Money Stuff: Everybody Wants Some AI Debt

Probably everyone in the financial industry, and possibly everyone in the world, has an opinion about the artificial intelligence boom. Many of them have views on whether we are raising too much money for data centers, on who is building the best AI models, and on the prospects of particular big AI-related technology companies like Nvidia or Meta or Alphabet or Microsoft or Oracle or OpenAI. And while a lot of those views are optimistic (AI companies are raising a lot of money), there is some diversity too (a lot of people complain about a bubble).

Some people in the financial industry, and probably outside of it, have thoughts about Olo Inc., “a leading open SaaS platform for restaurants” that Thoma Bravo bought for $2 billion in September. But most people don’t. It is perfectly normal never to have thought about Olo. I myself have thought about AI and data centers and Nvidia and OpenAI a lot, but I had never thought or even heard about Olo until I searched Bloomberg for recent direct lending deals and Olo’s acquisition financing popped up. [1] Also there is probably not that much diversity of opinion on Olo, or at least, not that much intensity of opinion. If you work on the Thoma Bravo deal team, or at the private credit funds that loaned Thoma Bravo money to buy the company, you probably have some thought like “yeah, Olo, pretty good, good restaurant Saas.” If you don’t, you are probably not going around thinking “I need to short this bubble in restaurant SaaS.” For the most part, either you’re in the deal or you’re indifferent. 

Stereotypically, in recent years, the words “private credit” (and “direct lending”) have meant something like that Olo acquisition financing. Private equity sponsors buy companies, and for reasons of speed and certainty and flexibility and relationships, they decide to borrow the money to buy the companies from private credit firms. Private credit firms are experts in things like having good relationships with private equity sponsors and evaluating the cash flows and deal structures of private-equity buyouts. They have that expertise and opportunity, and most people don’t, so they go around making loans to those buyouts.

And, stereotypically, those are buy-and-hold loans: The private credit lenders give the companies the money, and the companies pay them back over time. The loans do not trade in a secondary market, in part for relationship reasons (the companies and sponsors want to be able to deal directly with their original lenders if they run into trouble) and in part perhaps for cynical return-smoothing reasons (the lenders would prefer not to mark down their loans just because they trade at a low price in the secondary market) but also in part because, you know, who would want them? These are stereotypically loans to small-to-medium-sized private companies, and if you are not already a lender, you probably don’t know much about them. It might be hard to do due diligence and understand the companies, and it might not be worth it to do that just to buy a little bit of a smallish loan. 

But increasingly the words “private credit” mean AI financing: When AI firms borrow infinity zillion dollars to build data centers, they will sometimes do so in the bond market, but for structuring-flexibility reasons, they will often do it from private lenders. And when they do that, everyone in the financial industry, and possibly everyone in the world, will think a thought like “ooh AI, I would like to get in on that deal” or else “ooh that’s stupid, I would like to short that deal.” Everyone knows about this stuff, and there is a diversity and intensity of opinions.

A diversity and intensity of opinions makes a market, and so Bloomberg’s Caleb Mutua and Isabelle Lee report:

Artificial intelligence spending and the growth of the private credit market aren’t just spurring companies to borrow more, they’re also helping to generate fresh records for corporate-bond trading. …

Some of the borrowing for those investments is happening in private markets, as when Meta Platforms Inc. and Blue Owl Capital Inc. raised about $27 billion of high-grade debt for a data center in rural Louisiana last year. That in turn can spur more trading in private credit, where investors are increasingly looking for ways to sell out of positions, said Rehan Latif, global head of credit trading at Morgan Stanley.

“I view it very much as the biggest single opportunity coming into 2026,” said Latif in an interview. “Every single time a new market is created, there is a little bit of a lag before the secondary market kicks off. The reality is this is the right time for it to happen.”

Right yes if private credit is “junk-rated middle-market buyout loans,” then it will probably end up with liquid secondary markets, because in the long run everything ends up with liquid secondary markets. But if private credit is “$27 billion of investment-grade loans to Meta,” then it will definitely end up with liquid secondary markets pretty soon, because everybody wants that, or wants to bet against it.

