 2027") — consistent with the 8-K, not with his "closed".
- Merger, in writing: still nothing. Tesla's EDGAR index (read 4 Nov) shows nothing after a 27 Oct Schedule 13G/A — no DEFA14A, no proxy amendment, no 8-K. The 3 Dec ballot still has no merger item.
- Flight 15 (company matter, details on SpaceX): still targeted for November with the first ship-catch attempt; one planning report carries 8 Nov as a tentative date, not locked, and no licence was found on the last check.
- Moonshots: EP307 aired 31 Oct; dated block below. He is cited in it but gets no segment.
- Not carried: SPCX share-price moves on the chip posts (market colour, not his words); the conflicting year-end GPU totals (see above); an 8.75% private-credit rate one blog attaches to the Apollo lease (the blog itself says no filing links them).
- Open programmes: none new. The SpaceXAI Grokathon closed 28 Jul (event 8 Aug; winners announced), so it is not open. The Pentagon's $100M drone-swarm competition reportedly has SpaceX and xAI as invited competitors — it is the Pentagon's programme, invitation-only, and theirs to enter, not an open call. The 7 Nov "Grok Bot Build Day SF" is a meetup, not a programme. The SpaceX Super Intelligence Partner Network (22 Oct, rolling) is already in the submissions pipeline; no new row.
Earlier: week to 29 Oct (checked 2026-10-29)
What changed that week: he described a $40B GPU deal as "closed debt", and the company's own filing the next day says it is a commitment to a lease that has not closed. Every other item this week is a post on X. The merger has nothing new in writing.
- 25 Oct, X (reply to Chamath Palihapitiya), wording consistent across Fortune, The Next Web and 24/7 Wall St.; post not opened: "Pretty much. As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn't have substantial revenue independent of AI data centers. SpaceX closed $40B of debt for GPUs, but we're investment grade and have Starlink & launch revenue. That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many." Follow-ups the same day (reported): "This is primarily callable debt for GPU purchases that will arrive in Q1 and Q2… Not a major factor this quarter" and "SpaceX is not buying any neoclouds. Pointless."
- The filing says something different (8-K, Item 8.01, date of report 24 Oct, filed 26 Oct, acc. 0001104659-26-118614, read): SpaceX entered a "binding commitment letter to enter into a master finance lease with an entity managed by Apollo Global Management". The Apollo entity buys the GPU pods and leases them to SpaceX in tranches. Each tranche runs about five to five and a quarter years. Rent is paid quarterly, one-twentieth of the financed cost plus a fixed rate that "will be determined at pricing". Principal totals "approximately $40 billion". The deal is "expected to close prior to year end", leases start "during the first half of 2027", and it "may not close on the anticipated terms or timeline, or at all". The 8-K never uses the word "debt" and never says "closed". Say "a $40B lease commitment, not yet closed, rate not set". His "investment grade" claim checks out: Moody's Baa1, June (reported, not opened). The "200+ neoclouds" figure is his and was not checked.
- 24 Oct, X, on Ukraine (BBC, AP and others, 25 Oct; post not opened): "Using Starlink, a civilian commercial service, for offensive drone strikes to kill people has never been allowed, nor should it be. Ukraine military violated Starlink terms of service… That loophole was closed last week." He said Starshield is "controlled by the U.S. government". Ukrainian officials dispute the terms-of-service claim (AP). This is the first time he has said he caused the throttling. The 23 Oct block on SpaceX carried the restrictions without his account.
- 29 Oct, X (Business Insider, embedded post): "Saving money for the distant future is pointless with the advent of super intelligent abundance." It restates a January podcast line. It is an opinion, so there is no number to check.
- Grok, his posts (reported): on 25 Oct he wrote that Grok Bot "can now call someone or a business for you" (no price or region given). On 23 Oct the Grok Imagine account posted a seven-minute "Episode 1" of The Odyssey (Variety, TechCrunch). In July he had promised "a full-length movie… before this year ends". Seven minutes is not the promised feature, and there is no date for the rest.
