Musk Bubble Index

How much of Tesla, SpaceX, xAI and the rest of the Musk complex is earnings, and how much is story. One number from four things that can be checked: Tesla's multiple against the megacaps, the growth its price requires against the growth it delivers, 91 dated promises against what happened, and what the rest of the complex is priced at against what it sells.

Live prices · awaiting the desk's first read · index recomputed every 15 minutes · 10 feeds · method →

THE INDEX0 earnings · 100 hot air
02040608010089OF 100
Hot air

Earnings are a rounding error in the price. The valuation is the narrative.

  1. Priced as a business
    0–20
  2. Premium
    20–40
  3. Frothy
    40–60
  4. Bubble
    60–80
  5. Hot air
    80–100
TESLA FAIR VALUEvs the price today
Fair value · generous case
$47
as a megacap on next year's earnings
Market price
$403.89-0.08%
TSLA · $1.60tn market value
Air in the price
88%
of the price is not next year's earnings at a megacap multiple
$10$20$50$100$200$500$12As a carmaker13× trailing$22As a megacap, trailing24× trailing$47As a megacap, next year22× forward · FAIR VALUE$199True believer30%/yr for 5 yrs, 40× exit$404MARKET PRICE88% air

Trailing GAAP EPS $0.91 through 2026-09-30; consensus next-year EPS $2.16. Megacap median 24.1× trailing and 21.9× forward; carmaker median 13.3× trailing. Every rung is deliberately kind to the stock. How each is computed →

THE FOUR PARTSeach 0–100
Valuation gap · 30%90
Tesla's multiple vs the megacaps
Growth the price needs · 20%90
Reverse DCF vs delivered growth
Promise record · 30%88
Dated claims not delivered, by materiality
Rest of the complex · 20%88
SpaceX, Boring, Neuralink on sales
Trailing P/E444×Forward P/E188×Price / sales15.4×Growth price needs+32%/yrEPS growth, 3y−34%/yrPromises kept on time8 of 75
The readno desk read yet

The numbers above are live. The desk — a daily structured read of the Tesla, SpaceX and xAI wire against the promise ledger — files its first note shortly after deploy. Until then: Tesla trades at 444× trailing earnings, 16 times the S&P 500's multiple and 33 times a carmaker's, on revenue that fell last year, and the price requires earnings to compound at +32% a year for a decade from a base that shrank 34% a year over the last three.

Tesla trailing P/E vs the S&P 500 and the car industry — 2017 to todaythe full P/E history, by year →
Tesla now444×$0.91 trailing EPS
S&P 50028×Tesla is 16× that
Carmakers5–12×Tesla is 33× the median
Tesla peak1,384×Dec 2020
  • Tesla trailing P/E
  • S&P 500 trailing P/E
  • Carmakers, 5–12×
  • Hot air
NO TRAILING EARNINGS5×10×20×50×100×200×500×1000×20172018201920202021202220232024202520261,384× · Dec 2020444×28×

Monthly closes divided by the last four reported quarters of GAAP diluted earnings per share, split-adjusted, using only figures public by the end of each month. Hatched: no trailing earnings to divide by. The red area is the part of Tesla's multiple that is not the market's own multiple — the hot air, month by month. Q4 2023 includes a one-off $5.9bn tax benefit that flatters the twelve months after it; Q2 2026 includes a $0.22-a-share unrealised gain on Tesla's SpaceX stake.

When do the earnings meet the price?interactive
Earnings growth, every year
+25%a year, compounding
A normal multiple
25×earnings — the S&P sits near 26×, megacaps near 24×
The price stands still and the earnings catch up in
20369.0 years of +25% growth, from 2027

Tesla would be earning $16.17 a share by then, against $2.16 forecast for next year. Any share-price rise in the meantime pushes the date out.

5×10×25×50×100×200×normal: 25×20272032203720422047

For reference, Tesla's earnings per share over the last three years actually compounded at −34% a year. At that rate the answer is never.

The reckoning calendarall pending claims →

Not predictions. These are the dates on which claims the valuation rests on become checkable — put there by the companies themselves.

