Bloomberg Businessweek Weekend - August 7th, 2026

25:36Not saying anything like that's happening. But I'm just saying you have a fiduciary responsibility. And you're right. You go back to Enron or WorldCom. Sure. I think the point I'm making is with Google, they probably thought there would be more customers. I imagine with Azure and with AWS, they thought would be more large players. But the problem with Anthropic and OpenAI is they've raised$200,$300 billion of funding. But they've actually raised more because OpenAI and Anthropic got all of their infrastructure built for them by Microsoft, Google, and Amazon. They didn't have to pay, I think, in the Sam Altman, Elon Musk trial, one of the Microsoft executives said that they cost$100 billion, so call it like$70,$80 billion of infrastructure.

26:17So the problem is that nobody else can get as big as them. No one else can get that much compute. No one else could afford that compute and have the chance to do the pre-training runs necessary. Except now China's coming up behind them. And it's unclear how anyone really deals with any of the problems I've been listing for years, which is unsustainable, unprofitable, and also not really finding the ROI and AI. Ed, play this out for us because I think a lot of people think, okay, for there to be some sort of ROI on this, one thing has to happen. And like the best case scenario for all this money being spent is that productivity increases.

26:57Fewer people are needed to do more things. There are some serious implications if that were to come true and to the labor force. And Dario Amadei of Anthropic has talked about this in the past. Maybe he's talking his book. I don't know. The other side of this is, well, if that doesn't come true, then what does it mean for these stocks that have gained so much on hopes that they would be responsible for some of this productivity increase? Like, how does the shoe drop? What happens? Well, the thing is, if you think about what Amazon, Google, and Microsoft have done, and met to some extent, but they're not selling compute capacity yet, is they have gone from being these cash-heavy, these cash machines, they just spill out money, low cash burn, high revenue, low assets, into these bulbous GPU-filled asset mongers, who are just full of these semi-built data centers for two customers or three customers at best so that they can do what?

27:49Rent them out so that they can rent their models. And it isn't really clear what the plan is at this point. And the problem is, for me to be right, it doesn't even have to go that badly. OpenAI and Anthropic have to grow so large to be able to make all of this data center capacity good. I mean, Google's, I think the UBS estimate was like$76 billion in 2027 of Google Cloud's revenue will come from Anthropic. How's Anthropic going to afford that? They burn tens of billions of dollars. So it's not just that these companies are unprofitable and unsustainable, but they have to grow so very large to make AI pay off, because otherwise there just isn't demand for compute at scale.

28:29Last time you were on with us, we got a really incredible response, to be honest. And a lot of people who weren't typical viewers or listeners of our show saw what you did and listened to what you did. And it really seemed like there's this, what you're saying is resonating with a lot of people. Like there's a, it was almost like there's this anti-AI fervor that's out there. And I'm just curious why you think that is. So I'm not sure it's anti-AI, don't get me wrong, but I think it's also anti-financial shenanigans. I think everyone sees the circular financing. I think they see that Microsoft, Google, and Amazon gets basically all of their AI revenues, either through products they're pushing on their customers or indeed compute spend from Anthropic and OpenAI.

29:12And the average person's existence right now is so expensive, so hard, so difficult. Getting a mortgage as a regular person is so difficult. But if you're standing up a theoretical data center in 36 months full of NVIDIA GPUs, the banks fall over themselves to give you the money. Cool, we've just raised, what, a 9 % bond? I mean, you can raise anything if you have a data center. And I think regular people can see that AI does not deliver what people promise. They can see the opulence of the people at the top of the AI industry. And they can also see that they're being lied to and being deliberately scared on top of all of this egregious circular financing.

29:48So you think people are actually lying? Like, or do you think people... Specifically. I don't know. Like, is it the companies at the hyperscalers, the CEOs, the bankers? Like, do you think, you know, and to be fair, we really should reach out to everybody. But I mean, is that what you're saying? Or do they not really know? I think they're overstating things. I think lying would suggest a certain malice, what have you, I don't want to accuse anyone of, but I believe that they are massively overstating what AI will do. You'll notice that AI people tend to speak in the future tense. They tend not to say, oh, well, today it can.

30:25It's always AI will, AI will. Oh, we're going to get the singularity. Oh, AI will do this and that. That's because when they talk about what's happening today, it's pretty mediocre outside of code. And on top of that, these things are horribly unsustainable and unprofitable. And on top of that, they've got these destructive data centers, these massive eyesores that poison black communities, these massive eyesores that need billions of dollars at a time when it's hard for a regular person to get a dime from the banks. So, yeah, I think that there is beyond just the misleading, this general sense of unfairness that AI taps into.

