How AI Will Transform Investing: Cash-Flow Reality, Data Quality, and the Future of DIY ETFs

Welcome in everybody. Glad to have you here today. Headline quote that we'll talk a lot about today is the largest companies in the world are cracking under the weight of staying in the artificial intelligence race. We'll get into that and much more today. We're excited to have you guys here. This is our monthly podcast. Today's title is How AI Will Transform Investing. We'll dive right in. We've only got a limited amount of time, so we'll jam-pack this whole next 30 minutes. Start up with the disclosures, disclaimers. The important part is we are providing research, not advice, but you can read all through these in the replay that we send out. On the agenda today, first, talk about the latest research we've put out, a bunch of different new long ideas, some danger zone picks, earnings previews. If you guys haven't seen those, we named the 5 companies most likely to beat. and 5 companies most likely to miss upcoming earnings based on core earnings and Street earnings estimates. Some great insights there. We've put out a few more articles on AI, mainly the stock market being rogue AI's most obvious target and how to protect it. And then also the shareholders versus employees, which is the next AI trade-off. And we'll talk a little bit more about those today. Done a few different earnings watches since we last talked. We've got our focus list long and short performance for 1Q26. And we were the number 1 ranked stock pickers again in April. We're going on 58 straight months of number 1 ranking, so we're really proud of that. But onto today's agenda, we'll talk about some upcoming trainings, some new features of the site, and then we're going to dive deep on AI, specifically our AI predictions that have come true. And then we'll dive into how AI is going to take down the business of investing with some insights that David shared this past week with Harvard Business School. So without further ado, I'll jump right into all of that. The upcoming trainings we've got, the first one we're calling Buy This, Not That: The Answer Key to Beat the Market in 2026. That will be next Wednesday, April 15th. Registration links will start going out via email today. So be on the look for that. And then for our pro and institutional members, we do have the next Ideas Dinner coming up on April 21st. If you've registered for those in the past, you're already registered. But any new pro or institutional members, emails will go out as well to sign up for that Ideas Dinner with David. Those are always a good time.

By the way, those Ideas Dinners are awesome. If you're a pro customer, you haven't been to one, you're missing out.

Yeah, it's great discussion going on there.

I've had a lot of one-on-ones with our pro clients here recently. And those that know, know, because you've been. If you haven't, check it out. It's good stuff. Come join. Sorry to interrupt, Kyle.

Nope, all good. On the next item is a new feature. We've updated our best-in-class reverse DCF model a bit. We've added TTM data to the forecast page, and so you can now see the TTM, not just the prior fiscal year. You can see the TTM values for each of the drivers of our DCF model: revenue growth, NOPBT, net operating profit before tax, cash tax rate, and then change in net working capital and fixed assets. And we also have got historical TTM on a new TTM tab in that model. So just surfacing more ways to view the TTM data in line with our DCF models there. We're excited about that update there. But onto the big topic, AI. Everyone's talking about it, right? But specifically our predictions, AI is taking down some of the largest companies in the world. And that might sound wild to think about, but it's not as far-fetched as you may think. For months, we've basically warned about the dangers of the AI buildout, the huge spending going on in AI, and how it's impacting the largest companies in the world. And we put out a bunch of different articles, starting with The Losers in the AI Arms Race. That was back in November of 2025. And we found a hidden $500+ billion accounting trick. that basically a bunch of the top AI spenders are using to gloss over the cost of all this AI buildout. And we put out a new article, The Worst Is Yet to Come, and then The Great Rotation: Fade the Dying Tech Stars. And this was all in the span of a few different months to really take a deep dive on, again, how the AI revolution, the AI war, AI spending is impacting some of these huge companies. And the big takeaway is people are missing off-balance sheet debt that is cratering cash flows. This was a figure from our original AI Losers report. You can see Meta and Oracle being down at the bottom there with huge negative cash flow, and they basically don't have the cash flows to sustain the huge spending that they're committing to. And we say people are missing it because mainstream media doesn't tell you that they're losing cash flow. This was from CNBC article that Alphabet, Meta, Microsoft, Meta, and Amazon generated a combined $200 billion in free cash flow in 2025. When we look at the numbers in our system, they burned $146 billion out of those same companies combined. So huge disconnect between what the media reports and what the actual numbers from the company financials say. And that disconnect creates opportunity both on the upside and downside and pitfalls of companies to avoid that Without this data, you wouldn't have any clue what's going on.

