Money Stuff: OpenAI Is Good at Deals

The basic situation is that if OpenAI announces a big partnership with a public company, that company’s stock will go up. Today OpenAI announced a deal to buy tens of billions of dollars of chips from Advanced Micro Devices Inc., and AMD’s stock went up. As of noon today, AMD’s stock was at $213 per share, up about 29% from Friday’s close; it had added about $78 billion of market capitalization.

How do those negotiations go? Like, schematically:

OpenAI: We would like six gigawatts worth of your chips to do inference.

AMD: Terrific. That will be $78 billion. How would you like to pay?

OpenAI: Well, we were thinking that we would announce the deal, and that would add $78 billion to the value of your company, which should cover it.

AMD: …

OpenAI: … 

AMD: No I’m pretty sure you have to pay for the chips.

OpenAI: Why?

AMD: I dunno, just seems wrong not to.

OpenAI: Okay. Why don’t we pay you cash for the value of the chips, and you give us back stock, and when we announce the deal the stock will go up and we’ll get our $78 billion back.

AMD: Yeah I guess that works though I feel like we should get some of the value?

OpenAI: Okay you can have half. You give us stock worth like $35 billion and you keep the rest.

I don’t know. This is not, you know, correct in any respect. (In particular, the total cost of the deal was not disclosed, “but AMD said it costs tens of billions of dollars per gigawatt of computing capacity”; I’m using the probably-too-low $78 billion number for narrative convenience.) But there is some crude justice to it. This deal between OpenAI and AMD was obviously going to create a lot of stock-market value: The announcement of the deal would predictably increase the market value of AMD, and it’s not like it decreases the market value of OpenAI commensurately. [1] Why not use that value to subsidize the deal? Schematically, OpenAI could buy AMD stock to predictably profit from the stock-price bump it created. Just going out and doing that in the market would be awkward — it might look like insider trading — but buying the stock from AMD is fine. So that’s what they did:

OpenAI will receive warrants for up to 160 million AMD shares, roughly 10% of the chip company, at 1 cent per share, awarded in phases, if OpenAI hits certain milestones for deployment. AMD’s stock price also has to increase for the warrants to be exercised.

The warrants vest based on operational and stock-price milestones (some of them require the stock to hit $600 per share), but 160 million shares times the $213 price at noon today is about $34 billion. In rough numbers, OpenAI is getting back half of the value it created for AMD.

I have to say that if I was able to create tens of billions of dollars of stock market value just by announcing deals, and then capture a lot of that value for myself, I would do that, and to the exclusion of most other activities. OpenAI can do that, and does, though not to the exclusion of everything else. Bloomberg’s Ryan Vlastelica writes:

OpenAI may not be publicly traded, but the world’s most valuable startup is increasingly making waves in the stock market.

Just last week, OpenAI sent shares of e-commerce companies Shopify Inc. and Etsy Inc. soaring after unveiling an instant buy option in ChatGPT. Then a blog post detailing new features the company is using internally sent a fresh wave of jitters through software stocks like Atlassian Corp., already reeling from fears about AI disruption.

On Monday, it inked a deal with Advanced Micro Devices Inc. that could generate tens of billions of dollars in new revenue for the chipmaker. AMD shares soared 38%, its biggest intraday percentage gain since April 2016, and the news sent shockwaves through the rest of the industry. Nvidia Corp. and Broadcom Inc. both slid.

I am always impressed when tech people with this ability to move markets get any tech work done. OpenAI still seems to be building a lot of AI products; if I were them I’d just be day-trading Shopify stock.

The essential idea is that, if you find something bad at a public company, you can turn that into money. There are at least five broad categories of ways to turn bad stuff into money:

1. Business. Like, if you find that the leading car company makes cars that explode, you could build a car that doesn’t explode, and advertise it by saying “our cars don’t explode, unlike the other guy’s,” and people would buy it and you would make money. (Or if you find that the leading sneaker company makes sneakers that are awesome and cheap but made by workers in horrible conditions, you could make somewhat more expensive sneakers made by workers whom you treat well, and try to appeal to consumers who care about this difference.) This is, in the broad sweep of capitalism, the main way to profit from other companies’ badness, and surely the most socially beneficial. [2] You see a company doing bad stuff and you try to do better stuff.
2. Finance. If you find out that a public company is doing bad stuff that nobody knows about, you can short its stock. Then, when the bad stuff comes out — probably because you publish it — the stock will go down and you will make a lot of money. There are variations on this. (If you find out that a company has breached its credit agreements, you can buy its debt and buy even more credit-default swaps, etc.)
3. Law. In the US in the 21st century, this one is really really important: If you find out that a public company is doing bad stuff, you can sue. You can sue it for the bad stuff. You can sue it for securities fraud: As I often say around here, every bad thing that a public company does is also securities fraud, and the damages are big. (This ties nicely to short selling: You find out the bad thing, you short the stock, you announce the bad thing, the stock drops, you take profits on your short, and you organize a shareholder class action to sue for fraud.) Or you can bring the case to the government, which might sue the company and give you a cut (as a whistleblower or qui tam relator). (Again this ties to short selling: If the government goes after the company, the stock will drop.)
4. Government. In the US in 2025 specifically, there is a variation on Path 3 of the form “if you’ve got the ear of Donald Trump, you can make him go after a company that does something he dislikes, and probably turn that into money for him at least and maybe for you.” Somehow TikTok being a vehicle for Chinese propaganda turned out to be very good for Larry Ellison? 
5. Media. People love stories about bad stuff! If you have a juicy story about bad stuff, you can put it in your magazine or television broadcast or podcast or newsletter, and people will like it and you can sell ads or subscriptions.