30:58And on top of that, if this all goes pear-shaped, these people are going to realize that there was an authority crisis happening, that so many people got beguiled by hyperscalar promises. And what it ultimately is, and I'm quoting Ed Elson of Prof G Markets here, our media, I believe, has a cult-like worship of the wealthy, that they believe that whatever the wealthy says will come true. And in the past with the tech industry, that's kind of come true, except it stopped really coming true about 10, 11 years ago. And we exited the era of hypergrowth. And that's all AI is. AI is an attempt to restart hypergrowth for hyperscalers who don't have a new Google search, who don't have a new iPhone, and certainly do not have a next Amazon Web Services.

31:41You know, the conversation narrative is changing. Don't you think it will continue to change and it might be uncomfortable in terms of how it plays out in financial markets? Yes, I think this conversation is only going to accelerate. OpenAI didn't cut prices because they found some mystical way of making things cheaper. It makes something 80 % cheaper. They saw the danger from China and they saw the competition from Anthropic and they said, well, we're allowed to burn billions of dollars, so why don't we just cut prices and then make it up in volume for an unprofitable product? I think the ROI conversation is only going to accelerate too because we should have really had it years ago.

32:16We really should have had it immediately. But again, people believe everything the tech industry says, and they just thought, well, they wouldn't say this and be wrong, would they? Where, in your view, does Elon Musk and SpaceX fit into this conversation? I bring it up because we learned this afternoon that Elon Musk's net worth has fallen to$684 billion, which, yes, is a lot of money. It has erased, though, the IPO gains from SpaceX. And we haven't been with you. Still the world's richest. You've got a rich guy on the bluebird. you you haven't joined us since SpaceX IPO'd but there's a data point there for at least in the short term reception to a public company that has pretty significant exposure with AI well I think SpaceX is kind of the proof point you need we have someone who can sink unlimited capital into this who can hire anyone who can theoretically stand up as much capacity as possible breaking multiple laws at the same time not getting the permits.

33:13And what did we get for it? We got Grok. And what is Grok? Well, it's a third, fourth, fifth tier LLM that really only some people use by accident when they turn on Twitter. So we have this thing where we've had our third anthropic in open AI. We've seen someone else try it. We've had what should be the proof point that AI is an industry, that we can have many AI labs and, oh, a thousand flowers will bloom. And what we have is manure. We have a company that loses billions of dollars to do what? I don't know. So should the U.S. be in an arms race with China for AI? No. I think that the arms race with China in and of itself is a marketing ploy.

33:53Oh, no. What's China going to do? Make a cheaper and better LLM? Uh-oh. It already happened. Nothing happened. Nothing happened. What about the security risks that these LLMs or some of these agents are exposing? those from OpenAI and those from Anthropic? Well, I think the biggest risk with OpenAI and Anthropic's agents is they don't appear to do basic security practices. They don't appear to take care of how they're using their systems. OpenAI say, I actually question this entire story, that their agent ran autonomously for multiple days, burning what sounds like unlimited compute. Either this company has run so terribly that they were running up millions of dollars of bills to randomly do stuff, and also they don't watch what it's doing.

34:38Software does what it's told to do. We don't know the prompt. We don't know the training. And they're not releasing the model. But it doesn't change the fact that these agents reportedly found weaknesses in code that if not exposed, that could be vulnerable. What I'm saying is there is an idea that if this gets into the wrong hands, then systems could break down, just very briefly. One thing, it's already in the wrong hands, open air and anthropic. They've shown they do not have the responsibility to make security tools. They should not be making them. They don't know what they're doing. It's blatantly obvious.

35:12And on top of it, it's they brute forced a hacking agent. They shoved as much compute power into it as possible. You can also pay hackers to do that. It's illegal. Also, this all sounds illegal. I'm no lawyer. I'm no judge. But I don't know why they're allowed to do this. Yeah, these things are dangerous if they're allowed to be trained on cybersecurity measures and execute against them. I find the whole thing repugnant because everyone is saying, ooh, look at the scary LLM versus looking at the companies that run it. We've got to run. 20 seconds. Anything that would change your mind and make you say this is real real quickly.

35:44Not really. No. Okay. Thank you. Thank you. Thank you, thank you. There was a lot of conversations this week, and it was great to get your input, Ed. Thank you. Ed Zittrin, he's CEO, Easy Primary Research, right here in our studio.