  1. Nov 2026Teslaon the calendar
    Cybercab Dallas launch

    Promised for 'next month' on October 21. The first expansion of the driverless Cybercab beyond Austin.

  2. Nov 2026SpaceXon the calendar
    Starship Flight 15

    Attempted catch of both booster and ship. The orbital data-centre plan assumes a cadence this flight is nowhere near.

  3. Dec 2026Teslaon the calendar
    Tesla annual meeting

    Shareholders meet amid a board review of a Tesla–SpaceX merger. A merger would need a majority of all outstanding shares.

  4. Dec 2026Teslaon the calendar
    End of 2026

    Due date for 'very widespread' robotaxis, Cybercabs to customers, Optimus for sale outside Tesla, and the Mars window SpaceX already skipped.

  5. Dec 2026Teslapromise due

    No external customer, price list or delivery date exists. On the October 21, 2026 earnings call outside sales were put 'about a year' away — which is to say, not in 2026.

  6. Dec 2026Teslapromise due

    Six metros in Texas and Florida without a safety monitor and the Bay Area with a driver as of October 2026; no new city since July. Tesla has named Las Vegas, Phoenix, Jacksonville and San Antonio as next, without dates.

  7. Dec 2026Teslapromise due

    A car with no steering wheel can be sold in the US only under a federal exemption capped at 2,500 units a year, and NHTSA opened an audit of the Cybercab's self-certification in September 2026.

  8. Jan 2027Teslaon the calendar
    Tesla Q4 and full-year 2026 results

    Whether 2026 deliveries fell for a third straight year, and whether robotaxi revenue shows up anywhere in the accounts.

  9. Mar 2027SpaceXon the calendar
    Starship uncrewed lunar landing demo

    The test SpaceX chose over Mars 2026. Artemis III depends on it.

  10. Jun 2027Teslaon the calendar
    Customer cars unsupervised

    The Q2 2027 date given for owners' Teslas to drive themselves in robotaxi zones — the claim first due at the end of 2025.

The Musk complex — what each company is valued at, and what it sells$3.60tn on $136bn of revenue · 27× sales
CompanyValued atRevenue, 12mSales multipleEarnings multipleProfit needed at 25×The pitchThe business
TSLA · live
$1.60tn$104bn15×444× / 188× fwd$63.9bnAn AI and robotics company: robotaxis, Optimus, and energy, with cars as the starting point.A carmaker whose deliveries have fallen two years running, plus a profitable energy-storage business and a robotaxi service in six metros. Earnings down 57% from their 2022 peak.
SPCX · live
$1.97tn$32.0bn62×loss-making / 62× fwd$78.9bnLaunch, Starlink, orbital AI data centres and a frontier AI lab, with a million satellites to come.The world's dominant launch provider and a 13-million-subscriber satellite broadband business, listed in June 2026 at $1.8 trillion, now carrying xAI's losses on its income statement.
↳ xAI
all-stock acquisition by SpaceX, Feb 2026
$250bn$3.2bn78×—$10.0bnA frontier AI lab that will rewrite the corpus of human knowledge.Grok, a chatbot bundled with X, and the Colossus data centres. Lost about $6.4bn on $3.2bn of revenue in 2025 before SpaceX bought it, all-stock, at $250bn in February 2026.
↳ X (Twitter)
value assigned in the xAI merger, Mar 2025
$33.0bnn/d——$1.3bnThe everything app.A social network with about 550 million monthly users including Grok, bought for $44bn in 2022, marked at $33bn when folded into xAI in March 2025, now a line inside SpaceX.
The Boring Company
$3bn Series D led by UAE investors, Sept 2026
$23.0bnn/d——$920mTunnels that end traffic, dug at a fraction of the cost of anyone else's.About fourteen stations of tunnel under Las Vegas with human-driven Teslas in them, a second project started in Nashville in 2026, and a long list of cancelled cities.
Neuralink
$650m Series E, Jun 2025
$9.0bnnone∞—$360mBrain–computer interfaces that cure paralysis and blindness and then upgrade everyone else.A clinical trial with about two dozen implanted patients who can move a cursor. No approved product and no revenue. Secondary-market marks of $30–40bn have been reported but not priced in a round.
The complex$3.60tn$136bn27×7,865×$144bnTesla, SpaceX (which now contains xAI and X), Boring and Neuralink, summed. Lines marked ↳ sit inside SpaceX and are shown, not added.