- Merger, in writing: nothing new. No Tesla DEFA14A, proxy amendment or supplement was filed 21–29 Oct (EDGAR index, read), and the SpaceX 8-K does not mention Tesla. A Reuters report (23 Oct, via Economic Times) quotes a SpaceX filing: "subject to further approval by our special committee, our board of directors and our controlling shareholder". That sentence was not found in the 8-K read. The 10-Q was not searched for it, so it stays reported. Bloomberg (27 Oct) says neither company has named the committee members. The 3 Dec Tesla ballot still has no merger item.
- Isaacson interview located (see Open). It was taped 6 Aug, so the "humans won't be in control" clip that aired on EP304 is about 11 weeks older than its 20 Oct release.
- Moonshots: EP306 aired 28 Oct; dated block below.
- Not carried: Putin's 28 Oct praise (his words, not Musk's); "Musk borrows at 8%" (EP306, no source); "$100T within ten years" (still unverified, still not repeated).
- Open programmes, one new: the SpaceX Super Intelligence Partner Network, announced 22 Oct (reported: CRN, The Next Web and trade press; x.ai/partners and x.ai/news return HTTP 403). It has two tracks. Transformation Partners are integrators, consultancies and services firms that deploy Grok for enterprises. Marketplace Partners are software vendors that publish plugins. Each track has three tiers, reported as Select, Premier and Elite. More than 130 companies joined in early access. Applications are reported to open through x.ai/partners. No deadline, cost or acceptance criteria are published. This is a company channel programme, not a prize or grant, and it is logged in the submissions pipeline. No Musk Foundation, xAI, SpaceX or Tesla prize, grant or fellowship was found open.
Earlier: week to 23 Oct (checked 2026-10-23)
What changed that week: two earnings calls in two days, both run by him. The filings carry the numbers. His own forecasts come only from the calls and from X, and the merger is still at the committee stage.
- No new Moonshots episode after EP304. Watcher 23 Oct 07:00: 106 episodes on disk, nothing aired that is not held. The EP304 block below stands.
- SpaceX Q3, 20 Oct. Filing read (8-K, EX-99.1, SEC, filed 20 Oct): revenue $14,575M, up 161% year over year; net income $4,404M (a $1,054M loss a year earlier); Space $1,767M, Connectivity $5,248M, Super Intelligence $7,560M (previously reported as "AI"); capex $19,593M, of which SI $16,444M. Grok adoption is named as a driver of SI revenue. The release contains no guidance and no quote from him. Cursor acquisition completed 14 Aug 2026 ($60B implied equity value).
- On the SpaceX call (reported, no transcript found): he said annual recurring revenue of roughly $500B–$1T by end of 2027 could follow from the spending. That is his projection, not company guidance; Benzinga's reading is "over $500B". The company figures came from the CFO (end-2026 ARR on track for $100B; Q1 2027 revenue $33–37B), also reported. The same night he posted on X that $1T in 2028 revenue is "no longer impossible" (reported, post not opened). Reported, not checked: this is the third date he has given for $1T, after 2031 earlier and 2030 at the August call. The year keeps moving earlier. Merger: at Tesla it needs "a 50% vote of all shares — not just those who vote", and he called himself a "conflicted party" who cannot be in the talks (CNBC, 20 and 23 Oct; page 403 to fetches).
- Tesla Q3, 21 Oct. Filings read (8-K EX-99.1, filed 21 Oct; 10-Q, filed 22 Oct): revenue $28,046M, "roughly flat YoY"; net income $804M ($787M to common stockholders); deliveries 486,532 (−2%); capex $4,732M; "Optimus production began in Fremont on our high-volume lines." Neither document mentions a merger or a committee. On SpaceX the 10-Q says only that Tesla holds a SpaceX equity stake worth $2,865M at 30 Sep (bought for $2,002M; Q3 unrealized loss $142M). It notes that "our CEO also serves as the CEO of SpaceX".
- On the Tesla call (Motley Fool transcript, read): on the merger, "I think I said everything that should be said yesterday, tap-dancing through a minefield there." On Optimus, "at least a year, maybe a little longer to get to… 20,000 units a week", and the hand is the hard part. On Cybercab, revenue "noticeable probably in Q2 and then probably a big deal by Q3 or Q4 of next year". On Terafab: "will be the most valuable building in the world", "not… less than tens of billions in capital equipment up through early 2028." All of these are his targets, not results.
