The Money Stuff column titled Obviously SpaceX and Tesla Will Merge, dated Wednesday, October 21st, 2026. Section one: SXSI/Tesla One model of Elon Musk five years ago is that he had a bunch of disparate interests and was really passionate about all of them. (Start of Footnote one: Another model is that he had a bunch of interests that seemed disparate to an outside observer but were all really about colonizing Mars. I don't think this would change the analysis too much. End of Footnote one) One minute, he'd be like I want to make self-driving electric cars, and another minute he'd be like actually I want to colonize Mars, and then he'd be like I want to teach a monkey to play Pong with telekinesis, and then he'd be like actually I want to make Twitter more right-wing, and then he'd be like I don't like traffic so, tunnels. And he was enthusiastic and persuasive and had a good track record of succeeding with kooky-sounding ideas, so you'd be caught up in whatever crazy-sounding thing he was telling you. If you had a billion dollars and happened to talk to him about that day's enthusiasm, you'd give him the billion dollars to do the thing. In some sense, the billion dollars was not for brain implants or tunnels or Twitter or whatever; you were giving him the billion dollars as a bet on him, as a bet that whatever weird thing he touched he could turn to gold. But in another sense, the tunnels mattered; his enthusiasm for specific weird projects was contagious. Perhaps you found tunnels boring but were really excited about brain implants, so you gave him money for some projects but not others. (Start of Footnote two: Sorry. End of Footnote two) I don't know that this was the dominant model, though. More likely, on Monday you gave him a billion dollars for space rockets, and then on Wednesday he came back to you and was like I need a billion dollars for tunnels, and you were like oh, right, of course, the previous billion dollars was for the rockets, I can see why you'd need a different pot of money for the tunnels. In any case, by doing all the different projects in different companies with (possibly) different investors, he maximized his fundraising surface area. All of his investors got concentrated high-upside bets on whatever the relevant enthusiasm was. And then he went and did all of the things. There were some conflicts of interest involved in having one guy run all these companies doing all of these things. For one thing, there are only so many hours in a day, and if he's spending 20 hours a day on Twitter he can't spend 20 hours a day designing cars. Also, making Twitter right-wing might alienate a lot of electric-car buyers. But I don't want to overstate this. These were very different things! A car company just does not compete with a space-rockets company, which does not compete with a brain implant company, which does not compete with a social media site. Oh there's overlap; maybe the car company will borrow the rocket company's rockets to make its cars fly. But those overlaps are mostly cute and cool; if the flying cars work, no shareholder is going to go around complaining I think the car company overpaid for the space company's rockets. Running six wildly different companies is personally taxing for Elon Musk, but the investors' bet is that he can handle it. Otherwise, the conflicts are fine. One model of Elon Musk today is that there is only one thing. The one thing is I guess AI, artificial intelligence, though now he is calling it SI, super intelligence. Arguably the one thing is really physical infrastructure for AI. There is a multitrillion-dollar buildout of AI infrastructure going on, and it turns out that one thing that a lot of Elon Musk's disparate companies are good at is building complex physical infrastructure quickly. That was useful for building cars and sending rockets to space, so they got good at it, and now it is critical for AI. SpaceX, which now contains Musk's rocket business, and his artificial intelligence lab, and also Twitter, had an earnings call yesterday, and here's how Musk put it. (Start of Footnote three: Later he talked about blades for turbines for power plants: We're casting the blades and vanes for the turbines. So SpaceX already has a very advanced castings capability that was developed for rocket engines. So we have, I think, actually the most advanced casting capability on earth because the demands for castings for rocket engines exceed the demands for anything else you could possibly imagine. So taking that incredible team that has, I think, objectively, the most advanced casting technology in the world and saying, hey, can you guys also make some blades and vanes for terrestrial turbines, is much easier than what they have done in the past. And when we talk to the turbine suppliers, they tell us that the blades and vanes are 12 to 18 months behind the rest of the turbine, meaning they can deliver everything except the blades and vanes 12 to 18 months before they can deliver the blades and vanes. So casting is this really difficult bottleneck. And if anything, it seems to be getting worse. So that means that we can bring power online 12 to 18 months faster than anyone else. End of Footnote three) So talking about SpaceX Super Intelligence, SpaceX SI, sexy, whatever. So I think people are starting to realize the constraint on deployment of super intelligence is more in power and cooling and infrastructure than it is in chips. So what we've done at SpaceX is we've taken a team that has been able to achieve a market share in mass-to-orbit that is almost 90 percent of all