Tesla and SpaceX are priced live. SpaceX acquired xAI, which had absorbed X, in an all-stock deal in February 2026 and listed in June 2026; its trailing figures include xAI's losses. Boring and Neuralink carry their last priced rounds; reported secondary-market marks for Neuralink of $30–40bn are not used. “Profit needed at 25×” is the net income each valuation would require to trade at a normal multiple. For scale: Apple, the most profitable company on earth, earned about $110bn last year.

The promise recordfull ledger →
88%
of dated claims past their date were not delivered as described, weighted by how much each one matters to the valuation
  • Delivered8
  • Delivered late7
  • Partial11
  • Broken33
  • Abandoned11
  • False when said5
  • Not yet due16

91 claims since 2006, each with the date it was made, the date it was due, and what happened. Delivered late by 2.9 years on average when delivered at all. The rows that were kept are in there too; they are what makes the rest mean something.

Always next year — every dated promise, when it was made vs when it was duehover a dot
20062008201020122014201620182020202220242026200620082010201220142016201820202022202420262031+due the year it was saidtodayA Tesla will be able to drive itself for 90% of miles within a year. — said 2014-10-02, due 2015-12: BrokenFull autonomy is about two years away. — said 2015-12-21, due 2017-12: BrokenA Tesla will drive from Los Angeles to New York with no one touching the controls by the end of next year. — said 2016-10-19, due 2017-12-31: BrokenFull self-driving will be feature complete this year. Next year you will be able to fall asleep in the car and wake up at your destination. — said 2019-02-19, due 2020-12: BrokenNext year for sure we will have over a million robotaxis on the road. — said 2019-04-22, due 2020-12-31: BrokenI remain confident that we will have the basic functionality for level five autonomy complete this year. — said 2020-07-09, due 2020-12-31: BrokenI would be shocked if we do not achieve full self-driving safer than a human this year. — said 2022-01-26, due 2022-12-31: BrokenRobotaxi unveil on August 8. Unsupervised FSD is coming this year. — said 2024-04-23, due 2024-12-31: BrokenUnsupervised FSD as a paid service in Austin in June, then many cities in the US by the end of this year, and customers' own cars driving unsupervised in many regions by year end. — said 2025-01-29, due 2025-12-31: PartialAutonomous ride-hailing will reach probably half the population of the US by the end of the year. — said 2025-07-23, due 2025-12-31: BrokenA Tesla will be able to drive 1,000 km on a single charge by 2017, and 1,200 km by 2020. — said 2015-09-24, due 2017-12: BrokenThe Model 3 will start at $35,000 before incentives. — said 2016-03-31, due 2017-07: Partial5,000 Model 3s a week by the end of 2017, and 10,000 a week at some point in 2018. — said 2017-08-02, due 2017-12-31: Delivered lateThe Tesla Semi enters production in 2019. — said 2017-11-16, due 2019-12: Delivered lateThe new Roadster will be in production in 2020: 0–60 in 1.9 seconds, 620 miles of range, $200,000, with a $50,000 deposit today. — said 2017-11-16, due 2020-12: BrokenThe Cybertruck starts at $39,900, with production in late 2021. — said 2019-11-21, due 2021-12: BrokenA $25,000 Tesla, fully autonomous, in about three years. — said 2020-09-22, due 2023-09: AbandonedModel S Plaid+: 520-plus miles of range, deliveries late 2021. — said 2021-01-27, due 2021-12: AbandonedTesla will make 100 GWh of its own 4680 cells a year by 2022, and 3 TWh by 2030, cutting cost per kWh by 56%. — said 2020-09-22, due 2022-12: BrokenTesla will be delivering 20 million vehicles a year by 2030. — said 2021-08-09, due 2030-12: AbandonedTesla expects to achieve 50% average annual growth in vehicle deliveries over a multi-year horizon. — said 2021-01-27, due 2025-12: BrokenThe Cybercab, with no steering wheel or pedals, enters production before 2027 — probably 2026 — and costs under $30,000. — said 2024-10-10, due 2026-12-31: DeliveredA Tesla Bot prototype next year, built to do dangerous, repetitive and boring tasks. — said 2021-08-19, due 2022-12: DeliveredSeveral thousand Optimus robots working in Tesla factories in 2025, then 50,000 to 100,000 in 2026, and ten times that in 2027. — said 2025-01-29, due 2025-12-31: BrokenOptimus will be in limited production doing useful work in Tesla factories in 2025 and sold to outside customers in 2026. — said 2024-07-23, due 2026-12-31: Not yet dueDojo, Tesla's own training supercomputer, will be operational next year and will be the fastest AI training computer in the world. — said 2021-08-19, due 2022-12: AbandonedSolar Roof installations will reach 1,000 a week within the next several months. — said 2019-10-25, due 2020-06: BrokenFalcon Heavy will fly in 2013. — said 2011-04-05, due 2013-12: Delivered lateI am confident we will put a man on Mars in 10 years, maybe 12 to 15 in the worst case. — said 