There are probably others, but those seem like the main ones. The key thing to notice is that you don’t have to choose. If you have bad stuff, you can monetize it as many ways as possible. [3] You can be a short-selling hedge fund and also file whistleblower complaints. Or, like Hunterbrook, you can be a hedge fund that’s also a newspaper: You can publish stories (media! [4] ), short stocks, lead class-action lawsuits and file whistleblower complaints. Use every part of the bad news.

This seems like a big enough niche that there’s room for more than one “hedge fund that is also a newspaper,” for different audiences or political persuasions. Here’s “The Antifraud Company”:

The Government Accountability Office estimates that American taxpayers lose $500 billion a year to fraud which is almost $1,500 per citizen. Our thesis is that much of this fraud is organized lawbreaking by giant corporations. We think that's unacceptable.

That's why we raised over $5 million from Abstract Ventures, Browder Capital, and Dune Ventures to start The Antifraud Company.

The Antifraud Company operates as a new age journalism company, using the power of AI-enabled forensics, rigorous investigation, and public awareness to help arrest unlawful corporate conduct.

The Antifraud Company plans to investigate and publicize corporate fraud in critical government programs. Our business model is to additionally monetize our investigative work through a variety of government whistleblower programs.

Here’s their launch video, where they call themselves “a private-sector DOGE” and boast about how many congressional investigations they’ve already sparked. Honestly the video is very good and they’re going to go far. 

Speaking of Hunterbrook. One point that I sometimes make around here is that there is a continuum between “journalism” and “insider trading.” One thing that you might do in life is talk to people inside companies and try to get them to reveal secrets about those companies that they are not supposed to tell you. If they tell you a particularly juicy secret — a scoop, something like “we are about to announce a merger” or “we discovered a cure for cancer” or “we are doing a huge fraud” — then you might try to use that secret to advance your career.

If you are an investigative journalist, you can advance your career by publishing your scoop. This will be good for your employer (it will get more subscription or advertising revenue) and thus for you (you will get a congratulatory Slack from your editor). If, on the other hand, you are a hedge fund analyst, you can advance your career by trading on your scoop: You buy the company’s stock if the scoop is good, or short it if it’s bad. This will be good for your employer (it will make money on the trade and thus get more performance-fee revenue) and thus for you (you will get a big bonus). The difference is that, if you do this as a hedge fund analyst, you might go to prison for insider trading, while if you do it as a journalist you probably won’t.

If you are very skilled at getting people inside companies to talk to you and give you scoops, what career should you pursue? Well, you could be a journalist: You’d be good at it, but nobody goes into journalism in 2025 to get rich. Or you could be a hedge fund analyst, which is a better way to get rich, but if you use this skill set too much at a hedge fund you will go to prison and that’s not great.

Is there a way to get the best of both worlds? Conceptually what you want is a private newspaper for hedge funds. Like, you go to a hedge fund and say “hello, I can reliably get information for you from inside public companies. But if I did that on your payroll, or even as a consultant, it would be insider trading and you and I might both go to prison. If I did that at a newspaper, it would be investigative reporting, but you’d get my scoops at the same time as a lot of other people and you couldn’t make much money on them. But I’m going to do it at a special newspaper with very few subscribers and a very high subscription price. If you pay me $1 million a year, I’ll send my scoops to you and four other hedge funds, and you can trade on them, but it won’t be insider trading because it’s journalism instead.”

This is absolutely not legal advice in any way, but clearly something like this works. Lots of news organizations (disclosure, including Bloomberg News) provide information to paying subscribers, who can then trade on it because it came to them as a journalistic scoop rather than an insider tip. I suppose one could do some sort of line-drawing exercise: News disclosed to five subscribers who pay $1 million a year might look like a tip, while news disclosed to five million subscribers who pay $100 a year is surely journalism. But I do not know where the line is, or why you might draw it in any particular place.

