Lollapalooza of X
A thought experiment: fold Tesla into SpaceX, name the result X, hand the keys to Gwynne Shotwell, and point it at the most expensive line in Amazon's accounts, the last mile, then at the slowest one, the ocean in between. When four or five forces push the same way, one plus one can make eleven.
In 1995, Charlie Munger stood up at Harvard and gave a talk called The Psychology of Human Misjudgment. Near the end he described what happens when several forces push in the same direction at once:
“When you get these lollapalooza effects you will almost always find four or five of these things working together.”
Munger was talking about how minds go wrong. But the idea travels. Put the right pieces together and the result isn’t the sum of the parts. It is something much larger. One plus one makes eleven.
This post is a thought experiment. What if Tesla and SpaceX became one company? Call it X. And what if the first thing it went after was the single most expensive problem in modern retail?
No merger has been proposed. Elon Musk has hinted at it twice, and that is all. Everything below about “X” is our imagination; every number is real and sourced at the end.
Two companies, one founder
Start with what each company already is, as of October 2026:
On the money side, the two look very different:
- SpaceX booked $18.7B of revenue in 2025, $11.4B of it from Starlink. Second-quarter revenue grew 92%. It still lost $4.9B in 2025, after big AI spending.
- Tesla booked $94.8B of revenue in 2025, down 3%. Operating income fell to $4.4B from $7.1B the year before.
One is a fast-growing network that is spending heavily. The other is a huge manufacturer whose car profits are being squeezed while it builds robots.
Already half married
Here is the part most people miss: these two companies are already tangled together.
- In February 2026, SpaceX bought xAI in an all-stock deal that valued SpaceX at $1T and xAI at $250B. That brought the AI lab, and the X social network it already owned, inside SpaceX.
- Tesla shareholders narrowly backed an xAI investment in November 2025. Tesla put in $2B in January, and now holds SpaceX stock. It booked a $1.0B paper gain on it last quarter.
- On Tesla’s July call, Musk spoke of “more and more overlap” between the companies, then added that combining them would need “the appropriate process”. In September: “Who can imagine what action one might take.”
So a merger wouldn’t join two strangers. It would formalise a family that already shares a founder, a balance sheet and an AI lab.
The four or five things working together
Munger said to look for four or five forces at once. Here they are:
- Energy. Tesla deployed 46.7 GWh of battery storage in 2025. Starlink ground stations, AI data centers and robot fleets all run on electricity that has to be stored somewhere.
- Intelligence. xAI’s models sit inside SpaceX. Tesla has collected about 12B miles of driving data under supervised self-driving. Data on one side, compute on the other.
- Connectivity. Every robotaxi and every robot needs a link that works everywhere. Starlink has 12M paying customers and its own satellites.
- Manufacturing. Tesla builds millions of complex machines. SpaceX has flown one booster 37 times. Between them they know how to make hardware cheap and reusable.
- Capital. Starlink’s cash flow could fund the robot buildout without the quarter-to-quarter pressure on a car company’s profits.
Each one is strong alone. Together, each makes the others more valuable. That is the lollapalooza.
A company called X
In our version, the combined company takes the name of Musk’s son X Æ A-Xii, known to his family as X. (The name has history: it already belongs to the social network now folded into SpaceX.)
Two more changes in the thought experiment:
- Elon Musk becomes chair emeritus. He keeps setting the long-term direction and steps back from running the business day to day.
- Gwynne Shotwell becomes CEO. She has run SpaceX’s operations as president since 2008, through the reusable-rocket era and the build-out of Starlink.
Why would Wall Street likely cheer? Because the two biggest discounts on Musk companies are key-person risk and governance. Musk holds about 82% of SpaceX’s votes. A proven operator in the CEO seat, with the founder as chair, tells investors the machine can run without one man’s attention, which is exactly what large funds pay up for.
The founder sets the destination. The operator runs the railroad.
Amazon’s most expensive mile
Now point X at a target. Look at the most expensive line in Amazon’s 2025 accounts:
Amazon spends more to ship its goods than it earns in operating profit. And the costliest stretch of every journey is the last one, from the local depot to your door. Capgemini’s research puts the last mile at about 41% of total supply-chain costs.