2011-04-22, due 2026-04: BrokenCrew Dragon will carry NASA astronauts by 2017. — said 2014-09-16, due 2017-12: Delivered lateTwo cargo ships to Mars in 2022 and four ships, two of them crewed, in 2024. — said 2017-09-29, due 2022-12: BrokenStarship will fly passengers between any two cities on Earth in under an hour — New York to Shanghai in 39 minutes, London in 29 — within a decade. — said 2017-09-29, due 2028: Not yet dueStarship reaches orbit in less than six months, and flies people within a year. — said 2019-09-28, due 2020-03: Delivered lateA thousand Starships, three launches a day in each Mars window, a million people on Mars by 2050. — said 2020-01-16, due 2050: Not yet dueStarlink speed will double to about 300 Mb/s and latency will drop to about 20 ms later this year. — said 2021-02-22, due 2021-12: BrokenThe first Starships to Mars will launch in two years, when the next transfer window opens. Crewed flights in four years if those landings go well. — said 2024-09-07, due 2026-12: BrokenStarship will land NASA astronauts on the Moon in 2024. — said 2021-04-16, due 2024-12: Delivered lateAutonomous high-density urban buses and the Tesla Network, where your car earns you money while you are at work. Both to be unveiled next year. — said 2016-07-20, due 2017-12: BrokenThe Chicago Express Loop: downtown to O'Hare in 12 minutes, at no cost to taxpayers, built in three years. — said 2018-06-14, due 2021: AbandonedThe Las Vegas Convention Center Loop will move 4,400 passengers an hour in autonomous electric vehicles. — said 2019-05-22, due 2021-01: PartialI see a path to Twitter exceeding a billion monthly users in 12 to 18 months. — said 2022-11-27, due 2024-05: BrokenA general amnesty for suspended accounts will begin next week. — said 2022-11-24, due 2022-12: DeliveredBy the end of next year, X will handle your entire financial life. You will not need a bank account. — said 2023-10-26, due 2024-12-31: BrokenNeuralink will have a product to treat serious brain injuries within about four years. — said 2017-04-20, due 2021-04: BrokenNeuralink hopes to have its first human patient before the end of next year. — said 2019-07-16, due 2020-12: Delivered lateA SpaceX-built mini-submarine will help rescue the boys trapped in the Tham Luang cave. — said 2018-07-08, due 2018-07: BrokenTesla and SpaceX will make ventilators if there is a shortage. — said 2020-03-19, due 2020-04: PartialWe can cut at least $2 trillion from the federal budget. — said 2024-10-27, due 2025-12: BrokenGiga Shanghai will be producing Model 3s by the end of this year. — said 2019-01-07, due 2019-12: DeliveredThe Model Y will be the best-selling car of any kind in the world in 2022 or 2023. — said 2021-01-27, due 2023-12: DeliveredStarlink will offer service to customers after about six launches, in 2020. — said 2019-05-15, due 2020-12: DeliveredAbout 500 robotaxis in Austin by the end of the year. — said 2025-10-22, due 2025-12-31: BrokenOptimus Generation 3 will be unveiled in the first quarter of 2026. — said 2026-01-28, due 2026-03-31: BrokenRobotaxis will be very, very widespread in the US by the end of this year. — said 2026-01-22, due 2026-12-31: Not yet dueGrok 5 will be out by the end of the year, with a real chance of being AGI. — said 2025-09-18, due 2025-12-31: BrokenCustomers' own cars will be able to join the robotaxi network, driving unsupervised, around the second quarter of next year. — said 2026-10-21, due 2027-06-30: Not yet dueCybercab launches in Dallas next month, more cities before the end of the year, and Cybercabs delivered to customers by the end of 2026. — said 2026-10-21, due 2026-12-31: Not yet dueOptimus will reach outside customers in about a year, and the Fremont line will reach 20,000 a week in at least a year, maybe a little longer. — said 2026-10-21, due 2027-12-31: Not yet dueTesla is aiming for about five million vehicles a year by around 2030. — said 2026-10-21, due 2030-12: Not yet dueSpaceX will put one gigawatt a year of AI compute into orbit by the end of 2027, then ten, then a hundred, eventually a terawatt a year. Starship will be lifting about a million tonnes a year within three years. — said 2026-06-10, due 2027-12-31: Not yet dueSpaceX annual recurring revenue will be somewhere between $500 billion and a trillion dollars by the end of 2027. — said 2026-10-20, due 2027-12-31: Not yet dueAbout 100,000 Starlink V3 satellites and a hundred times today's bandwidth within three years. — said 2026-10-20, due 2029-10: Not yet dueThe Terafab chip plant will be in volume production in the third quarter of 2028, with a first test chip by mid-2027. — said 2026-10-20, due 2028-09-30: Not yet dueNext year, I will give one of my SpaceX shares to every kid in America — the first of many. — said 2026-10-19, due 2027-12-31: Not yet due
  • Broken
  • False when said
  • Abandoned
  • Delivered late
  • Partial
  • Not yet due
  • Delivered
  • · dot size = materiality to the valuation

