It does seem like the preferred way to raise money for an idea in artificial intelligence is with institutional venture capital. Some once quite far-fetched AI ideas have paid off richly and rapidly in recent years, and there’s a lot of venture capital looking for more. We’ve talked about Mira Murati’s AI startup, Thinking Machines, which apparently raised $2 billion from top venture firms without telling them anything about its product or business plan. It had people with good AI credentials, it was doing something in AI, good enough!

Probably if you’re one or two tiers down from Murati, in terms of your AI pedigree, you can still get a pretty good deal from good VCs. Not everyone can, though. Presumably there are people with (1) decent AI credentials and (2) a desire to raise a pot of money to do something in AI, who nonetheless can’t get $1 billion or even $100 million from VCs even in this hot market.

There are other sources of money? For instance, retail investors in the public market seem interested in AI. You can’t necessarily start a company, raise no venture capital, and do an immediate initial public offering with a proposal like “ehhhh gonna do some AI stuff, trust me.” But there are plenty of existing small US public companies that don’t have much going on. You could take control of one, announce “ehhhh gonna do some AI stuff,” and the stock will pop: Before, the company was mostly an empty shell, or a deli, but now it is a speculative bet on your AI plans and talents. It’s an AI lottery ticket, and a lot of people are in the market for AI lottery tickets.

Then you do an at-the-market stock offering — or a floating-strike convertible — to sell stock, at new AI-inflated prices, to retail investors. Then you have $100 million to do your AI thing. Then you can do your AI thing. Will your AI thing be as good as Murati’s AI thing? I mean. It is tempting to think that the top venture capitalists have better sorting and vetting capacities than microcap retail stock investors do. But everyone’s  moving pretty fast and making lottery-like bets, and what do I know?

Bloomberg’s Zsana Hoskins reports:

Allbirds Inc., the maker of once-viral wool sneakers that suddenly announced a pivot to artificial intelligence earlier this year, has officially renamed itself Smartbird Inc. and is tapping the former leader of a Danish AI infrastructure group as its new chief.

Nadia Carlsten is taking over as chief executive officer, replacing Joe Vernachio who has resigned from the company and the board of directors. 

Smartbird is among the many struggling companies that have sought another lease on life by seizing onto the buzziest trend, AI in this instance. Such name and business switches were legion during the dot-com bubble. Newly branded crypto-treasury companies surged when that was all the rage. And in February, a tiny karaoke company’s stock soared — and triggered a fierce selloff across the logistics industry — when it trumpeted its AI tool for trucking companies.

Going forward, Smartbird plans to focus on the business of supplying AI infrastructure as a service, a field that has exploded in recent years given the computing resource-intensive nature of AI models and tools. Carlsten previously served as CEO of DCAI, short for the Danish Centre for AI Innovation that was established to run Denmark’s first AI supercomputer.

Here is Smartbird’s announcement, which notes that Carlsten has “served as the Vice President of Product at SandboxAQ, a Google spin-off developing AI models and platforms” and “held leadership positions at Amazon Web Services, including as Head of Product for the AWS Center for Quantum Computing,” and that she’s getting restricted stock units worth about $9 million at current prices.

We talked about Allbirds in April, when it announced its AI pivot (and changed its name, briefly, to NewBird AI, before settling on Smartbird). It also announced a $50 million floating-strike convertible, which is essentially a way to sell stock to retail investors to take advantage of a meme- or AI-driven stock rally. That convertible got upsized to $100 million this week, and All/New/Smartbird is also doing a separate at-the-market stock offering. Also, as previously planned, Smartbird sold off its sneaker assets last week, for $40.7 million in cash. Now it is a pure-play AI company. What sort of pure-play AI company? I am sure that is none of my business. Its press release says:

Smartbird delivers dedicated AI infrastructure as a managed service, providing the performance, control, and scalability of ownership without the capital investment and operational complexity. The company is in active discussions with prospective customers across its target verticals and is currently designing its first cluster deployments.

