| Why would you shoot rockets into space? There are some classic answers. "To slip the surly bonds of earth and touch the face of God." "To boldly go where no man has gone before." "Not because it is easy, but because it is hard." "To make humans a multiplanetary species." To extract resources from Mars or spice from Arrakis or unobtanium from Pandora. To become younger than your twin. The 21st century answer, though, is mostly "to put satellites into orbit around the Earth to add functionality to our phones and televisions and computers and navigation systems." We have some screens, down here, and the screens work better if there are some electronics up there sending them signals. It's about improving technology down here, not exploration and adventure up there. Space flight is a complex and capital-intensive sort of consumer internet technology. These days the hot capital-intensive technology is artificial intelligence, particularly large language models that require enormous amounts of computing and electrical power to train and use. And so the natural answer, in 2026, to the question "why would you shoot rockets into space" is: "AI data centers." You've got a screen, and you want to ask the screen "please write my essay for 10th-grade English" or "help me code up a consumer internet app," and a vast space program will make the screen's answers faster and better. The world is just interesting, you know? The great marvels of human civilization are weirdly prosaic up close. Anyway one thing that Elon Musk is doing is this (emphasis added): SpaceX is requesting permission to launch as many as 1 million satellites into the Earth's orbit in order to pull off Elon Musk's latest grand vision of putting data centers in space to do complex computing for artificial intelligence. In a filing with the Federal Communications Commission made late Friday, SpaceX said it's creating the solar-powered network in order to "accommodate the explosive growth of data demands driven by AI." "Launching a constellation of a million satellites that operate as orbital data centers is a first step towards becoming a Kardashev II-level civilization — one that can harness the Sun's full power — while supporting AI-driven applications for billions of people today and ensuring humanity's multi-planetary future amongst the stars," SpaceX wrote in the filing. It's both, right? Chatbots today, intergalactic colonization tomorrow. In 3026 our nth-great-grandchildren will encounter a race of intelligent aliens in a distant galaxy, and the aliens will be like "how did you get here," and our descendants will say "well it started with some slop feeds" and the aliens will be like "what." Lots of very smart people have had high-minded ideas about intergalactic travel for centuries, but perhaps it will turn out that a critical step in getting there is building an app to generate short-form pornographic video on demand. Another thing that Elon Musk is doing, or thinking about doing, is this: Elon Musk is in advanced talks to combine SpaceX with xAI, according to people familiar with the matter, underscoring how the billionaire's vision for an artificial intelligence-powered world has grown too costly for any one of his entities to shoulder alone. Musk's rocket and satellite group and his artificial intelligence firm have informed some of their investors about the plans, the people said, asking not to be identified because the information is private. They may announce an agreement as soon as this week, some of the people said. ... A deal would combine two of some of the largest closely held companies in the world. xAI raised funds at a $200 billion valuation in September, while SpaceX was set to go ahead with a share sale in December at about an $800 billion valuation. The central catalyst for a merger is AI's insatiable need for capital. The rate at which xAI has been burning through cash — around $1 billion a month — has compelled Musk to further blur corporate boundaries, pool capital and rethink whether his moonshots should stay separate. SpaceX has also "discussed the feasibility of a tie-up between SpaceX and Tesla, an idea that some investors are pushing," Bloomberg has reported: "Tesla's ability to manufacture energy storage systems could help SpaceX use solar energy in space to run the data centers. Musk has also discussed using SpaceX's Starship rockets to carry Tesla's Optimus robots to the moon as well as to Mars." Combine 'em all! It would be satisfying if humanity's ambassadors to distant galaxies turn out to be humanoid robots made by a corporate successor to Twitter Inc. It just makes sense, not in an "actually Twitter is a distillation of human consciousness" way but in an "actually the greatest modern human achievements are funded and motivated by selling targeted internet advertisements" way. But also: The M&A rumination swirling around Musk's companies suggests there's more afoot than just clever financial engineering. It's a sign, if not an admission, that the billionaire's grandiose vision requires more capital than any one of his entities alone can provide in their current form. While executives at xAI have told investors that the company has the resources to continue spending aggressively, losses have swelled as the startup quickly consumes the funds it's raised in massive recent financing rounds. … Such outlays help explain why a historic initial public offering of SpaceX stock is now being framed not only as a bet on rockets, but as a way to fund space-based data centers that would serve as critical AI infrastructure. It also explains Musk's increasingly global hunt for capital, from Nvidia-backed GPU financing — where investors fund special vehicles that buy Nvidia chips and lease them to AI firms, using the hardware itself as collateral — to sovereign wealth funds in the Gulf. If Elon Musk's goal is to raise as much capital as possible, what's the best way to do it? One extreme answer, the one that he has mostly chosen in the past, is: Every time he has a new idea, he should start a new company and raise money for it separately. This has several advantages: - Every new idea he has is cool and will attract a lot of interest. "Ooh, brain implants," "ooh, rockets," "ooh, artificial intelligence," etc., [1] people think, and give him money. A mature electric car company is old hat, and "we're gonna do brain implants at a mature electric car company" is confusing. Launching a new company for each idea maximizes attention for the idea.