36:03This is the Bloomberg Tech Minute brought to you by ChatGPT. Now with ChatGPT work. I'm Carol Masser. Globetrotters hunting for airfare bargains are in for a rude awakening, as the days of stumbling across a cheap seat on a popular flight could soon disappear. Bloomberg's Juan Ha reports that airlines from Delta to Virgin Atlantic are adopting artificial intelligence to change seat prices more quickly by weighing dozens of variables in real time, helping capture more revenue while shrinking pricing gaps that once allowed travelers to find bargain fares. Machine learning models can more accurately forecast demand by analyzing historical booking patterns, seed inventory, and seasonal trends, while also continuously tracking competitors' fares and capacity changes to update prices in near real time.

36:51The technology could lead to higher fares on busy routes as airlines pack flights closer to capacity, but may also result in lower fares on off-peak and lower-demand routes. That's the Bloomberg Tech Minute brought to you by ChachiPT. Put ChachiPT to work on your most ambitious ideas and projects. Get started at ChachiPT.com today by selecting Work Mode, available on Plus and Pro Plans. Whether you're planning a big tech event, launching a new campaign, or just stocking up on team gear, finding the right promotional products makes all the difference. 4imprint offers thousands of options, from on-trend apparel and premium drinkware to tech, totes, and giveaways, so you can find the right fit for any audience, purpose, or budget.

37:37You can customize it all, your logo, your message, your look, and many items come with no setup charge to help you save. And if you're really watching the bottom line, you'll find standout choices at every price point so you can make a real impact while staying on budget. Plus, you'll get expert help, fast turnaround times, and their 360-degree guarantee. So you can be 4imprint certain your order will arrive on time and look exactly right. Whatever your goal, 4imprint makes it easy to find your perfect promo match. Explore the possibilities today at 4imprint.com. 4imprint. 4certain. Let's talk about healthcare for a second.

38:13It doesn't always work the way people expect it to. If you've ever waited on a prescription refill or had a hard time getting the care you needed, you know the feeling. The system should just work better for everyone. That's exactly what the people at Optum are trying to do every day. They're a healthcare company linking patient care and pharmacy services and using data and technology to drive the whole system so care is connected, not complicated, for patients and providers. Things like making it easier to get care that looks at the whole person, from primary care doctors to mental health support and even in-home care, and then using technology to make sure they all work together.

38:50technology designed to help doctors spend less time on busy work and more time with their patients. And those prescriptions? Optum is working to bring costs down, save patients money, and make it easier to get refills. Little by little, Optum is helping make healthcare work as one for everyone. Head to business.optum.com to see how. You're listening to the Bloomberg Business Week Daily Podcast. Catch us live weekday afternoons from 2 to 5 p.m. Eastern. Listen on Apple CarPlay and Android Auto with the Bloomberg Business app. Or watch us live on YouTube. Former OpenAI researcher Leopold Aschenbrenner's fund, Situational Awareness, was riding high on a 439 % return until a brutal July tech sell-off triggered massive margin calls.

39:39Ken Griffin came to the rescue as he's wont to do with beaten down assets and snapped up some of them at a discount. Citadel's flagship fund surged 5 % in July after the firm bought most of Situational Awareness' public stocks at that discount. And the transaction helped boost year-to-date gains at Citadel to 12%. That's according to a person familiar with the matter. And then, after all of that, and really, to be quite fair, just a few days after his hedge fund came close to a collapse, Oshin Brenner is back in the game, plunking down$400 million on a privately held company, according to people familiar with the matter.

40:16Bloomberg Opinion's Aaron Brown knows a thing or two about market risks. He's former chief risk manager at AQR Capital Management, the hedge fund of Cliff Asness. Aaron argues that situational awareness's massive return was a, quote, warning, not a triumph. Aaron, nice to have you here with us. Welcome, welcome. You do know a thing or two about risk. You write that math sets a speed limit on how fast a portfolio can compound. What is that math? take us to Bell Labs, and take us to 1956. Thank you for having me, Carol. Yeah, so this is John Kelly from Bell Labs, a physicist, also a fighter pilot and a bridge player, you know, really fascinating guy.

40:58And he discovered that, you know, most people assume that taking more risk means you increase the possibility of very good and very bad outcomes. But what he discovered is there's a limit. And beyond that, all you do is increase the probability of very bad outcomes. The Kelly point. And for those of you who are familiar with his work, probably in gambling context or investing context, the optimum is half of the Kelly limit. You go halfway to the cliff, and that's where you get your maximum growth. 439 % under any reasonable economics, any analysis. You know, we only have partial information about situational awareness.