It is a tradition around here that, when I go on vacation, Elon Musk does some stuff. Last month, I went on vacation, and Musk immediately went and got himself $140 billion:

Elon Musk won reinstatement of his 2018 pay package as chief executive of Tesla Inc., after the Delaware Supreme Court reversed some findings of a judge who said the billionaire improperly influenced board members who formulated the record-breaking corporate compensation package.

The high court [on Dec. 19] concluded the world’s richest person is entitled to a stock-based compensation plan now valued at about $140 billion. When Tesla directors first authorized the payout, it was the biggest ever for a US executive. It’s since been eclipsed by a separate plan that could be worth an additional $1 trillion for the Tesla chief executive officer if he hits future performance targets. 

Musk’s 2018 compensation plan has been on hold after a single investor — who owned nine shares — successfully sued to block it in Delaware, where the electric carmaker was incorporated at the time. Tesla investors twice overwhelmingly voted to back the plan, which surged in value as the stock soared close to $500 a share from around $20 seven years ago.

In a unanimous ruling, the five justices on Delaware’s highest court said canceling the CEO’s entire payout improperly left “Musk uncompensated for his time and efforts over a period of six years.” However, they upheld the finding of a lower court judge that Tesla’s board was riven with conflicts of interest when it established the pay package for the company’s billionaire co-founder.

Here is the Delaware Supreme Court’s opinion, which is a bit unsatisfying theoretically but very practical. Basically:

1. Sure, yes, obviously Tesla’s board was and is in Musk’s pocket, and there were conflicts of interest when it gave him his giant slug of options in 2018. Because of those conflicts, Tesla had to demonstrate that the grant of options was “entirely fair” to minority shareholders, and it failed to do that.
2. But ehhhh what are you gonna do about it? Take back the options? Seven years later? That’s hardly fair either.

This strikes me as basically a correct analysis. As I wrote two years ago, when the Delaware Chancery Court struck down the pay package, “there is something awkward about a judge second-guessing the CEO compensation decisions of the board of directors of a $600 billion company,” particularly when shareholders (twice!) approved those decisions.

It is also in its way a very Elon Musk analysis. Of course Tesla did not follow every technical nicety of corporate law or fiduciary duty; that’s not how Elon Musk operates. Of course the transaction was conflicted and the board gave him an unprecedented and probably unnecessary pile of money. But (1) he’s Elon Musk, (2) it’s Tesla and (3) it all kind of worked out: Tesla really did create a trillion dollars of value for the shareholders, so they have no cause to complain and almost none of them did. (The plaintiff owned nine shares!) So, you know, no harm, whatever. 

A few years ago, Musk decided to merge Tesla with SolarCity, a solar power company that he controlled. Shareholders sued, saying that the transaction was obviously conflicted and Musk was just bailing out his SolarCity investment with Tesla shareholders’ money. A Delaware judge criticized Tesla’s process, but nonetheless rejected the lawsuit. I wrote:

Those are the two sides of Elon Musk, aren’t they? On the one hand he infuriatingly refuses to comply with the basic expectations of law and society; on the other hand Tesla’s stock keeps going up. ... He did a deal that was rotten with conflicts of interest, and while he could have followed the proper procedures to do the deal right, he didn’t, because he is too impetuous and doesn’t really believe in rules or social norms. But also the deal worked out great and confirmed his reputation as a business genius. And so the Delaware judge spends some time criticizing him for doing the deal wrong, but his heart isn’t in it. Musk is too much of a lovable business scamp for any judge to stay mad at him. He does all the wrong things but they keep working out right.

I feel like in the intervening five years he has become a less lovable business scamp, but nonetheless that analysis holds up. [2] Tesla did not legally give him those options, or not quite, but still he gets to keep them. That’s how this works.

Elsewhere in Elon Musk, I have to say that I grudgingly respect that his approach to life seems to be essentially aesthetic. His aesthetic is, it seems to me, very bad, but sincerely held. Here is a Wall Street Journal article about him that drove me somewhat insane:

For the longest time, Musk has talked about the possibility that we’re living in a computer simulation. …

“I do have this theory about predicting the future, which is that the most interesting outcome is the most likely,” Musk said during a podcast appearance a few weeks ago. 