We’ve talked about this a few times before, and I once guessed at where the line might be:

If you find stuff out and email it to five hedge funds, that feels like you are a “consultant” and it’s insider trading. If you email it to 100 newsletter subscribers, 90 of whom are hedge funds, you’re probably okay. Not legal advice!

I then went on to talk about the leading case, which involved Reorg Research. Reorg is a “corporate intelligence company” covering distressed debt, and there was a dispute (involving disclosing sources, not insider trading) about whether it was in the business of journalism or the business of, you know, business. Many journalistic organizations (including Bloomberg LP) supported Reorg, and a New York court found that Reorg was doing journalism and could use a journalism shield law to protect its sources. I wrote:

At the time, it had “about 375 subscribers who ‘manage trillions of dollars in assets’” and paid “between $30,000 and $120,000 a year” for subscriptions. So the line between journalism and insider trading is probably lower than 375 subscribers.

That was a couple of years ago, and Reorg is now called Octus. (“The leading voice in credit intelligence. Octus delivers proprietary data and actionable intelligence.”)

Recently, in the course of her reporting, a Reorg correspondent got a good scoop: “London-based private equity firm Investindustrial is preparing an approximately $3 billion bid to take US food processing company TreeHouse Foods private, said sources.” That’s public now; she posted it on her LinkedIn last week, and TreeHouse’s stock shot up when she did. But she also posted:

Octus subscribers got it first over a dinner, maybe while dining at Eataly, majority owned by Investindustrial.

If you go to dinner at Eataly with a handful of your hedge fund friends, and someone at the dinner is like “I talked to someone at Investindustrial and they’re working on a bid to take TreeHouse Foods private,” and you go trade on that news, is that insider trading? Not always!

We talked a few months ago about “multiverse finance.” This is a prediction-market-inflected idea from Dave White at Paradigm about trading financial assets contingent on certain events. So if you have a thesis like “interest rates will go down if the Democrats win the midterms,” you could buy a contract that (1) resolves into $1 million of Treasury bonds if the Democrats win and (2) resolves into $0 if the Democrats lose. The prices of ordinary financial instruments reflect entangled market-implied probabilities and expected impacts of many possible real-world events; the prices of prediction markets isolate market-implied probabilities of particular events. Multiverse finance could combine these things so you can get “asset prices conditional on one particular event I care about.”

I don’t know how practical that is but it’s fun to think about. Of course, “prediction markets” in 2025 mostly means sports gambling. There are not a lot of financial assets whose value is contingent, in any obvious way, on some particular team winning some particular game. But there are some. Tickets to the final match of the World Cup, for instance, will vary in value based on the teams that make the final. More to the point, tickets to see England in the final will be worth a gazillion dollars if England make it to the final, and nothing if they don’t. So if you could buy that — “tickets to the 2026 World Cup Final, conditional on England playing in it” — that would be, uh, cool I guess? Fun to think about? Like:

1. Buying that now should cost you less than buying just an unconditional ticket to the World Cup final match, because the conditional ticket is worthless in the majority of future states of the world (the ones where England don’t make it to the final [5] ). So this could be a way to get reasonably affordable World Cup final tickets.
2. You might want exactly that conditional ticket: You might really want to see England in the final, and not care about seeing the final if it’s anyone else.
3. Let’s face it, you might enjoy sports gambling, and the conditional ticket nicely combines (1) an actual consumption good (the ticket) with (2) a sports gamble (you only get it if your team wins).

Apparently this is how the actual 2026 World Cup ticketing system works, because all of sports has become 100% about gambling. Also because it’s 2025 this gambling takes place on the blockchain sure sure sure sure sure sure sure. Bloomberg’s Emily Nicolle and Giles Turner describe the system:

FIFA last year started selling so-called Right to Buy tokens, providing buyers the right to instead purchase guaranteed entry to a World Cup game. ...

Some of these tokens are tied to a national squad — but come with greater risk. The Right to Final token, for example, offers fans an opportunity to purchase a ticket to the World Cup final match, but only if the team whose collectible they’ve bought makes it that far. ...

The Right to Final: England token cost $999, and those tokens are now sold out. If England makes the final, those token holders will be entitled to a ticket to the match, but it’s not yet clear how much more they will have to pay to secure an actual ticket.

It’s not even a contingent ticket, just a contingent right to buy a ticket. Anyway the point of that story is that Swiss regulators are investigating whether this is gambling:

Switzerland’s gambling regulator has opened a preliminary probe into sales of blockchain-based tokens backed by FIFA which fans can exchange for tickets to games at the 2026 World Cup.

Gespa, which also oversees lotteries and sports betting, is examining whether FIFA is selling the potential to win tickets to a sporting event, or offering something closer to gambling.

Of course it’s gambling! And it’s also something else! Like everything these days.