The last mile is expensive for a simple reason: people. A driver, a van and a stop every few minutes. Take away the driver and the cost changes completely:
Cost per mile
Human ride-hail today versus ARK's 2030 projection for Tesla's Cybercab. A projection, not a result.
Tesla has started making Cybercabs in Texas, with capacity above 125,000 a year. Its robotaxis carry passengers with nobody in the car in six cities. The same vehicle, a depot robot to load it and a robot at the curb to carry the box: that is an autonomous last mile.
Amazon knows this. It already runs over 1 million robots in its buildings. Its Zoox robotaxis started paid rides in Las Vegas in August. Prime Air drones fly in 11 metros. Morgan Stanley estimates warehouse robotics alone could save Amazon $2–4B a year by 2027.
But the doorstep is the hard part, and it is where X would be strongest.
What a cheaper mile is worth
Here is a simple illustration (see the note at the end). Suppose about 41% of Amazon’s shipping bill is last-mile cost:
Against $80B of operating income, a 25% cheaper last mile is a 13% profit boost. Halve it and profits rise by a quarter.
Whoever cuts that cost captures some of the prize. If X sells autonomous delivery, it gets a share of the savings on billions of parcels. And robots don’t stop at parcels. Morgan Stanley sees a $5T humanoid robot market by 2050, with over a billion units. Goldman Sachs recently raised its 2035 forecast to 6.5M humanoid robots a year.
A combined X would own every piece of that machine: the robot, the brain, the battery, the network and the factory.
The longest mile: same day, anywhere on Earth
The last mile is the most expensive. The longest mile is the slowest.
Amazon delivered 13 billion items the same or next day in 2025, but almost all of them came from a warehouse in the same country. Cross a border or an ocean and “fast” still means days: a freighter flight, customs, a truck to a depot, then the last mile.
Now add the one piece of X we haven’t used yet: Starship.
The idea is not new. In 2017 Musk showed a rocket that would carry people between cities through space, and said most long-distance trips would take “less than half an hour”. What is new is that the rocket now flies. On 28 September 2026, Starship reached orbit for the first time and released 26 Starlink satellites.
The Pentagon is already testing the idea for freight. Its Rocket Cargo program wants to move about 100 tons, a C-17 load, anywhere in the world in under an hour. In 2022 it gave SpaceX a $102M, five-year contract to show it can be done.
New York → Shanghai
London → Dubai
Los Angeles → New York
Typical nonstop flight times versus SpaceX's 2017 Earth-to-Earth figures. The rocket times are a pitch, not a timetable.
Put the pieces in a row and you get a chain that nobody else owns:
- Starship carries a container between continents in under an hour.
- Cybercab-style vans take it from the landing site to local depots, with no driver.
- Optimus robots load, sort and carry the last few metres to the door.
- Starlink connects every step, over oceans, deserts and ports.
That is same-day shipping, worldwide. Amazon would be the obvious first customer. It already runs its own fleet of cargo jets, and global speed is the one promise it still can’t make. But X would not need to own the shop. It could sell the longest mile to Amazon, to every other retailer and to every carrier, the same way Starlink sells bandwidth.
And beyond: the airlines
Cargo goes first because boxes don’t complain about g-forces. But once rockets carry freight safely and often, the next question writes itself. In 2017 Musk said a seat should cost “about the same as a full-fare economy plane ticket.”
Long-haul business travel is the richest slice of the $693B airlines earn from passengers. An executive who can fly New York to Shanghai and back in a day will pay a lot for it. Even a small share of that market is a very large business.
And the same vehicle, refuelled in orbit, is the one built for the Moon and Mars. Earth-to-Earth flight is just the closest destination.
The last mile cuts Amazon’s costs. The longest mile shrinks the planet. That is the case for one plus one making eleven.
What could go wrong
A lollapalooza can work in either direction. Munger’s original talk was about how several biases together lead smart people into bad decisions. So here is our checklist:
- No merger has been proposed. It would need independent boards, fairness opinions and shareholder votes on both sides.
- Tesla made $0.4B of operating income last quarter. The car business is funding the robot bet on thin margins.
- SpaceX lost $4.9B in 2025. AI spending is large, and Starlink has to keep paying for it.
- The two companies are already valued at about $3.6T on roughly $113B of 2025 revenue. A lot of success is priced in.
- Cybercab rides run in one geofenced city. Regulators still decide how fast autonomy spreads.