“Smartbird is entering the market at a pivotal moment in the evolution of AI infrastructure,” said Dr. Carlsten, CEO of Smartbird. “AI is rapidly becoming mission-critical for organizations across every industry, yet many organizations lack a practical path to deploy and operate the dedicated infrastructure these workloads require.

There is a clear opportunity to meet the growing need for enterprise-grade AI infrastructure that delivers control and performance without the capital and operational burden of hardware ownership. With a differentiated strategy, significant capital, and the opportunity to build an exceptional team, we are uniquely positioned to capitalize on one of the most significant infrastructure opportunities of the next decade.”

The stock was up about 50% this morning. I don’t know, there is something charming about this. Every day, institutional venture capitalists are throwing millions of dollars at vague AI companies run by former Google AI employees with PhDs. Regular retail investors don’t have access to those startups. But they have access to this! The public stock market, for reasons of its own, has created a version of AI startup investing. That’s neat.

The main modern way to invest is to buy the index, or most of it. Most investors should assume that they’re not smarter than the market, or at least that they’re not much smarter than the market; they have at most specific narrow advantages over the market. Buying all of the stocks gets you the market return. Buying most of the stocks, but adjusting a bit — buying more of a few stocks that you like; avoiding a few stocks that you don’t like — might get you a bit better than the market return. The baseline is owning the market, and then you tinker from there. My Bloomberg Opinion colleague John Authers writes today:

Thirty years ago, a fund manager would say they “owned” a stock. Now, they tend to say they’re “overweight.” Once, if you didn’t like a stock, you simply wouldn’t buy it. With the rise of indexing and benchmarking — judging active funds by comparing them to an index — that no longer follows. Put 5% of your fund into Nvidia Corp., and that’s a big bet on the world’s largest chipmaker. You’d be desperate for the company to succeed. But with Nvidia now 7.9% of the S&P 500, a 5% holding becomes a big “underweight,” and the manager bizarrely would prefer a company they own to do badly. This was labeled the “curse of the benchmarks” by Paul Woolley of the London School of Economics. …

Thus investors move ever more in crowds, out of a basic instinct for self-preservation. Just like wildebeest on the savannah, there is more safety in the middle of the herd than as an outlier at the front or back, where they will be far more vulnerable to predators. That tendency toward herding gnaws away further at market efficiency.

“Quasi-indexing,” you might call this style of investing. But it is not the only way to invest. Another, more old-fashioned way to invest is to buy, like, 20 stocks that you like and think will go up. Your starting point is not owning the market; your starting point is a blank sheet of paper. This has perhaps three advantages over the quasi-indexing style of investing:

1. For the quasi-indexing style to work for everyone else, somebody has to be doing the first-principles work of figuring out what companies are good and how much to pay for their stocks. You start from a blank piece of paper so that all the indexers and quasi-indexers don’t have to.
2. If you are good at this style, it is very restful. You only have to buy your 20 stocks, hold them, and wait for them to go up.
3. If you are good at this style, it is very lucrative. Most of the market’s returns come from just a few stocks. You can just buy those, and not the other ones, and then you’ll significantly outperform the market.

We talk occasionally about professional investors who follow roughly this approach and have nice lives. Bill Ackman at Pershing Square Capital Management has a lot of stuff going on, but the main thing he has going on is buying and holding like a dozen stocks that go up. Chris Hohn at the Children’s Investment Fund is another guy who makes concentrated long-term bets, is a billionaire, and goes around saying stuff like “close your eyes, and in the darkness you can know God and God is consciousness.” Plenty of other hedge fund managers don’t have time to close their eyes; they might miss a crucial market move.

There is a significant selection bias here, though. If you are a professional concentrated buy-and-hold investment manager and you buy the 20 stocks that go up, you will make a lot of money and get on lists and have profiles written about you in which you can muse about the nature of consciousness. If you are a professional concentrated buy-and-hold investment manager and you buy 20 other stocks — ones that go down, or even ones that go up less than the overall market — you mostly won’t.