- Conceivably there are different investor bases for different ideas. If you run an AI fund, you might not want to invest in Musk's AI-flavored car company, but you might invest in his pure-play AI company.
- The employees of each pure-play company will be compensated in that company's stock, and will have good incentives to work hard on that company's single-minded project.
- There is a strange but valuable competing-for-Musk's-attention dynamic among the companies and their investors, which might make it easier to raise money. If you are an investor who loves rockets and wants to invest in every SpaceX funding round, and Musk comes to you for money to buy a social media site, you might reasonably think "well, I don't like this social media site, but if I give Musk money for this he'll let me into the next SpaceX round."
- "A basket of options is worth more than an option on a basket": If Musk starts 10 companies, and eight succeed fabulously and two fail, then he gets the fabulous success of the eight, while his downside on the other two is limited. Even a catastrophic failure at one company won't directly affect the fortunes of the others.
- Contrary to that last point: If one company does run into trouble, and Musk wants to bail it out using another company's money, he can. He doesn't have to, but he has a free option. If he wants to move money from one company to another, he can go to the donor company's board and say "hey gotta send out a few billion dollars" — or employees or computer chips or whatever — and the board will be like "whatever you say boss." Keeping the companies separate does not really hinder him from pooling their resources, because he's Elon Musk. As Byrne Hobart writes: "Part of Elon's model is to control a portfolio of companies, have each one take actions that expand the variance of outcomes, and then, whenever one of them faces downside from this, either use financial resources from borrowing against another one or the human resources within the companies to fix it." [2]
The opposite extreme answer is of course "combine them all into one mega-company." I sometimes call the hypothetical mega-company the Musk Mars Conglomerate, because its notional purpose seems to be to colonize Mars. The reason for taking that approach could be that there is some limit to how much money even Elon Musk can raise, and some limit to his ability to move it between his companies: - If he wants to raise a lot of money for SpaceX or xAI, it might help to cross-collateralize them with each other and/or Tesla. An option on a basket is cheaper than a basket of options.
- If he wants to pivot all of, say, Tesla's resources into building data centers for xAI, there will be some frictions; shareholders and directors might have complaints. Musk's separate companies have unusual freedom to work together, but it is not infinite. Whereas if they were all merged, he could do whatever he wanted, and everyone would just think "ah yes the chief executive officer of this big company is allocating capital to its best uses." If his skill is allocating capital, it is good to free him up to do more of that.
- Similarly, if you are investing in any Elon Musk company, what you are buying is not so much "cars" or "rockets" or "robots" or "social media" or "brain implants," but rather that ability to allocate capital to the next big thing. If you invest in any particular Musk company, there's a risk that the next big thing will happen in a different one and you will be left behind. But if all of the Musk companies are combined, you can be confident you'll get the best of him.
One way to think about what is going on here is that, for most of Elon Musk's career, "tech" was an essentially equity-funded sort of business. Musk went around launching various cool projects using money from people who were excited about the upside of those projects. Compared to other technology entrepreneurs, Musk did much more capital-intensive projects — cars and rockets, not websites — but the investor psychology was similar. His investors were venture capitalists, or public shareholders who thought a bit like venture capitalists; they were willing to take big risks for big wins. That model is naturally suited to launching a bunch of cool companies. Each new company is cool, so it attracts hype and investment. Each has one exciting idea, so if it works out you'll get rich. Each is unencumbered by the other companies, a pure bet on an exciting long-shot proposition. If the brain implant company works, great; if it fails, your investment goes to zero; each investment is its own silo with its own exciting idea and its own all-or-nothing payoff. But modern artificial intelligence turns out not to be that sort of business. I wrote last year, about OpenAI, that "it turns out that the supply of science-fiction-suspension-of-disbelief capital is really quite large but not trillions of dollars": Trillions of dollars is going to tap out every source of venture capital, every investor who is willing to put in money now for an uncertain but potentially huge future return. To get trillions of dollars you will need boring debt investors, people who need their money back, people who are willing to put in money now for 9% a year and repayment in five years. Very different psychology.