Elsewhere, Bloomberg’s Ira Boudway interviewed Robinhood Markets Inc.’s head of sports gambling. Wait, sorry, no, that’s, “JB Mackenzie, the retail trading platform’s vice president and general manager of futures and international,” whose “job includes overseeing Robinhood’s event contracts business.” Mackenzie explains that sports gambling and the stock market are basically the same thing when you think about it:

I don’t know if customers define them as entertainment or not. You have people that are just staunch believers in companies. You’ve got people who are Tesla bulls. They believe in Tesla. With these prediction markets, on the sports side, it’s just a slight flip because you already have that affinity because you were a Jets fan with your dad.

Yep! You can be a Jets fan and bet on the Jets, or you can be a Tesla fan and bet on Tesla’s stock, what’s the difference really. I tend to think that capital markets have some purposes outside of gambling and fandom, but I recognize that that is an old-fashioned view.

Sorry this is just good comedy:

US House Republicans have told Securities and Exchange Commission Chair Paul Atkins that they are investigating the loss of text messages from former SEC Chair Gary Gensler from when he led the agency.

The SEC’s Office of Inspector General’s findings in early September cast doubt on whether the Gensler-led SEC acted with transparency and integrity while serving between 2021 and 2025, House Financial Services Committee Chairman French Hill said in a letter to Atkins on Tuesday. …

The letter, also signed by House Ranking Members Ann Wagner, Dan Meuser, and Bryan Steil, said that Gensler sued several financial firms for “widespread record-keeping failures,” collecting more than $400 million worth of fines to settle charges alone in 2023. 

The deleted text messages highlight a clear double standard, the House Republicans claimed.

They are not quite right about that, though. As we discussed a lot over the previous four years, one of the Gensler SEC’s top enforcement priorities was fining financial firms whose employees used personal cell phones to text about work. The theory was that SEC regulations require banks, brokers, etc. to retain copies of their business communications. Most business communications these days are electronic, so firms try to comply with that rule by keeping archives of employee emails, Bloomberg messages, etc., for the requisite amount of time. But sometimes bank employees would send business texts from their personal phones or by WhatsApp: They’d text colleagues to discuss strategy or to plot fraud; they’d text clients to provide market updates or to say “I’m stuck in traffic but I’ll be at the meeting in 10 minutes.” The Gensler SEC took the view that these texts were all business communications subject to the record-keeping rules, but banks did not (at the time) reliably keep copies of all their employees’ personal texts, so they were breaking the rules. “Off-channel communications,” as the SEC called them, were forbidden.

I thought this theory was kind of nuts, but I cannot deny that it brought a lot of money into the SEC, so, sure. And now har har har it turns out that Gensler was texting about SEC enforcement from his personal phone. Except he wasn’t! He was texting from his SEC-issued work phone, which did keep copies of his communications. The problem is that the SEC accidentally deleted them:

The SEC Office of Inspector General (OIG) investigated how nearly a year’s worth of text messages from Gary Gensler were permanently lost between October 2022 and September 2023, during the height of the agency’s crypto enforcement action campaign.

In a report released on Wednesday, the OIG revealed that the SEC’s IT department “implemented a poorly understood and automated policy that caused an enterprise wipe of Gensler’s government-issued mobile device,” which deleted stored text messages and operating system logs. 

The loss was worsened by poor change management, lack of proper backups, ignored system alerts and unaddressed vendor software flaws.

Honestly much worse! Imagine the fines if the Gensler SEC had inspected a bank and the bank was like “well, good news, nobody at our bank has ever texted about work from a personal phone; they all use their work email like they’re supposed to; bad news is we deleted all the emails.”

I am a puzzle hunt guy, and the premier New York financial industry puzzle hunt is Midnight Madness (formerly Compass), and it was this weekend. Sadly I did not participate this year, though we did have the organizers on the Money Stuff podcast a few months ago. Anyway congratulations to this year’s winning team, Regenwolk 3000, as well as the runners-up Citadel Insecurities (lol) and Midnight Marauders.

Fifth Third to Buy Comerica in Year’s Biggest US Bank Deal. Tricolor Records Show Same Cars Tied to Thousands of Loans. If Trump’s Tariffs Are Ruled Illegal, Businesses Expect Refund Chaos. Hedge Funds Targeting Fire Insurance Hit a Wall in California. Why Insurers Are Taking Your Money to the Cayman Islands. ‘Gold-plated Fomo’ powers bullion’s record-breaking rally. How China Secretly Pays Iran for Oil and Avoids U.S. Sanctions. Kremlin-backed crypto coin moves $6bn despite US sanctions. Ferrari-Loving Trader Burned Wall Street With Bond Bet Leveraged 11,000-to-1. Centerview Partners to face trial over junior banker’s long hours. Why America’s Fridges Are Overflowing With Sauce. Bill Ackman Says He’s ‘Uncancelable.’ Treasury Department Has Plans To Mint Dollar Coin Featuring Donald Trump’s Likeness.

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!