41:40We have a 13F from April that doesn't have the shorts. We have, you know, Wall Street trader chatter. But any reasonable suggestion says 439 % meant they were well over the Kelly limit, meaning sooner or later you have this happen to you. You blow up. Could be years, could be tomorrow. Aaron, the column's about situational awareness, but it's also about the time period that we're in. And you referenced what's happening in South Korea and specifically with some of those levered ETFs and the chip names there. To you, does this illustrate sort of where we are in maybe a market cycle, a hype cycle?

42:16What does it tell you?

42:21Well, I'm in an AI bolt myself. And I have some considerable investments in AI. Nothing we're talking about today would affect that. But having a long-term vision that AI is going to be very big doesn't give you a reason to take unlimited risk. And whether we're talking about South Korean retail investors, for that matter, New York retail investors, or the situational awareness fund, you have to think about the long term. You have to think about, do I survive long enough to collect on my bets? So I have a question for you. You know, investment folks, right, and companies often have, you know, risk managers.

43:06And I understand, you know, so. So he's laughing. Am I wrong? Is there not someone to say, OK, you're in over your skis here? You know, like so what's your read on this this firm? I mean, it's still a 10 billion dollar hedge fund, so it's not like it's collapsed. And Citadel was happy to take, you know, but we understand that that's what they do. But I don't know, like how do we kind of step back here in terms of internally what this company is doing or this hedge fund is doing? Okay, so Citadel has tremendous risk management, has some of the best risk management on the planet, which is why they're in a position to do this kind of thing.

43:44I don't know anybody at situational awareness, but the fact that they had either four or seven, I've seen different media reports, total financial professionals, makes me suspect, and plus their investments, make me suspect they did not have a risk manager or did not pay attention to him or did not have a qualified one because it just does not seem like a risk managed portfolio. And all of the public statements we've heard from them only mention expected return, you know, future outcomes. None of it mentions risk. So yes, they should have had a better risk manager. And that 10 billion is pretty misleading.

44:24First of all, I think that they're still carrying Anthropic at 5 billion. I don't think they've marked it down at all from its peak, and it's certainly worth less than that. Second, you know, they started the year at$1.5 billion, we think. They grew to well over, you know,$10 billion. And so I suspect most of the investors in SA on a dollar basis have lost quite a bit of money. If you were in January, you know, if you were part of that$1.5 billion, you're still up, I think, 30 % for the year, Wall Street Journal reported. but most of the people got in closer to the peak and are probably well underwater today.

45:04We're speaking with Aaron Brown, columnist for Bloomer Opinion, former chief risk manager at AQR Capital Management, also the author of Wrong Number, How to Extract Truth from a Blizzard of Quantitative Disinformation. I like that you brought up that private stake in Anthropic because I don't want to give the whole column away, and I encourage everybody to go read it. I was just sending it around to some guests who've joined us in the past. You make the point that the The Anthropic investment, it makes sense that they still have that because they couldn't take margin on that. They can't transfer shares of that.

45:36So it's like at the end of the day, it's sort of the safest thing for them because they couldn't bet against it. Well, they can't lever it. There are people who will lend you money against it, but they're not daily margin. So you're not getting the kind of leverage they have for public investments. And it's possible that they didn't lever it at all. My guess is they didn't. If you have public stock, you're going to lever those. You don't have to go to your private. But yes, a company like Situational Awareness, with their approach to the market, they should be making private investments and not levering them.

46:13What about the banks that were lending the money? What's the due diligence on that? And they were well-known banks. We keep citing like JP Morgan. in. I'm just curious how that typically works out. Well, it typically works out like this one did. They get all their money back. Yeah. Okay. But the due diligence ahead of it, Aaron, like, do they just, is there something that they look at ahead of it in terms of? Yeah. Yes. Yes. They do that very carefully. And Archegos was, you know, a couple of years ago, So that was the exception. That was where they all got burned because they went ahead over their skis, as you put it.

46:58This is exactly how it is supposed to work. The banks always should do okay. And what they were looking at is they were looking at the market for this stock. They knew Citadel was around. They knew there were other people around who would be in a position to buy on a dip. And they quickly got out before they got hurt. And as I say, that's how it's supposed to work. And that's how it usually does work. That's why these companies are so big and rich. But you also make the point in the piece that Citadel learned this lesson the hard way. Like they're looked at right now as coming and swooping in at this time.