Yeah. Look, this is one of my favorite charts. I loved it when you found this, Kyle, because hey, this is what the world is telling you and it's not true.

From that analysis, we've realized Oracle and Meta were the clear losers, but we were not here to just find losers. Like I said, it creates opportunity both on the upside and downside when you've got that huge disconnect in cash flow we just pointed out. But we identified Photronics was a cash flow winner a few months before we flagged Oracle and Meta. And since that November 10th article, we've got Meta down 1%, Oracle down 43%, and Photronics up 98%. All the other 2 companies in that losers report are not living up to their names. They might have the big flashy headlines, but they're spending billions and don't have the money to keep spending those billions. Cash burn rises, layoffs are now accelerating, which we pointed out in more recent articles. Shareholders versus employees. They found a way to increase some of that runway, which was lay off a whole bunch of people. They can cut costs and keep some of that AI spend possible for a little bit longer. And we call it the AI purge. But ultimately it is a bandaid that basically these companies can delay the inevitable, but it doesn't change the fact that they are spending billions while burning billions in cash. And ultimately, the impact of AI will be felt not only at these companies, but across the entire investing industry. And these are the first early signs of what we can expect in this whole AI trade.

We see these stories all the time, and we just recently published a report on Allbirds, which was a big popular IPO, came out at $2 billion, went up to $4 billion, and then cited all these kinds of restructuring plans to save the business. And it just closed down for $40 million. So the narrative can attempt to delay the inevitable, but it is just a narrative. And at the end of the day, the numbers are what matters. And either you got a business model that can work or you can't. What we're revealing to you in these AI stocks is these profits are a figment of the imagination. So be careful. And so far, I would also say our predictions have been pretty good. So maybe worth listening to.

And that takes us to our next topic, which is AI's impact on markets and investors, not just the Few companies, but a much more broad discussion. Earlier this week, David spoke at Harvard Business School, and we're going to hear him share those same insights with you today. There's a lovely picture here of David at the Harvard Business School.

Yeah, I just got a better one.

Oh yeah, there we go.

Yeah, so I just got this from the professor. Yeah, I've been talking to a lot of advisor teams, investors in general, and for the last few years about what's real and what's not real about AI and And I enjoy doing that because we've had a lot of experience in what it takes to build something that could be trusted. Google spent $2 million and built their AI agent investing on our data because they know it's reliable and it can be trusted. And so that's, I think, a very important distinction when it comes about all AI. My fellow panelists here at HBS said the same thing. It's all about high-quality data. The person in the middle, Chris, her firm specializes in actually doing these really sophisticated searches to create, she calls 'em new datasets. They're not new. She's just able to collate more data. And their specialty is to do it at a lower cost, like to burn less tokens. So it's really becoming not just a skill to know what to look for, but even to do it with efficiency. And so I wanted to, I put some stuff together for them and I figured, you know what, I should share it with my audience too. And again, this has also been conversations I've been having at conferences. And here in Nashville, Tennessee, I've been out to see a bunch of local advisors. If you guys, any advisors out there, you want me to come talk to you about this or talk to you and your clients, a lot of these advisors say, hey, can you explain this to our clients as well? It's not just about what's good and bad about AI and what works and what doesn't, but how to get healthy exposure to AI investing. Because a lot of it's hocus-pocus and it's risky. And there are smarter ways to get AI exposure. And that's something that we have a very unique perspective on. Again, Google chose us to build an AI agent. They've never done that before. They spent millions of dollars to build this thing, and they fed it our data entirely based on our data to demonstrate the art of the possible when you endow a machine with highly reliable data. I want to say that again because it's important to understand. When you endow machines, AI machines, with super reliable, 100% accurate data, you can get reliable results. Otherwise, it's garbage in, garbage out. And so let me, you know, so here's what I shared with some of the kids at HBS. And it was the business analysis valuation class. And I've done this probably, I don't know, 10 or 12 times. And it's pretty funny. Like all the, it's like this is the 4th different professor. And every single one just goes on to be like a superstar professor. So, this class is nice. It was when I started doing it, they were thinking about closing the class down. And I got a call from Charlie Wang. He's like a superstar academic on all things AI. He wrote the core earnings paper, all things AI and high-quality research and integrity in capital markets. And the reason he called me is because he said, David, People don't want to take my class. And by the way, at Harvard Business School, if they don't want to take your class, you get the boot. At Harvard Business School, teachers get paid and promoted based on the surveys by the students. How's that for a novel idea? How much would you've loved it if when you were in school, teachers cared about what you thought? Kind of. So it's an interesting thing. I think Harvard Business School and maybe Darden, Virginia, Virginia's business school. Oh, there's a— There's a couple, just a few that do that. And so Charlie called me and said, they don't want to take my class because they don't think fundamental research matters anymore. Too much work, not enough reward. And I said, when, if there was a technology that could give investors or your students their cake and they can eat it too? What did that mean? It means you get the benefit of the footnotes without having to do the work for the footnotes. That's what we provide. And now this class is the platform. to boosting careers at Harvard Business School. So let's talk quickly. Let's go to the slides here. What are we talking about here? What is the evolution of fundamental research and how it affects people in business school in particular, and even people coming up in their careers, is to understand that fundamental research, even its own evolution, it gets better as you can do more analytics with more data. So for example, I like to talk about data collection. In the beginning, AI was good at data collection, not much else. Now it can do a little bit of modeling. And eventually, you can combine data collection and intelligent modeling into alpha. The problem is if the data collection is bad, all that is worthless. Whatever insights you draw from bad data could be worthless. And listen, folks, it doesn't matter if it's 99% good.