mass-to-orbit of Earth, where China is about, I think, roughly 6 percent, and the rest of the world combined, including the rest of the U.S. is about 4 percent. So it's a team that has been able to achieve results that are an order of magnitude greater than the rest of all competitors in space combined, and applied that same team to the deployment of terrestrial compute. So this is a bit like the New York Yankees playing in Little League. It's not fair. But the net result is being able to deploy power cooling infrastructure at a rate that is vastly faster than anyone else, and may at some point be faster than everyone else combined. It's not obvious that a company that is good at putting satellites into space would also be good at building data centers for large language models, but it seems to be true. SpaceX has also talked about putting the data centers into space, though I don't know that that's essential to any of this. Meanwhile, Tesla Inc., the car company, is also increasingly an AI company. The cars are supposed to be self-driving; that's pretty AI. Once you are in the business of building autonomous electric machines that move around on their own (the cars), you might think it's just a short hop to building autonomous electric robots that move around on their own and have opposable thumbs, and so in fact Musk is building a robot army at Tesla, absolutely. The synergies between brain implants and AI seem pretty clear, and Twitter is a good distribution channel for an AI chatbot. Presumably the tunnel-drilling machines could help with building data centers? If all of the companies are AI companies, though, there are a lot more potential conflicts. They all need the same scarce Nvidia chips to power their AI, for instance, at least until Tesla builds its own chip fabrication plant. (At which point SpaceX SI will presumably be a big customer.) If they are all building AI models, the company with the best AI researchers will have the best models. There are tons of possible synergies: If you build the best general-purpose AI model, maybe it can drive the cars and operate the robots and upload people's brains and serve right-wing takes on Twitter. AI is a very general-purpose technology (Anthropic is, like, a law firm and a biotech company?) and so it can be applied to all of Musk's formerly disparate businesses. In this world, Elon Musk's critical job is prioritization. His job is not to launch six kooky new ideas; his job is to allocate scarce resources to the highest-value areas to win the race to build AI, colonize space, etcetera. If there are limited chips, they should go to the projects that will have the largest net present value, and Musk should decide which projects those are. Meanwhile, the fundraising environment has also changed. With a multitrillion-dollar AI buildout, Musk's ability to get venture capitalists and his billionaire buddies to kick in a billion dollars to fund a kooky new idea is not that relevant. What is relevant is getting credit investors to kick in tens or hundreds of billions of dollars to fund infrastructure. Credit investors do not want concentrated bets on weird stuff with high upside; they want to get paid back with interest. A giant company with a diversified set of bets is a better credit risk than a spiky company with telekinetic monkeys. Give us 100 billion dollars to build data centers in space, but if for some reason we don't build data centers in space you can repossess our car factories and data centers on earth is a good pitch to credit investors. Anyway, of course SpaceX and Tesla should merge. It's crazy that they are separate companies! One guy runs two gigantic physical AI companies with huge overlaps! What a nightmare of conflicts of interest and red tape! Just merge them, so that the one guy can come into the office each day with a simple mandate: Make the choices that maximize overall value for the giant AI complex, and for the shareholders of that complex. If some chips or engineers or drills or whatever should be working on robots, work on the robots; if they should be working on data centers, work on the data centers. The shareholders have come this far; surely they trust Musk to make those decisions. It's just that right now, working on the robots is good for Tesla and bad for SpaceX; working on the data centers is bad for Tesla and good for SpaceX. Musk can't put both of their interests first. There's an easy fix. I do wonder what the special committee meetings are like: Elon Musk openly endorsed the prospect of combining SpaceX with Tesla Inc. in what would be the biggest-ever corporate merger, uniting assets from electric cars to AI data centers to rocket and satellite capabilities. The world's richest man said Tuesday the boards of both companies had formed special committees of independent directors to consider the idea of a combination. There's a lot of merit to combining the companies, Musk said on a conference call after SpaceX announced earnings that beat estimates. Like it's easy for me to be like this is a dumb structure and you should smoosh them together, but the devil is in the details. SpaceX and Tesla both have pretty volatile stocks; since SpaceX went public in June, its market capitalization has ranged from about 1.4 trillion dollars to about 2.6 trillion dollars, while Tesla's has ranged from about 1.2 trillion dollars to about 1.6 trillion dollars. Presumably any deal would be an all-stock merger-of-equals-ish in which neither company gets a premium, but you still have to pick an exchange ratio. You have to get banks to provide fairness opinions. Do