It is an idea rooted in the thinking that if our reality really is a simulation, we would get shut down if we were boring. 

“Another way to think of it is like we could be an alien Netflix series and that series is only going to get continued if our ratings are good,” Musk said. “If you apply Darwin to simulation theory then only the most interesting simulations will continue. Therefore, the most interesting outcome is most likely because it’s either that or annihilation. 

“So,” he added, “really, we have one goal: Keep it interesting.” 

This requires a real theory of mind about the hypothetical aliens who are running the hypothetical simulation. Are Musk’s various social-media fights and other adventures interesting? I find them pretty exhausting and try to avoid them, even as he tweaks X’s algorithms to show them to everyone. But I cannot deny that, you know, I keep writing about them. He obviously finds them riveting. What do the aliens think? I mean. The world continues to exist? That is consistent with (1) the world is a simulation, (2) it will only continue if ratings are good and (3) Elon Musk’s personal contributions to the interestingness of the world are good for ratings. It is also consistent with other hypotheses. 

Elsewhere:

The future of Tesla is an army of humanoid robots that Elon Musk says could eliminate poverty and the need for work. He has told investors the robots could generate “infinite” revenue for Tesla and have potential to be “the biggest product of all time.”

Musk has bet the company and his personal fortune on this vision of the world in which Optimus, as it is known, works in factories, handles domestic chores, performs surgeries and travels to Mars to help humans colonize the planet. Though today each robot is made by hand, Musk has proposed manufacturing millions of robots a year.

Okay. And:

Tesla took the unusual step of publishing a series of sales estimates indicating the outlook for its vehicle deliveries may be lower than many investors were expecting.

The carmaker posted estimates to its website showing analysts on average expect the company to deliver 422,850 cars in the fourth quarter, down 15% from a year earlier. That compares with a Bloomberg-compiled average of 440,907 vehicles, an 11% drop.

And:

Elon Musk’s Grok is facing mounting criticism and threats of government action around the world after the artificial intelligence chatbot created sexualized images, including of minors, on the social media platform X in response to user prompts.

See, if I were getting paid a trillion dollars to run a car company, I would try to (1) sell more cars and (2) not build a chatbot that creates sexualized images of minors, because I would think things like “I want a trillion dollars” and “I do not want to be fired from my lucrative job” and “I do not want to get arrested.” But Musk is optimizing for the fate of humanity, and selling more cars is boring. At least sell robots! 

The stock of Chevron Corp. closed at $155.90 per share on Friday. As of noon today, it was trading at about $164.36, up about 5%. There are call options on Chevron’s stock with a $160 strike price and an expiry of Jan. 16. About 1,346 of those contracts traded on Friday, with a reported closing price of $0.65 per share. (A contract is 100 shares, so that’s about $87,500 of trading volume.) At around noon today, I saw those contracts at $5.51, with more than 8,000 contracts in volume. If you had spent $87,500 on Chevron call options on Friday — as buyers collectively did — you’d have about $740,000 today. 

Why might you have done that? Well. You might have had a joint hypothesis like: “Nicolás Maduro will be removed as leader of Venezuela, and his replacement will be more open to US energy companies extracting Venezuela’s oil reserves, and that will be good for Chevron, which ‘remained in Venezuela after the seizure of foreign oil assets at the turn of the century’ and ‘is best positioned among global oil giants to immediately benefit from greater US access to the world’s largest crude reserves,’ and therefore Chevron’s stock will go up, and this will all happen before the options expire on Jan. 16.” You could have made a highly levered joint bet on (1) Venezuelan regime change, (2) its economic implications and (3) its timing.

And if you had done that you would have been correct. [3] Bloomberg’s Mitchell Ferman reports:

US oil stocks jumped Monday after President Donald Trump pledged to revive the Venezuelan energy sector following the capture of Nicolás Maduro over the weekend.

Chevron Corp., the only American oil major currently operating in the South American nation under special US permission, surged as much as 6.3%, the most since April.