Those people have to be out there, though. The whole premise of modern investing is that not everyone can pick the 20 best stocks. If all the other, bad, stocks are in the index, it’s because somebody made concentrated bets on them; somebody mistakenly thought they were the good stocks. Today Bloomberg’s Loukia Gyftopoulou profiles one of those people:

Based in Boca Raton, Florida, Polen [Capital] is on no one’s list of industry behemoths. But the nearly half-century-old firm offers a prime example of how missing one big winner can swing fortunes in the AI boom. In just four years, its assets have plunged by 60%, or almost $50 billion. They now total about $33 billion. …

Its six equity mutual funds mostly buy and hold a relatively small number of growth stocks. The flagship Polen Growth Fund owns fewer than 30. Instead of buying Nvidia, the fund stuck with software stocks like Adobe, Salesforce Inc. and ServiceNow Inc.

“We believe virtually all the upside opportunities we can currently see for the company are already priced in,” Polen wrote to clients, referring to Nvidia, in June 2023. The stock went on to soar almost 400% from there. The BVP Nasdaq Emerging Cloud Index, which tracks emerging companies primarily focused on cloud-based software, fell 3% during that period. ...

Today, the flagship fund’s investments are down 45% from the 2021 peak, despite the historic bull market. Its assets have dwindled to less than $2 billion from about $14 billion – a decline of nearly 86%, according to data compiled by Bloomberg. …

The damage is piling up. Clients have bolted, along with a slew of senior staff members, according to interviews with a dozen former employees and people familiar with Polen Capital. Recent departures include the chief operating officer, the chief compliance officer, the head of international finance and the head of client service, according to their LinkedIn profiles. Over the past two years, [chief executive officer Stan] Moss has eliminated roughly 100 jobs, or about half the workforce, some of the people said, asking not to be named discussing private information.

There is a coin-flip alternate universe where the Polen guy is a billionaire celebrity sitting down with an interviewer to be like “lemme tell you my theories about God,” but in this universe he picked the wrong stocks and “declined to be interviewed for this story.” I truly do not think you can hold that against him? I mean, you can; this is not investing advice, and feel free to invest with the people who picked the right stocks.^[1] But there is honor in trying, and the structure of the market does kind of require someone to pick the wrong stocks.

The thing about Sam Bankman-Fried, other than the fraud conviction, is that he’s a nice guy and fun to talk to. So here is a New York Magazine story about what he’s up to in prison, and it is clear that both the writer and Bankman-Fried’s fellow prisoners find him charming, even though he still manages to play videogames while talking to people:

Even in the understimulating, internet-less environment of prison, Bankman-Fried, 34, has found ways to indulge his old propensity for distracted multitasking. For the cost of about $135 at commissary, he bought a SCORE 7T tablet. The name stands for Secure Commitment to Offender Re-entry and Enablement and it bills itself as “a robust, purpose-built, locked-down computing device that helps persons-in-custody (PICs) pass their time constructively, improve themselves and prepare to re-enter society.” … The tablet’s gaming options are “mostly really dumb,” Bankman-Fried says, with the passable exception being Shattered Pixel Dungeon, which he has now played roughly 6,000 times, almost always as the “Mage” and, according to one of his prison friends, with an unheard-of nine “challenges” running simultaneously. Often when we are talking on the phone, Bankman-Fried confesses, he is on his tablet playing the game, just as he was on his computer half-playing Storybook Brawl virtually until the moment he was arrested at his Bahamas apartment complex in December 2022.

In other ways, too, he remains Sam Bankman-Fried:

“There is an undercurrent of people who argue that I should be released from prison in less than 25 years because to do otherwise would be cruel,” he writes, summing up this view as “SBF should be released at an age where he can still settle down to a quiet life somewhere, they say; get married, have kids. Write stories, sometimes, about readjusting to society as an ex-con. Slow down; get off his meds. Detox from a frantic working environment. Be healthy and safe.” To be clear, he adds, “I do think I should be released in less than 25 years!” Just not on those grounds.