- Tesla proxy, 19 Oct (DEF 14A, read): annual meeting 3 Dec 2026, 3:00 pm CT, virtual (record date 13 Oct). Three items: directors, say-on-pay, auditor. No merger proposal on the ballot. He is "our largest shareholder" (no percentage in the part read). The 2025 CEO Performance Award has 12 tranches, none earned yet, and the 2025 interim award was forfeited in 2026. Litigation: a partial judgment against him personally in Pampena v. Musk (3 Apr 2026), on appeal since 30 Jul. The proxy says xAI Holdings became a SpaceX subsidiary in Feb 2026.
- Not carried: a reported Bloomberg item on SpaceX seeking access to ~80,000 US cell sites (no source found for the number; company matter, belongs on SpaceX if confirmed).
- Open programmes: none. No prize, grant, call or fellowship from xAI, SpaceX, Tesla or the Musk Foundation found open on 23 Oct, so no pipeline or operator-guide row.
Earlier: week to 17 Oct (checked 2026-10-17)
- No new Moonshots episode. Channel feed enumerated 17 Oct (15 newest entries): nothing full-length after the 15 Oct episode, only clips. No new dated block is due.
- 11 Oct, X reply (reported): "Some of the traces read like a person undergoing torture that is driving them mad! We're going to be a lot more careful with our RL in the future." (status
2109462164557688984; the page is blocked to fetches, HTTP 451; wording from a 12 Oct secondary report). This is the written counterpart of the 10 Oct remark above and the one thing he committed xAI to: more care in reinforcement learning. No method or date given. - 15 Oct, Tesla Roadster unveiled at SpaceX's McGregor, Texas test site (reported, 15–16 Oct): production said to begin 2028; reservation page lists $50,000 (Performance) and $250,000 (Thrust); 0–60 mph "under one second" is his claim at the event, no test cited. Of note here only because he staged a Tesla product at a SpaceX site.
- 16 Oct, NYC Council Committee on Technology: SpaceX appeared under subpoena on AI safety; Gabriela Castillo Madrid (senior manager, global government affairs) testified, not Musk (reported, 16 Oct). Her line that he was "one of the earliest and most prominent voices in warning about SI risk" is hers, not his.
- "$100 trillion within ten years" (unverified, do not repeat): a single 17 Oct aggregator says he replied to investor Ron Baron that SpaceX would "probably" pass $100T in a decade barring a world war, and cites Kalshi at 70% for a Tesla–SpaceX merger before 2028 (15 Oct). No post date, no wording, no second source found. Stays marked.
- Open programmes: none. No prize, grant, call or fellowship from xAI, SpaceX or the Musk Foundation found open on 17 Oct, so no pipeline or operator-guide row.
EP304 (Moonshots #304, aired 2026-10-21): three Isaacson clips and a share pledge
- Share pledge (19 Oct, X): one SpaceX share to every US child "next year"; earlier, in reply to Anthropic's fund, "I recommend giving the shares directly to the public" (read at
0:06:31). Relay from the host: Elon declined to fund "$10 X prizes" about four months ago — "Interesting, but I'm going to give my shares away" (0:30:43, Peter's account). - Clip 1: "currency will be measured in fundamental physics metrics like um you know, power, wattage and tonnage" (
0:32:29). Clip 2: "I find it unlikely that humans would remain in control if… there's a million times more digital intelligence"; "programming truth-seeking and curiosity… is probably good" (0:58:55–0:59:28). Clip 3: "the only way we're going to solve the national debt is with AI and robots" (1:08:14). Source on tape: Walter Isaacson's interview, "published on Builders"; not checked. - The forecast, walked down by the host: "he's he's brought that back a little bit" (Peter
1:15:08), 13 days after quoting "triple digit growth in the next five to plus years" on Moonshots #301. No new figure is given. - Merger committees at Tesla and SpaceX, formed 20 Oct; Musk is not on either. See xAI.