ConocoPhillips and Exxon Mobil Corp. also rose. The three largest oil-service companies — Halliburton Co., SLB Ltd and Baker Hughes Co. — all jumped more than 5%.

Trump said US oil companies will spend billions of dollars to rebuild Venezuela’s crumbling energy infrastructure and restore the nation’s oil sector to its former glory.

So not just Chevron. “American military adventurism will be good for Halliburton” is a thesis with a decades-old pedigree, and some more general view like “regime change in Venezuela will be good for oilfield services” seems pretty straightforward. 

There are other theses. Venezuela has been in default on its sovereign bonds for a while, and it is plausible to think that regime change might lead to restructuring of those bonds and some recovery for the bondholders. And so “the country’s dollar bonds rose by between 7 and 9 cents on the dollar across maturities in thin London trading on Monday.”

Or, of course, you could have had some wrong joint hypothesis. “Regime change in Venezuela will bring more of its huge reserves into the global market and drive down the price of oil”: Not a crazy thing to think, but in fact oil is up a bit today and there is a lot of uncertainty about the timing and impact of any changes in Venezuela.

And so if Donald Trump had come to you on Friday and said “hey, we’re going to grab Maduro this weekend, don’t tell anyone,” you would have had material nonpublic information, but it would have been up to you to know how to use it. You could have correctly predicted the economic implications of that information, and made a lot of money, or you could have incorrectly predicted its implications, and lost a lot of money. (Also I am ignoring the legality of all of this!)

But that would require both (1) the inside information and (2) some ability to develop a model of its economic implications. What if you don’t have that, or don’t want to bother? Well, there were betting markets just on whether Maduro would be deposed. You could have made that simple bet. Somebody did. Axios reported:

Traders on Polymarket appeared to anticipate Maduro's capture late Friday night, before President Trump announced it early Saturday morning.

The market for whether or not Maduro would be out of power climbed shortly before 10pm ET on Friday, after hovering in the low single digits for weeks, per the Wall Street Journal.

What appears to be a newly created account appeared to invest $30,000 Friday in Maduro's exit. After Maduro went into custody Saturday morning, that same investor netted $436,759.61.

Was this based inside information? I have no idea. Byrne Hobart writes:

[I previously wrote] that such a market would barely function, because market-makers would assume that any trader in those markets had information the market-maker didn’t. And yet, $30k of liquidity is a lot more than I’d expect to get on such an asymmetric bet. What that piece didn’t anticipate was just how big prediction markets and recreational gambling would get. There are enough punters on geopolitical markets that liquidity provision can still work, and even if there aren’t professionals involved in every market, those same recreational traders are still providing liquidity. 

On the one hand, the point here is that there are enough recreational gamblers on Polymarket that, if you have inside information about a CIA-backed coup in Venezuela, you can bet $30,000 on it at 14-to-1 odds. On the other hand, the point here is that there are a lot of recreational gamblers on Polymarket, so it’s always possible that the winning bettor here was one of them: Perhaps someone just felt like betting some money on Maduro’s exit without inside information, because that is the sort of gambling that people do now.

My basic view is that there is some natural limit on the liquidity here. Markets with economic consequences — the market for oil, or for oilfield services stocks, or for Venezuelan sovereign bonds — are bigger and deeper than markets for recreational gambling. People have diverse reasons for buying or selling oil or stocks or bonds, and those reasons make a market, and if your reason for buying or selling those things is “I have inside information about a coup,” then someone will take the other side of your trade, in large size, for unrelated reasons of their own. Whereas the market for pure bets on a coup is mostly limited to (1) recreational gamblers and (2) insiders.

Anyway Axios goes on:

Rep. Ritchie Torres (D-N.Y.) will reportedly introduce a new bill — the Public Integrity in Financial Prediction Markets Act of 2026 — to limit federal elected officials and some political figures from engaging in prediction markets.

It does seem bad for government officials to have incentives to do unexpected things to make themselves a quick profit, though maybe Elon Musk would disagree.