Over email, he tells me what is wrong with this line of thinking: “It presumes that my goals in life are to find some semblance of personal happiness and fulfillment, rather than still trying to have positive impact on the world.” Because he maintains he did nothing wrong, Bankman-Fried rejects the idea that the story of his life should follow some standard redemptive arc. Not only does he not wish to exit prison as a changed person, it is more or less his greatest fear.

Speaking of coin-flip alternate universes. When Bankman-Fried’s crypto exchange, FTX, collapsed in 2022, and Bankman-Fried’s cobbled-together balance sheet for FTX came out, I called it “an Excel file full of the howling of ghosts and the shrieking of tortured souls.” Because it was full of crypto tokens that were linked to FTX and that went to zero. You know what else it was full of, though? Anthropic stock that would be worth something like $75 billion today. Basically Bankman-Fried’s problem is that he took a lot of demand deposits from FTX customers and put them into weird related-party tokens and illiquid venture capital investments, which (1) the government argued that he was not allowed to do, (2) he has consistently argued that he was allowed to do, (3) you really shouldn’t do as a matter of prudent financial institution management but (4) worked out great! I mean, not for him. Still, his argument all along was that FTX never had a solvency problem, only a liquidity problem. Obviously everyone always says that, but in hindsight it turned out to be bizarrely true.

Elsewhere: “US senators from both political parties are urging President Donald Trump to reject FTX co-founder Sam Bankman Fried’s pleas for a pardon after the cryptocurrency fraudster applied for clemency last week.”

Hedge funds are pretty good at aligning incentives between their managers and their investors (who are generally called limited partners, or LPs). The manager manages the investors’ money and tries to make it go up; for this, the manager traditionally gets a management fee (traditionally 2% of assets each year) and a performance fee (traditionally 20% of returns each year). This does not create perfect alignment of interests. For one thing, 2% of a large pool of assets is real money, and a hedge fund manager might reasonably think — and I have argued — that her main job is not getting high returns (and earning 20% of the returns) but rather keeping the investors’ money forever (and earning 2% of it, plus some optionality if the returns are good).

But pretty good alignment:

1. Getting 20% of good returns is good.
2. Even if the manager’s goal is just to raise and keep a lot of money and earn 2%, good returns are a good way to raise and keep the money.
3. The hedge fund manager will normally have a lot of her own money in the fund, and her own wealth will often depend more on getting good returns on that money than it does on the fees.
4. The manager is a hedge fund manager; she has a certain self-confidence and risk tolerance. Even if the expected value of being cautious and clipping 2% of a lot of money is higher than the expected value of doing lots of trades to try to earn high returns, she might try to earn high returns anyway, because that is what she is socially conditioned to do. If she wanted to clip management fees on a giant pot of money, she’d be a long-only manager.

On the other hand, sometimes a hedge fund manager will sell shares in her management company — that is, in the stream of fees charged to limited partners — to outside investors. Those outside investors — “GP investors,” you might call them; they invest in the management company, which acts as the general partner in the hedge fund — do not have quite the same incentive alignment as the actual hedge fund manager, in two ways:

1. They don’t necessarily have their own money in the fund. You can be a GP investor in the management company without being an LP investor in the fund.
2. They are not managing the fund themselves; they do not necessarily share all of the manager’s self-confidence and risk tolerance.

Stereotypically, investors in asset management businesses prefer stable recurring management fees over lumpy at-risk incentive fees. Bloomberg’s Hema Parmar, Preeti Singh, and Layan Odeh check in on Bridgewater:

Three years into Nir Bar Dea’s effort to revitalize Bridgewater Associates, the hedge fund is leaner and churning out better performance than it’s had in years. Still, some shareholders are cutting the firm’s valuation and others are pulling out.

Even as the firm’s flagship macro fund was posting a record 2025, two of the firm’s seven institutional owners decided to sell their shares back to Bridgewater at a discount to their purchase price, according to people familiar with the matter. A third, the Teacher Retirement System of Texas, is also looking to sell its stake after slashing the value of its holding by 9% last year.