discounted cash flow valuations of the data centers in space and the robot army, fun stuff. I assume you have to pay the bankers for those fairness opinions, but how much? This would in theory be a 3.5 trillion dollar merger, and bankers are traditionally paid a percentage of deal value; a 0.1 percent fee would be 3.5 billion dollars. I would take the under on that. The bankers are not, I suspect, going to be doing a lot of heavy lifting on valuation, negotiation, due diligence or financing. This is one guy merging his two companies; he could negotiate the terms with himself in five minutes. Technically he won't, though; he said on the earnings call: Because I'm a conflicted party, I can't actually be in the discussions of the special committees. So it's sort of a, it's an odd kabuki dance where two groups of people are talking, and I can't be in the conversation. Just loom over the conversation a bit. Also, I assume it's a merger of equals, but is that right? (Start of Footnote four: To be clear, merger of equals does not mean each company gets exactly half; it means neither company is the big acquirer or the teeny target, so everyone swaps their stock for stock in the combined company and there's not a big premium paid to the target. Also even in a merger of equals there is usually one acquirer for legal purposes. (Not always; every so often there's a top hat merger where a new holding company acquires both companies.) I assume the acquirer would be SpaceX, in part because it has a bigger market capitalization but mostly because of Musk's greater control there. End of Footnote four) Tesla has a single class of stock with one vote per share; Musk only controls about 20 percent of the stock. SpaceX has dual-class stock; Musk has 84 percent of the vote plus extra rights to appoint the board and generally do whatever he wants. Presumably SpaceX would be the surviving company in the merger, so that Musk could ensure perpetual control of his robot army. I feel like a Tesla special committee might reasonably demand a premium for that: Right now, Tesla shareholders have some rights and SpaceX shareholders have none, so Tesla shareholders should get some extra value in a merger for giving up those rights. But of course an actual Tesla special committee might not think like that. No, giving up voting rights to Elon Musk is good for Tesla shareholders. Also, Musk has a trillion-dollar contingent pay package at Tesla, and a nearly-trillion-dollar contingent pay package at SpaceX, and merging the companies and keeping both pay packages would arguably overpay him. Does this matter? Feels like the sort of thing an independent special committee might traditionally want to consider, though maybe it's silly here. We're colonizing Mars! What's an extra trillion dollars among friends? I say in the headline that obviously they will merge, but I don't actually want to be too confident. Musk says: the special committees have not reached agreement on the terms of any transaction, and obviously, there's no guarantee that they will. (But they will?) The deal would require shareholder approval: easy at SpaceX, where Musk controls the vote, but ever so slightly harder at Tesla, where public shareholders would need to approve. (But they will?) The real impediment is regulatory: These are two giant companies with lots of critical businesses and a politically charged owner; every country's antitrust regulators will take a close look. Like I said at the beginning, their businesses were traditionally quite different and did not have much overlap, which suggests antitrust approval should be fine, but like I said after that, their businesses are increasingly converging, which suggests it might be challenging. Musk said on the call: I do think there's a lot of merit to combining companies as we are pretty much in the super intelligence singularity and there's a lot to be gained from cooperating on developing super intelligence software, chip design, the Terafab, obviously, robotics, mass-to-orbit with the Starmind constellation And I think that's right. In a singularity, all correlations go to one, or something. If Elon Musk is going to run a gigantic AI infrastructure company, it should be one company with one set of shareholders. That would solve a lot of governance complexity and make everyone a lot more chill about Musk's potential conflicts of interest. They just have to do one last conflicted transaction to get there. Section two: Withdrawal strategies The defined-benefit pension is a good product. What you want, when you are saving for retirement, is a certain amount of income in retirement. A pension gives you that. You will get five thousand dollars a month of income in retirement (in today's dollars), yes, good, that's what you want. I mean, maybe you want a higher number, but that is the kind of number you want. You want a number of dollars that you will be able to spend each month during the indefinite, hopefully long, period in which you are not working. Other forms of number are bad. You will be able to spend five thousand dollars a month in retirement for 137 months: not ideal. You will begin retirement with 1.3 million dollars, but that number will fluctuate with market conditions, and you can draw down as much as you like each month until you hit zero: what, no, terrible. More subtly, the defined-benefit pension is a good funding model. If you are not a person saving for retirement, but rather a financial intermediary looking to raise money to invest in stuff, running a pension is pretty good. A defined-benefit pension, for a 35-year-old