An important problem with being a dedicated short seller is that stocks mostly go up, so you are fighting against the current. If you are fabulously good at finding bad companies, perhaps your short portfolio will underperform the market by 10% per year. The S&P 500 index was up 16.4% last year. If you had bet $100 against the bad companies you identified, you’d have $94 now. There are other, more favorable ways to monetize this skill, but it is a hard business.

Buying stocks that go up is an easier business, because stocks mostly go up. If you are very good at finding bad companies, perhaps you can apply that skill to buying stocks that go up? “Buy all the stocks, except the bad ones” is one plausible (and common) approach. Also, though, maybe “investigate a bunch of stocks to see if they’re frauds, and buy the ones that are the least fraudulent”? Or, for that matter: “Be an activist short seller, identify bad companies, short their stocks, publish withering short reports about them, develop a reputation as a vicious short seller, and then occasionally find good companies, buy their stocks, publish glowing reports about them, and have everyone think ‘oh wow if this vicious short seller likes the stock it must really be good.’”

Anyway the Wall Street Journal reports:

Hedge fund Hunterbrook Capital launched less than two years ago with $100 million and what it called a “news-driven strategy.” An affiliated team of journalists would publish exposés on corporate wrongdoing, and the fund would short the stocks of their targets ahead of time.

In 2025, it learned that good news can be just as profitable as bad. 

Hunterbrook Capital gained 23% after fees last year through September, according to an investor letter viewed by The Wall Street Journal. That was higher than indexes tracking hedge-fund performance, as well as the roughly 15% return of the S&P 500 over the same period. Many of its best-performing positions were bullish bets on stocks based on reporting from its sister newsroom, Hunterbrook Media.

Take flying-taxi maker Joby Aviation. The hedge fund had previously bet against similar companies it thought were overhyped and potentially fraudulent. Then, a Hunterbrook Media reporter discovered an abnormally long Joby flight appear, then disappear, from public flight logs.

He booked a room with a view of the hangar and photographed a new aircraft Joby was testing. Hunterbrook Capital bought Joby shares before the scoop was published, betting it would make Joby bears have to cover their shorts, and sold them a few weeks later after a rally.

“We naturally assumed that the flying taxi company was a fraud, but then we watched its taxi fly so we went long the stock,” reasonable. Anyway Hunterbrook’s whole schtick is that they are not just a hedge fund, and not just an activist short seller, but rather a real independent media company that publishes stories rather than stock reports. I suppose they could get into publishing glowing puff pieces to see if that moves any stocks.

My general rule around here is that you should not be involved in any sort of corporate rap video. A case that I had never considered, though, is: What if you work for a cannabis company? On the one hand, no, corporate rap is bad. On the other hand, I feel like, historically, selling marijuana and rapping are often complementary skills. If you tell me “I sell mortgages and rap,” I will assume you are a bad rapper. If you tell me “I sell marijuana and rap,” I’m gonna give you a chance.

But then, I am old, and when I started listening to rap selling marijuana was illegal. Now you can sell cannabis and be very corporate indeed. So here’s the chief executive officer of a European cannabis conference company rapping. Is he Snoop Dogg? He is not. But amid the corporate jargon you get the phrase “a medley of super heavy nugs.” The video is also great. Everyone is wearing a lanyard.

Incidentally I have previously written a few times about a corporate rap video produced by an opioid company, which was a terrible idea, but I suppose technically the same loophole applies. “I rap and sell opioids,” okay, maybe.

Bridgewater, D.E. Shaw Among Top Hedge Fund Gainers of 2025. Citadel’s Flagship Hedge Fund Climbed 10.2% Last Year. Japan’s ‘ Dementia Money’ Is a Warning to the World. Deutsche Bank shares exceed book value for first time since 2008. The upstart exchange drawing traders to the world’s best-performing stock market. Small-Time Crypto Investors Are Facing Violent Attacks. “Households with at least one GLP-1 user counterintuitively spend more on eating out than they did before family members started the drug.” US Auto Sales Poised to Slip as Middle-Class Buyers Retreat. “He rated a new roll with spicy tuna he recently tried a ‘negative million.’” America Is Falling Out of Love With Pizza. A Booming Live Music Industry Looks for Its Next Generation of Roadies. “The secondary market for bankers is highly illiquid.”

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!