Bar Dea’s bet has been that a smaller, more focused version of its flagship Pure Alpha — long the biggest hedge fund in the world — would turn around performance that has lagged rivals for a decade. While that’s worked, overseeing less cash also risks reducing the predictable management fees that drive much of the value of ownership stakes. …

“Because Pure Alpha underperformed for stretches while still producing meaningful fee revenue due to its scale, some stakes investors may have been relatively comfortable with the status quo,” said Bruno Schneller, managing partner at multifamily office Erlen Capital Management. “What Nir Bar Dea appears to be doing is effectively forcing a strategic reset.”

Limited partners obviously prefer a smaller fund that achieves higher returns. Employees probably prefer that, because it is more prestigious to work at a good hedge fund than at a large one. But outside shareholders are going to put a higher multiple on steady management fees on a large pot of money than they will on uncertain performance fees on a smaller pot of money.

An extremely good scam is:

1. Get a job as a high-level supervisor in the US Central Intelligence Agency.
2. Run the sort of super-secret spy programs that, for legitimate reasons of national security, require giving you millions of dollars of cash with no explanation and no oversight of how you spend it.
3. Launch another, similar, but fake super-secret program that also requires giving you lots of unsupervised unexplained cash.
4. Keep it.

The Wall Street Journal has a story on David Rush, a CIA official arrested for allegedly taking $40 million of gold bars from the CIA:

He operated a highly classified intelligence program approved by Congress several years ago to use large quantities of cash to obtain critical information about American adversaries, according to people familiar with the matter, and held a rank that is the CIA’s equivalent of an army general. ...

Separate from the real national security project Rush was operating, he appears to have created a fake classified program, known as a special access program, people familiar with the matter said. The fake program was supposedly related to the continuity of government operations, the people said, which generally allow Washington to continue functioning in case of a catastrophic emergency.

Rush allegedly conducted a fake briefing with two co-workers on the supposed program, which he claimed was run jointly with the Pentagon. He convinced one of them that it required tens of millions of dollars in funding through a contract; that money was then transferred to a military contractor who provided the bars, the people said. …

The exquisite security precautions of such special access programs, also known as black programs, helped enable the scam, the people said. Since the two who were read-on to the program weren’t permitted to discuss it with other employees or supervisors, Rush appeared to have been able to move more than 300 two-pound gold bars to his home before anyone noticed.

Why don’t senior CIA officers do this all the time? Three possible answers are:

1. They do, but they are spies, so they don’t get caught;
2. The CIA’s vetting procedures are so reliable that most senior CIA officials are far too upstanding to steal money from the CIA; or
3. If you’re the sort of person who is going to invent a totally fake top-secret spy program, you’re probably the sort of person who is constantly making up wild stories, and eventually one of your lies will get you caught. And then they’ll search your house for misplaced gold bars.

So the Journal reports:

The Federal Bureau of Investigation began looking into Rush, authorities said, after the CIA referred allegations that he had filed false time cards. He had allegedly requested pay for time as a deployed Navy reservist, when in reality he had left the military a decade earlier, authorities said. ...

Rush not only wasn’t an elite military test pilot, as he had claimed to the CIA, but he also held no pilot credentials at all. The college and graduate school he said he had attended told the agent they had never heard of him.

I just. “I have a secret spy program that requires me to take $40 million of gold bars but you can’t tell anyone” is an outlandish lie, but in the right circumstances (your office at the CIA) it is apparently believable and certainly lucrative. “I am an elite military pilot with a graduate degree” is just, like, you are addicted to making stuff up, and you’re going to get caught.

VW resorts to sealed bids to avoid conflicts in $10bn engines sale. DE Shaw Moves to Close $5 Billion Lithic Fund to New Money. White House to Woo Wall Street Dealmakers With $400,000 DC Jobs. Beijing’s New Message to Its Citizens: Your Money Belongs at Home. The Great Middle Management Purge Is Here. World Cup player Elye Wahi arrested for alleged fixing offences on eve of tournament. Xi’s Enforcers Are Hunting Down Officials Who Consult Mystics and Borrow Too Much. Japanese fair trade officials raid 6 ice cream makers on suspicion of price fixing. Say Hello to the Newest N.Y.C. Area Code: 465.

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!