And so the financing action in AI is, yes, some multibillion-dollar equity raises at enormous valuations, but also a lot of debt securitizations for data centers. At Bloomberg Markets Magazine, Paula Seligson reports that "last year, AI-related companies and projects tapped debt markets for at least $200 billion," and "projections are in the hundreds of billions of dollars of issuance for 2026 alone." That model is arguably more suited to a giant diversified conglomerate. If you are going to lend Elon Musk hundreds of billions of dollars to build data centers, you will be less excited about pure all-or-nothing payoffs and more interested in things like steady cash flows (from selling cars?) and cross-collateralization. A bigger and more diversified Musk Mars Conglomerate might be less exciting for Musk's traditional shareholder base, but more soothing for the more conservative lenders, lessors and other investors who will be needed to fund all those data centers in space. If the whole economy is becoming data center securitizations, then the Elon Musk complex might have to become a data center securitization too. Is building autonomous war robots securities fraud? | Two themes around here are: - "Everything is securities fraud": When a public company does a bad thing, it will be sued for securities fraud, because every bad thing can be recast as a fraud on shareholders.
- "What if the thing that stops artificial intelligence from enslaving humanity is some weird wrinkle of the legal and financial system, like insurance requirements or copyright law or the ban on onion futures?"
The obvious synthesis is: What if the thing that stops artificial intelligence from enslaving humanity is that everything is securities fraud? That is, what if AI would develop autonomous warfare capabilities that could wipe out humanity, but that development will be thwarted because (1) that's bad and (2) bad things are securities fraud? The Washington Post reports: Google breached its own policies that barred use of artificial intelligence for weapons or surveillance in 2024 by helping an Israeli military contractor analyze drone video footage, a former Google employee alleged in a confidential federal whistleblower complaint reviewed by The Washington Post. … At the time, Google's public "AI principles" stated that the company would not deploy AI technology in relation to weapons, or to surveillance "violating internationally accepted norms." The whistleblower complaint alleges that the IDF contractor's use contradicted both policies. The complaint to the SEC alleges that Google broke securities laws because by contradicting its own publicly stated policies, which had also been included in federal filings, the company misled investors and regulators. I suppose that, short of Terminator-style time travel, we'll never be able to prove the counterfactual, but I tell you what: If "everything is securities fraud" really is what saves humanity from being wiped out by evil war robots, I will expect at least a share of that Nobel Peace Prize. The weird situation in US sports gambling is: - Some US states prohibit sports gambling, while other states allow it in heavily regulated form.
- Kalshi, a prediction market registered with the US Commodity Futures Trading Commission, offers sports gambling in all 50 states, without complying with any state's rules.
- States complain, but Kalshi argues that it is regulated by the CFTC, and that CFTC regulation preempts state regulation: The states can't interfere with the CFTC's enforcement of CFTC sports-gambling rules with respect to CFTC-regulated sports gambling sites.
- But the CFTC doesn't regulate sports gambling. It has never explicitly said that sports contracts are legal on CFTC-regulated trading platforms, and in fact it has occasionally suggested the opposite. It does not explicitly approve the sports contracts; Kalshi "self-certifies" them. It has no specific rules related to sports gambling; no specific guidance about market manipulation or insider trading or customer suitability in the sports space. It's just a gap.
- Nonetheless, Kalshi's arguments mostly work in court, and it keeps listing sports contracts. There is federally regulated online sports gambling, but it is not actually federally regulated.