worker, has the form this person pays you some money every month for 30 years, and at the end of the 30 years, you will need to start paying her back five thousand dollars a month. That's a lot to work with! Your obligations are predictable (you know the schedule) and extremely long-dated (you don't have to pay anything for 30 years, and then you only have to pay out a little bit of the money each month). You can invest in all sorts of spicy stuff, earn a lot of illiquidity premiums, and take a certain amount of risk because you have such a long time to earn back any losses. You have very stable funding, so you can be a patient provider of capital, so you can get paid a lot for that capital. The defined-contribution retirement plan is less good. For employees, it has the form put as much as you want into whatever investments you want, and in 30 years you will have whatever you put in plus whatever the market return on that stuff is, and you can draw down what you want each month. Just a mess. The one thing you care about, your monthly income in retirement, is a complicated math problem with uncertain inputs. For financial intermediaries, it's also not great. The customers get a menu of investment options, so their money is not locked up in one thing for the long term, so you have less ability to take risk and harvest illiquidity premiums. Also, though, the customers can log into their accounts and see a total dollar amount. When the dollar amount goes down, they will get sad, and maybe move their money somewhere else. No no no, you tell them, the dollar amount going down is fine: The dollar amount went down because expected rates of return went up, so the lower dollar amount today will support the same monthly payment when you retire in 30 years; everything is working out perfectly. They will not believe you. The dollar amount went down. I've said all of this before. The solution is pretty obvious: Financial intermediaries can roll their own defined-benefit pensions. A roll-your-own defined-benefit pension looks approximately like this: 1. Employees put money into their retirement plan every month. 2. The money goes to a target-date fund that invests in a mix of investments designed to provide for retirement starting in, say, 2055 (or whatever the employee's expected retirement date is). Because the target date is a long way off, the fund does not need a lot of liquidity, and can invest in fairly spicy stuff. 3. In 2055, the employee retires and stops contributing to the plan, and the target-date fund becomes an annuity, providing payments of X dollars per month for the rest of the employee's life. (Or some variation: X dollars per month plus an adjustment for inflation; X dollars per month for 30 years; whatever.) 4. Before 2055, when the employee logs in to her retirement account, her total amount of money in the fund is displayed in a relatively small font, and the expected annuity in retirement is displayed in a relatively large font. This is all pretty straightforward and lots of retirement plans offer some form of it. You don't have to go all the way, though. You can gesture at this by, for instance: 1. Just having a normal four oh one kay plan, but 2. Displaying the amount of money in the fund in a small font, and displaying the expected annuity in a large font. You don't even have to sell the customer the annuity. You can just solve the math problem for her. You can calculate the expected annuity; you know, you can probably spend about 4 percent of your money each year, so if you have 1 million dollars in your account that's like three thousand three hundred thirty-three dollars per month. (This is called the 4 percent rule and is a pretty standard rule of thumb, though there is an endless literature about how it can be bad.) A helpful data point, and arguably a more relevant number than the total dollar amount. Does this allow you to lock up funding and take longer-term risks and harvest illiquidity premiums? I dunno. The Wall Street Journal reports: Capital Group, the country's largest manager of active target-date funds, will start offering research-backed withdrawal strategies for holders of its funds. The investment manager landed on two different approaches, depending on a person's goals and expenses, after studying the hypothetical performance of a retirement portfolio across market scenarios. It is set to publish recommended annual withdrawal ranges and the research supporting them on Wednesday. For the 11 million four oh one kay plan participants that hold nearly half a trillion dollars in Capital Group's target-date funds, the withdrawal approaches are intended to be an informational starting point for figuring out an often stressful problem: how much to withdraw from a nest egg, without running the risk of depleting it too quickly. Some of Capital Group's biggest retirement fund competitors, including Vanguard Group, BlackRock, and Fidelity, have recently announced new target-date funds that attempt to tackle the retirement-income puzzle by embedding the option to buy an annuity, giving those who opt-in some level of guaranteed income for life. Capital Group has yet to roll out such a product. Michelle Black, the principal investment officer of Capital Group's target-date retirement funds, says annuities aren't necessary if a smart drawdown strategy is employed. Thinking about retirement income, we wanted to create something where the beauty is in the simplicity, Black said in an interview. Our framework is easy to understand, and it's also something you can opt to make a change to at any time. Buying an annuity