Dots on the dashed line were due the year they were said. Height above it is how far out the promise reached. The self-driving claims form a staircase along the diagonal: the same promise, re-dated each year since 2014.

The claims the valuation rests onall 91, with verdicts →
12 of 12
SaidDueStatusThe claimWt
Jan 2026Mar 2026BROKEN
“Optimus Generation 3 will be unveiled in the first quarter of 2026.”
Tesla · Tesla Q4 2025 shareholder letter · verdict →
●●●
Oct 2025Dec 2025BROKEN
“About 500 robotaxis in Austin by the end of the year.”
Tesla · Tesla Q3 2025 earnings call · verdict →
●●●
Jul 2025Dec 2025BROKEN
“Autonomous ride-hailing will reach probably half the population of the US by the end of the year.”
Tesla · Tesla Q2 2025 earnings call · verdict →
●●●
Jan 2025Dec 2025BROKEN
“Several thousand Optimus robots working in Tesla factories in 2025, then 50,000 to 100,000 in 2026, and ten times that in 2027.”
Tesla · Tesla Q4 2024 earnings call · verdict →
●●●
Apr 2024Dec 2024BROKEN
“Robotaxi unveil on August 8. Unsupervised FSD is coming this year.”
Tesla · Tesla Q1 2024 earnings call and Twitter · verdict →
●●●
Jan 2022Dec 2022BROKEN
“I would be shocked if we do not achieve full self-driving safer than a human this year.”
Tesla · Tesla Q4 2021 earnings call · verdict →
●●●
Jan 2021Dec 2025BROKEN
“Tesla expects to achieve 50% average annual growth in vehicle deliveries over a multi-year horizon.”
Tesla · Tesla Q4 2020 shareholder letter · verdict →
●●●
Jul 2020Dec 2020BROKEN
“I remain confident that we will have the basic functionality for level five autonomy complete this year.”
Tesla · World Artificial Intelligence Conference, Shanghai (video) · verdict →
●●●
Apr 2019Dec 2020BROKEN
“Next year for sure we will have over a million robotaxis on the road.”
Tesla · Tesla Autonomy Day · verdict →
●●●
Feb 2019Dec 2020BROKEN
“Full self-driving will be feature complete this year. Next year you will be able to fall asleep in the car and wake up at your destination.”
Tesla · ARK Invest podcast · verdict →
●●●
Oct 2016Dec 2017BROKEN
“A Tesla will drive from Los Angeles to New York with no one touching the controls by the end of next year.”
Tesla · Autopilot 2.0 hardware announcement · verdict →
●●●
Dec 2015Dec 2017BROKEN
“Full autonomy is about two years away.”
Tesla · Fortune interview · verdict →
●●●
Desk log — the index, day by daynewest first

The history starts with the desk's first reading. No readings yet.