Just weird! This is one of those situations where it's tempting to be like "well that can't last," and then to think for a minute and conclude "actually sure it can." The CFTC, in the Trump administration, clearly likes Kalshi's sports contracts, and Donald Trump Jr. is an adviser to Kalshi. But it might be a little awkward for the CFTC to explicitly publish rules allowing sports gambling on Kalshi, since the legislative history around event contracts suggests that Congress never intended to allow sports gambling on commodity exchanges. Just ignoring it while Kalshi goes about its business, and seeing what happens in court, might be the simplest approach for the CFTC. But, nope: Wall Street's main derivatives regulator will write new rules for the multi-billion dollar prediction markets industry, the head of the Commodity Futures Trading Commission said Thursday. "It is time for clear rules and a clear understanding that the CFTC supports lawful innovation in these markets," Chairman Michael Selig said in prepared remarks. "Consistent with my commitment to fostering responsible innovation in crypto asset markets, I will continue to support the responsible development of event contract markets." It's not clear exactly what that means, but if you read Selig's speech it seems pretty bullish for Kalshi, Polymarket and other commodity/event/sports-gambling platforms: First, I have directed CFTC staff to withdraw the 2024 event contracts rule proposal that would prohibit political and sports-related event contracts and the 2025 staff advisory, which cautioned registrants about offering access to sports-related event contracts due to ongoing litigation. ... Second, looking ahead, and in the spirit of markets that trade on expectations, I have directed CFTC staff to move forward with drafting an event contracts rulemaking. ... Third, I have directed CFTC staff to reassess the Commission's participation in matters currently pending before the federal district and circuit courts. Where jurisdictional questions are at issue, the Commission has the expertise and responsibility to defend its exclusive jurisdiction over commodity derivatives. That is: First, the CFTC has tried to prohibit sports gambling in the past, and will now explicitly reverse that position. Second, the CFTC will make new rules about, one assumes, sports gambling. Third, it will get involved in those state court cases "to defend its exclusive jurisdiction over commodity derivatives," that is, to tell the states that they're not allowed to interfere with online sports gambling when it's federal online sports gambling. This remains a weird situation just because it does not seem to me that Congress intended, or that anyone really wants, the US federal commodity derivatives regulator to get into the sports gambling business. But it's less weird than the status quo. If we're going to have federally regulated sports gambling, at least it will actually be federally regulated. Just to close a loop: I wrote on Thursday that "I once heard that the Harvard major whose students were most likely to be rich 20 years after graduation was art history. I have no idea if this is true and can find no evidence for it, but doesn't it sound like it should be true?" Felix Salmon points out that it probably came, in some indirect way, from this 2012 report about "What the Top 1% of Earners Majored In." As Virginia Postrel wrote here at Bloomberg Opinion in 2014: [Art history is] stereotypically a field for prep school graduates, especially women, with plenty of family wealth to fall back on. In fact, a New York Times analysis of Census data shows that art history majors are wildly overrepresented among those in the top 1 percent of incomes. Perhaps the causality runs from art history to high incomes, but I doubt it.
Right. What I wrote on Thursday was that "people who treated college as a luxury consumption good were in certain ways more appealing to investment banks"; being a prep school graduate with lots of family wealth to fall back on makes you in some ways an appealing investment banking candidate. (In other ways, not, and a lot of banks like to hire "P.S.D.'s, as in Poor, Smart, with a deep Desire to become rich.") Investment banks might want the art historians, but the art historians don't really need the job. Astrology-based investing | To close another loop, we have talked recently about astrological investing, and here's the CLSA Feng Shui Index for last year, the Year of the Wood Snake. "Secret hissings point towards good rises in the market through the spring and winter, a long period of summerish lassitude, and small falls throughout the year," etc. One more loop. We talked last month about a guy who launched a crypto token backed by US pennies in "a vault you can visit," possibly his closet? The token was trading at about 19 cents as of noon today. That is a much better mNAV than Strategy Inc. has managed, though obviously in smaller size. The lesson is perhaps "the premium to net asset value of a digital asset treasury company [3] is partly driven by a flywheel of accretive dilution but is mostly driven by comedy," which we kind of already knew. How Chinese Speculators Set the Stage for Gold and Silver Crash. China's 'gold fever' sparks US$1 billion scandal as trading platform collapses. Metals Traders Lose at Least $144 Million as 'Hat' Flees China. Carvana's Red-Hot Growth Runs on a Cycle of Borrowed Money. The $100 Billion Megadeal Between OpenAI and Nvidia Is on Ice. 'Spy Sheikh' Bought Secret Stake in Trump Company. Rare SpaceX Bet Turns $1.1 Billion Fund Into a Retail Magnet. Oracle to Raise Up to $50 Billion in 2026 for Cloud Buildup. AI Boom Is Triggering a Loan Meltdown for Software Companies. How HSBC fashioned a $600bn debt machine. The Wild Markets Behind Polymarket's 'Truth Machine.' How a Silicon Valley Startup Became a Crypto Lifeline for Venezuela. When the Last Local Bank in Town Turns to a Global Salad Company for Help. Disney Is Close to Picking Parks Chief D'Amaro as Next CEO. Wealthy and Powerful in Spotlight After Latest Epstein Release. Trump Lawsuit Against IRS Puts Him on Both Sides of the Same Case. Minnesota Businesses Resisted ICE, and Now Face DHS Audits. "A corporate officer's sexual harassment of employees will not constitute a breach of fiduciary duties except under limited circumstances." Young girl topples tower made from 63,000 beer coasters created by artist in a bid to break Guinness World Record. Between the sheets at the college Excel championships. If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Or you can subscribe to Money Stuff and other great Bloomberg newsletters here. Thanks! |
No comments:
Post a Comment