47:31But post 2008, they suffered some serious losses. And AQR in 2007 did as well. Yeah. So yes, risk management is a lot of unhappy experience, but learning from experience. So, okay. So hindsight is 20-20. If this portfolio, and again, we don't have complete information, like you said, this is based around what has leaked in 13Fs, but what would have been the right way to build positions in companies that you believe in that wouldn't have overexposed them on the downside? Well, okay. So situational awareness, the investment thesis is that AI is going to be gigantic, is going to, you know, I won't say take over the world, but is going to be bigger than even most of the optimists think.

48:26But it has no thesis, at least as many of the public statements, about the path to getting there. So you have to think about that through and say, what are the scenarios where we're right, but we don't get to keep our positions? It also has, I think people are not aware of how complex its positions are. Again, this is looking at the 13F without the shorts, but we can see they're betting against a lot of these companies. They're picking and choosing. And so they've got longs and shorts and they've got a lot of puts on. So they're betting certain segments will do well and others are going to get beaten out.

48:59So this is a very complex bet. So you have to think, OK, what's the situation in which we're right? But what's the worst point between now and then? And can we survive it? it doesn't appear to me that they were asking that question or they weren't, you know, taking it seriously enough. I want to wrap up with, you know, you said earlier, Aaron, that you are an AI bull and you have positions, situational awareness. We were trying to figure out, is this kind of maybe a coal in the canary mine when it comes to the AI trade and narrative? What would you say? It's not? No, no, I don't think so. I mean, you know, we had to pull back in AI and, you know, a lot of people got hurt, but really the only headlines, disasters are the people who were over levered, either the ETFs or situational awareness.

49:47Most of the investors are there for the long term. You know, you don't see a huge sell-off. I don't see anybody changing their mind about AI. You had to expect. I mean, these stocks are extremely volatile and the events of the summer have been, you know, pretty much normal volatility for this sector. So if you were investing sensibly in AI, this summer was not an unpleasant experience for you. We're going to leave it on that note. We're so glad we could get you on. We read your column and thought it was super smart and just a different take. And it was something we wanted to bring to our viewers and our listeners.

50:21Aaron, thank you so much. I hope you'll come back. Thank you. I will. Okay. Good stuff. Aaron Brown, columnist for Bloomberg Opinion, former chief risk manager at AQR Capital Management. His book, Wrong Number. I'm glad he said he'd come back because if he would have said, I'm not going to come back. Yeah, I guess it wouldn't have been. I put them on the spot. That's not so fair. No, it's a good strategy. I like it. Now we're going to hold them to it and we're going to get them on for our next call. It's a story like we're still trying to figure out. Aaron, please come back. This is the Bloomberg Business Week Daily Podcast.

50:53Available on Apple, Spotify, and anywhere else you get your podcasts. Listen live weekday afternoons from 2 to 5 p.m. Eastern on Bloomberg.com, the iHeartRadio app, TuneIn, and the Bloomberg Business app. You can also watch us live every weekday on YouTube and always on the Bloomberg Terminal.

51:19When you're running a business, the best days are the ones where priorities stay on track. For midsize and large companies, risk can affect multiple parts of the organization at once, from property and liability to cyber and regulatory challenges. At that level, managing risk becomes an ongoing discipline. At the Hartford, the focus is on helping businesses manage risk before it turns into something more disruptive. And when losses do happen, that work is paired with insurance coverage shaped by years of underwriting, risk engineering, and claims experience. Learn more at thehartford.com slash risk mitigation.

51:55Policies provided by Hartford Fire Insurance Company and its property and casualty affiliates, Hartford, Connecticut. Whatever your goal, trade show giveaways, client gifts, or team gear, 4imprint has the promo products to match. With thousands of options, from apparel and drinkware to tech and totes, it's easy to find the right fit for your brand and budget with standout choices at every price point. And with their 360-degree guarantee, you can be 4imprint certain your order will show up just right, right on time. Explore more at 4imprint.com. 4imprint. 4certain. As industries evolve faster than ever, companies need an environment that accelerates strategic growth.

52:32And Michigan delivers on that promise. From emerging startups to global enterprises, Michigan offers what executives value most. A resilient, innovative ecosystem. Diverse communities that attract top talent and a quality of life that supports work-life balance. With our unified Team Michigan approach, businesses scale faster and compete at the highest level. Michigan, pure opportunity. Seize your opportunity at michiganbusiness.org.