- $0.20 a mile is a projection. Amazon is building its own robots, not waiting to buy them.
- Starship reached orbit for the first time only in September 2026. Rocket flights between cities need offshore pads, noise limits and regulators on both ends, and the Pentagon paused its landing-test plan at Johnston Atoll in 2025.
- Cybercab and unsupervised robotaxis spreading well beyond today's cities.
- Optimus robots doing paid work outside Tesla's own factories.
- Starlink still adding millions of subscribers a year.
- A Rocket Cargo demonstration: Starship landing real freight somewhere far from its launch pad.
- Tesla's margins recovering as robotaxi revenue grows.
- A real board process, instead of hints on podcasts.
The Kuberin view
Munger warned that the same combining forces that build great businesses also fool great investors. Enthusiasm, social proof and a charismatic founder can stack up into a lollapalooza of misjudgment.
So hold both ideas at once. The pieces of X are real: rockets that land, a satellite network with 12 million customers, robots coming off a production line, a rocket that could make same-day shipping global, and the biggest shipping bill in retail waiting to be cut. Put them together under an operator like Shotwell, and the whole could be worth far more than the parts.
But X does not exist. Until a board puts real terms on the table, it is a story. Price the pieces, not the story.
One plus one can make eleven. Make sure you’re not paying for twelve.
Calculation note. The last-mile illustration assumes the last mile is about 41% of Amazon’s 2025 shipping costs of $102.7B, about $42.1B. Capgemini’s 41% describes total supply-chain costs, not Amazon’s accounts, so treat it as an assumption. Savings: 10% ≈ $4.2B, 25% ≈ $10.5B, 50% ≈ $21.1B, compared with 2025 operating income of $80.0B. Combined value and revenue: SpaceX ≈ $2.1T + Tesla ≈ $1.5T ≈ $3.6T; revenue $18.7B + $94.8B ≈ $113.5B. Starlink growth: 8.9M subscribers at end-2025 to 12.0M at end-June 2026, +35%. The cost-per-mile chart compares today’s average US ride-hail cost with ARK Invest’s 2030 projection for Cybercab. The flight-time chart compares typical scheduled nonstop times (rounded) with the point-to-point times SpaceX showed in 2017; no Starship has yet flown between cities. Air cargo and passenger revenue are IATA’s 2025 figures. Market values are approximate, early October 2026.
Disclaimer. This is a thought experiment. Tesla and SpaceX have not proposed a merger, and the company “X”, its name and the leadership roles described are hypothetical. Educational content only. Not investment, legal, tax, or financial advice, and not a recommendation to buy or sell any security. All investments involve risk, including possible loss of principal.
Sources. Munger, The Psychology of Human Misjudgment (1995) · SpaceX 2025 launch record (Space.com) · SpaceX Q2 2026 results (SEC) · SpaceX IPO filing financials (Via Satellite) · SpaceX–xAI merger (CNBC) · SpaceX IPO (CNBC) · Booster reuse record (UPI) · Gwynne Shotwell (CNBC) · Tesla FY2025 results (SEC) · Tesla Q2 2026 results (SEC) · Tesla shareholder votes (CNBC) · Musk on overlap (U.S. News / Reuters) · Musk on All-In (24/7 Wall St) · Cybercab public rides (InsideEVs) · Robotaxi cost per mile (ARK, via Electric Vehicles) · Amazon 2025 10-K (SEC) · US parcel volumes (Pitney Bowes) · Last-mile costs (Capgemini) · Amazon’s million robots (Amazon) · Morgan Stanley on Amazon robotics (Yahoo Finance) · Zoox paid rides (CNBC) · Prime Air expansion (Amazon) · Humanoid robots, $5T by 2050 (CNBC, Morgan Stanley) · Humanoid robots, 6.5M by 2035 (24/7 Wall St, Goldman Sachs) · Musk family (Wikipedia) · xAI acquires X (CNBC) · Market values (CompaniesMarketCap) · Amazon same- and next-day deliveries (Amazon) · Starship Flight 14 (Wikipedia) · Starship reaches orbit (Space.com) · Earth-to-Earth pitch, 2017 (CBS News) · Rocket Cargo program (Wikipedia) · Rocket Cargo (Spaceflight Now) · Airline industry outlook 2025 (IATA)