is a decision you can't undo. Is Capital's research-backed withdrawal strategy the 4 percent rule? Sort of: Capital Group's two withdrawal strategies will be conveyed to account holders through employer materials and four oh one kay account portals. The first, a fixed-dollar, inflation-indexed strategy, is reminiscent of the 4 percent rule that many financial advisers have recommended for decades. Under that approach, Capital Group recommends taking a withdrawal between 4 percent and 4.5 percent of the total account balance at retirement, and then withdrawing that same dollar amount annually, with adjustments for inflation. The second, variable-dollar, approach involves taking a withdrawal of between 5 percent and 5.5 percent of the retirement account balance at the end of each year. This does seem a bit more complicated than you get five thousand dollars a month. Section three: Significant liquidity I'm sorry, this is just a cool thing to say: I'll be coming into significant liquidity, personally, as a result of exiting certain investments, he said. Those will generate billions of dollars of liquidity. That's Todd Boehly, quoted in a story by Bloomberg's Alexandre Rajbhandari and Allison McNeely about his company, Security Benefit Life Insurance Co. I don't think I've ever had occasion to say: I'll be coming into significant liquidity, personally, but it's something to aspire to. I think a lot about Jules Feiffer saying (of David Shaw!) they offered me real money, and I was in the market for real money. This has a similar ring to it. Anyway, we have talked a bunch recently about (as Rajbhandari and McNeely put it) Boehly's longtime business associate, billionaire sports mogul Mark Walter. They write: The two billionaires helped pioneer what has become one of Wall Street's most lucrative strategies: arranging for an investment firm to take over a life insurer, dial up annuity sales and plow the cash into higher-returning, and potentially riskier, assets. Along the way, the profits helped them build personal empires, including their joint ownership of Major League Baseball's Los Angeles Dodgers and other sports teams. The basic insight is that life insurers control big pots of long-term locked-up money, because they sell life policies and annuities that don't have to be paid back for years. So if you are a smart investor and you get control of a life insurer, you can use that funding to make cool investments, make a lot of money, use some of it to pay off the policies and keep the rest for yourself. Traditionally life insurers are run by conservative people who invest the long-term money in conservative stuff, but that was arguably a market inefficiency. Live a little! Insurers have safe stable funding, so they should be bearing some risk. Not too much! They shouldn't be buying lottery tickets; they do need to have the money when the insurance comes due. But a theory like I am a good investor who will make money on my investments, so I should go out and take over pots of money to invest is not unreasonable. You do have to be upfront about it. The problem with Walter is not that he used policyholder money to fund his own fun cool investments (in sports teams, etcetera); it's that his insurers didn't disclose that: They reported loans to affiliated businesses as non-affiliated transactions, making regulators think they were safer than they were. Rajbhandari and McNeely write: Boehly is in a much better position, thanks to Security Benefit's long-running disclosures. The firm, once controlled by Walter and now central to Boehly's Eldridge Industries investing conglomerate, has for years noted that it has billions of dollars of affiliated assets, amounting to 42 percent of its invested holdings. Such deals are allowed as long as they're publicly disclosed for additional scrutiny. About 7 percent of all US life insurers' assets were invested in affiliated entities at the end of last year, according to a recent report by the National Association of Insurance Commissioners. Right, yes, the whole point of buying the insurer is to invest the insurance money in Boehly's cool stuff. If you disclose that clearly and everyone knows it, it's fine. Section four: Things happen Global bond markets are adjusting to a less forgiving buyer base. Larry Ellison Went on Borrowing Spree to Bankroll His Son's Warner Bros. Bid. Anthropic S-1 Shows Red-Hot Growth, Risks and Massive Computing Costs. Musk Ordered Starlink Slowdown That Blindsided Ukraine. Nvidia CEO Huang Pushes Europe on All Industries Adopting AI. OpenAI agents disabled company's monitoring systems. Hackers Used Claude and Grok to Fully Automate Attacks on Banks. Nscale boss seeks to reassure staff amid AI market fears. Trump Says Data Centers Must Provide Their Own Electricity. NYSE Is Getting Trounced in the Battle for Wall Street's Biggest IPOs. Abbott Spins Off Embattled Business That Sells Baby Formula. How Coca-Cola Got Burned With a 5 billion dollar Bet on Costa Coffee. Crypto exchanges race to offer prediction markets to US customers. New York pensions chief says private equity firm sought to expense jet use. What's going on with chocolate is a parable for the modern world. Of course, most recognize that Borges was a visionless fool who failed to recognize the commercial potential of his library. That was the Money Stuff column titled Obviously SpaceX and Tesla Will Merge, dated Wednesday, October 21st, 2026, as read by Robo Matt. All credit for Money Stuff goes to the biological Matt Levine and Bloomberg. Subscribe to the money stuff newsletter and podcast at bloomberg dot com.