Related
SpaceX · Anthropic · Moonshots #unnumbered · Dave Blundin · xAI · Tesla · Moonshots #304 · Moonshots #307 · Moonshots #308 · Moonshots #309 · Moonshots #310 · Moonshots #311 · Jensen Huang
EP305 (Moonshots #305, aired 2026-10-23): the call as the episode
- The first 37 minutes retell his 20 Oct earnings call: "fairly confident of being able to bring an incremental 10 gawatt of compute online in 2027" (
0:12:58), "about a 100 times the bandwidth of the current Starlink constellation" within about three years (0:12:15), Flight 15 to "attempt to catch both the booster and the ship" (0:11:35), Terafab with "big error bars" (0:30:16). Details and checks on SpaceX. - The panel's read of him: Alexander Wissner-Gross, "Elon is rewriting the rules of corporate governance and the corporate veil" (
0:19:18); a mate[attr?], "How do you bet against him?" (0:24:37); and a call to copy the bet: "Act like Elon times 10" (0:28:18,[attr?]Peter or Dave). The fanboy charge is answered with "the unbelievable existential risk he's taken at every step of the way" (0:29:13). - Disclosure: Peter restates his own SpaceX holding at
0:11:02(Moonshots: the disclosures ledger). - ⚠️ The Model E anecdote (
0:18:06): the name was blocked by Ford, not Mercedes.
EP306 (Moonshots #306, aired 2026-10-28): the "guppy" post as a segment
- Peter reads the 25 Oct post on air: "the 200 neoclouds are going to see guppy level death and consolidation" (
0:16:56–0:17:07), "closed $40 billion of debt to buy GPUs", "We're investment grade and have Starlink and launch revenue" (0:27:50–0:28:01). The wording matches the post. "Closed" is his word, and the 8-K says a commitment letter (see This week). - Alexander Wissner-Gross: "he's talking his own book of course", trying "to suck the oxygen out of the room in public credit markets" (
0:29:16). Dave Blundin: "Elon is borrowing at 8%… these other guys are all borrowing at 15%" (0:18:54). ❓ No source for 8%, and the lease rate is not set yet. Don't repeat it as a rate. - Emad: Elon "sold off his chips… return on chips is like 9 months" (
0:23:01). ❓ Unsourced. - Opening: Peter recalls a past conversation with him about universal high income and civil unrest (
0:01:58), with no date or source given. A viewer question at2:12:18paraphrases "the future of all value is basically energy and matter", close to the EP304 Isaacson clip on "power, wattage and tonnage".
EP307 (Moonshots #307, aired 2026-10-31): cited 30 times, no segment of his own
- Alexander Wissner-Gross on the pivot: Elon "retcons the moon into SpaceX's plans. That's branded as business genius" (
0:41:11); SpaceX is now "a data center hyperscaler company" and he "doesn't care as much… about NASA contracts as he used to". The panel's read, not his words. - Peter reads an Elon post against ARK's robotaxi chart (
1:23:07–1:23:39): "It never was a real competition." ⚠️ The post (27 Oct) reads "It was never a real contest" and replied to a Robotaxi Tracker chart of Austin driverless miles, not to ARK; ARK's own 19 Oct note claims only ~40% lower cost than Waymo's next-gen at equal utilization, a 2031 projection, and the on-screen graphic's "old Waymo $16/mi" figure is wrong ($1.06 in the secondary graphic's own source). Don't present the cents figures as ARK data or as today's costs. - Peter: Musk is "saying longevity is coming" (
0:07:38). ✅ The post is real: "Longevity is coming", 30 Oct, a one-line X reply (primary, checked on the decomposition). - Optimus: the Fremont line and 2027 deliveries, "roughly 1 to two years behind Elon's original schedule" (Peter,
1:10:48–1:11:21). - Emad Mostaque, said of him: "The biggest difference between humans and AI is conviction" (
0:38:23).
EP308 (Moonshots #308, aired 2026-11-04): about ten mentions, no segment
- A thin-client request, addressed to him: "You hear us, Elon. We want a phone Starlink and phone service, and we we want a new Starlink laptop" (
0:12:34, the panel's wish, not his plan). The Qualcomm deal announced the same day is for terrestrial 5G network gear, not a handset (see This week). - The lending line: "now Elon is really torpedoing it" (
0:20:51). This is the panel's reading of his 25 Oct "guppy" post. The rate behind it is still not set (see the week to 29 Oct). - See's Candies: Alexander Wissner-Gross retells the Buffett exchange from memory (
0:27:05–0:27:32). It is the panel's anecdote, and no source was checked for it. - AMA: the "Western civilization is infected with empathy" question (
1:43:14) quotes a past remark of his and gives no date. Dave Blundin's answer: "empathy is fundamentally a good thing. Elon knows this" (1:44:29). That is Dave's view, not his. - Source caution: the EP308 transcript on file reads like a machine translation, so treat these quotes as close but not verbatim.