The wire it is reading72h · 10 feeds

The wire fills after the first poll.

Questions people askanswered from the live numbers

Is Tesla overvalued?

On every conventional measure. Tesla trades at 444 times its last twelve months of GAAP earnings, against about 28 for the S&P 500, 24 for the median megacap technology company and 13 for the median carmaker. Its revenue is $104bn and falling year on year; its market value is $1.60tn. Whether that is a bubble depends on whether you believe the robotaxi and Optimus businesses will arrive at the size and on the schedule Musk describes — and the ledger below records how that schedule has gone so far.

What is Tesla's fair value?

Valued as the median megacap technology company on next year's consensus earnings — the most generous public benchmark that exists — Tesla is worth about $47 a share. As a carmaker it is worth about $12. Even a believer's case, with earnings compounding 30% a year for five years and a 40× exit multiple, discounts back to $199. The share price is $404.

When will the Tesla bubble burst?

Nobody can date a bubble, and this page does not try to. What it does is date the claims the price rests on. A bubble ends either when earnings grow into the price or when the price falls to the earnings. At 25% annual earnings growth from next year's forecast — roughly what analysts assume — Tesla's multiple would reach a normal 25× in about 2036, if the share price did not move at all in between. The nearer reality checks are on the calendar: the end-of-2026 deadlines for widespread robotaxis and Optimus sales, the full-year 2026 results in January 2027, and customer cars driving unsupervised by mid-2027.

What is Tesla's P/E ratio compared with other car companies?

Tesla's trailing P/E is 444. Toyota, General Motors, BYD and Volkswagen trade between roughly 6 and 20 times earnings; Ford and Stellantis currently have no trailing earnings to divide by. On price to sales, Tesla is at 15.4× against a car-industry range of 0.1× to 0.9×. Tesla's multiple peaked at 1,384× in Dec 2020.

How many of Elon Musk's promises have come true?

Of the 91 dated, checkable claims in the ledger, 75 have passed their date. 8 were delivered on time as described, 7 arrived late by about 2.9 years on average, 11 were partly delivered, and 49 were broken, abandoned or false when made. Weighted by how much each claim matters to the valuation, 88% were not delivered as described.

Is SpaceX overvalued?

SpaceX listed in June 2026 and is valued at about $1.97tn on $32.0bn of trailing revenue — roughly 62× sales, with a net loss over the last four quarters because it now carries xAI. Starlink is a real and fast-growing business; the valuation also prices orbital data centres and revenue of $500bn to $1tn a year by the end of 2027, which Musk promised on the October 2026 earnings call and which would be a fifteen-fold increase in fourteen months.

How this is built

The index is arithmetic, not opinion: four components, each 0–100, weighted 30/20/30/20 and recomputed from live prices every fifteen minutes. The components and their inputs are stored with every reading so any number can be recomputed from the row that produced it. The full method, with today's inputs →

The fair values are deliberately generous. The headline values Tesla as the median of the six largest technology companies on earth, on next year's consensus earnings rather than this year's reported ones. If the number still looks small, that is information.

The ledger is hand-kept and in version control; every edit to a verdict is a commit with a reason. The desk — one structured model read a day over 10 free feeds — attaches evidence to entries and logs new dated claims, all marked unverified until an editor confirms them. It never edits a verdict.

Indy Finance is an independent research publication, not an investment adviser. Nothing here is a recommendation to buy or sell any security. The index is an editorial measure built from public data and a curated record of public statements; it can be wrong, and expensive stocks can stay expensive for longer than anyone expects. Prices and multiples from free, best-effort public feeds. See our editorial policy.