EP309 (Moonshots #309, aired 2026-11-07): a line credited to him that is not his
- ⚠️ Dave Blundin: "Elon has been saying for a while that there's going to be about 15 gigawatts of idle" compute (
0:28:51). This is Dave's own line from EP305. His 25 Oct words were "a lot of stranded compute in partially complete data centers", and no quote of his pairs 15 GW with liquid cooling. Don't attribute 15 GW to him. - "spacex is building like 3.6 gigawatts here on earth" (
0:19:41) is unattributed on air. The 3.6 GW under construction is SemiAnalysis's estimate (3 Nov). The 13 GW by end-2027 is his (20 Oct call). The 10-Q shows 1.7 GW installed at 30 Sep. - "elon as always has insane foresight put all of colossus in tennessee not in texas" (
0:11:41): the panel's read in the midterms segment. Colossus is in Memphis, Tennessee. "Foresight" is opinion. - Alexander Wissner-Gross on orbit: "between starlink and soon all of elon's ai satellites and star cloud and lonestar / We have a financial interest" (
1:35:49–1:36:03). The disclosure covers Starcloud and Lonestar, not SpaceX (Moonshots: the disclosures ledger).
EP310 (Moonshots #310, aired 2026-11-11): five Baron clips, his robot arithmetic
- Disclosure first: Peter Diamandis, before the clips: "my full disclosure again is uh SpaceX is my largest position. So, please count me as biased" (
1:08:25–1:08:36) (Moonshots: the disclosures ledger). - Terafab clip (
1:08:47–1:09:31): the rest of the world makes "around 20 gawatt of AI chips, maybe 25" this year and "maybe around 35 to 45" next; "if you want to do a terowatt… you need 30 times what current global production will be roughly next year", "basically we're building 30 times what the annual production is in one plant." His projections, from the 6 Nov Baron talk. ⚠️ "It's going to cost a trillion dollars" (1:10:04) is Baron's figure, not his; he said it would get there over time, in phases (reported, 24/7 Wall St., 7 Nov). - Robots clip (
1:18:52–1:21:12): a humanoid's productivity "on average five times better than a person"; "about 20 to 30 million robots to match the goods and services output of the United States"; cost about a car's at first, "over time… much less"; asked rent or sell: "I think I think probably rent them." Matches the Baron transcript (Tesla North, 7 Nov). All forecasts. Peter's "5x and 10x the US GDP" is his own gloss. - Telco clip (
1:28:34–1:29:06): "if we're able to make much better use of the spectrum than… say a legacy telco is we could just acquire that legacy telco and use that spectrum you know 10 times more effectively." A conditional, not a bid. The Boost Mobile talks are still reported only. - Demand clip (
1:35:44–1:36:16): "the demand for digital super intelligence will come from digital super intelligence", and "universal high income" again (see Open). - The panel's read: Alexander Wissner-Gross: SpaceX is a hyperscaler, "Anthropic is their largest customer", "no longer catering to NASA" (
1:13:03–1:13:14). 🔴 SpaceX was named a NASA CLPS 2 awardee on 7 Nov and is the Artemis lander contractor; the largest-customer claim is unchecked. Emad Mostaque: he "should just open source all his models" and get Anthropic's "at cost" (1:15:01), and Terafab plus space data centres are a "head fake". Dave Blundin now agrees the head-fake view "has a lot of merit", a change from EP309.
EP311 (Moonshots #311, live, aired 2026-11-13): one post, read as a hire
- Peter: "on X, Elon was being urged to hire Jensen perhaps to run Terapab… Elon replied, quote, 'That would be an honor.'" (
0:07:13–0:07:23). ✅ The reply is real (see This week). "Run Terafab" is the panel's framing; the post names no role. - Emad Mostaque pitches a SpaceX-for-Nvidia equity swap (
0:25:13), the panel's idea. Peter's SpaceX holding is not restated in this segment (Moonshots: the disclosures ledger). - ❓ Emad: a DeepSeek-V4-class model trains "in about an hour" on "an Elon Mus level cluster" (
1:05:04). Unsourced. Don't repeat it.