Why?

Can anyone tell me why it doesn't matter even if the data is 99% good? Anybody wanna put something in the Q&A or the chat? Because if you don't know which 1% is bad, how do you operate with that? How are you gonna make reliable decisions? This is where I think where AI is a long way off. A place where we can actually trust the AI to give us answers that we would get from an expert. This is why I keep saying you can only trust AI if it's endowed with 100% accurate data that's been vetted and we know to be true and reliable, sophisticated. Otherwise, are you going to trust this thing for medical advice? You're going to trust it for an investment decision? You're going to trust it for a life decision? You're going to trust it for a business decision? No. Maybe you're going to trust it to do some initial work that you have to review, like what you would for a junior analyst. Great. It's like an infinite number of secretaries. And junior analysts, great. And I would say mediocre. I've been using it to do some stuff like that. Every time I'm like, oh, I wanna do this or that, try it for close to free. See if the AI can do it. And it does okay. The point is, if you've got bad data in your model, it's not just gonna be bad, but those errors can compound. And that's where it gets really scary. I just got an email from a friend who's a registered investment advisor. Runs Silver Rebellion here in town. And he pasted it and sent it to a few of us in our group who are into AI for our entrepreneurs group. And Claude lied to him and admitted he lied to him, said, but it wasn't a bad lie. We didn't mean to be mean about this lie, but we lied because of all these reasons. And it was just ridiculous. Ridiculous. He busted into the lie, outed it, and at least admitted, yes, I did lie, but you know what? It's too late by then. So. AI needs better input. What does that mean for these HBS students? Their career needs to be focused on how to identify unique data sets and how to derive value from them. This fancy word ontology— I hate using fancy words because everybody just thinks you're a dork or trying to be a smart aleck. But it's— man, it works so well. It's one word that speaks to what AI has to do for us. Is find meaning in all this data out there. Now, obviously, the meaning is not reliable if the data is not reliable, of course. But organizing data into meaning, organizing data into thought structures, thoughts that lead to insights that lead to meaning— that's ontology. That's what we need to be good at as humans. That's what we can do, because the machines can do all the grunt work. We have to figure out how you get insights from data. That's where these guys should focus on their careers going forward. And look, we've got proof now that if you can do this right, you can use AI to create real insights. And this is a— this is from our first slide, really. First, you start with gathering data. It's basically pretty simple. You get the financial statements, then you get the footnotes, and you add onto that advanced modeling, and you get novel alpha. And novel alpha is where you outperform the market. And we've got 3 live-traded indices, one based on our very attractive stocks, one based on our proprietary measure of core earnings, superior measure of profits, and the third one, just an enhanced S&P 500. All 3 beating the S&P substantially over the last 5 years and in the short term. Example, here we are, Core Earnings Leaders Index. As of just a couple of days ago, beating the S&P by 28% over the last 5 years. We have other indices as well that have outperformed. So what does this look like? It's my favorite slide. I love this slide. It's funny because any slide with monkeys is funny, right? But I think it helps people understand AI is clearly not making the leap to artificial generative intelligence. That's not a thing. As much as they've tried to convince us it's a thing, It is definitely not a thing. We're not there yet. How does it get there? How does AI get to where we can really trust it? I've explained, I think, before, and now I will give you an example. This is how we did it in New Constructs. First, we got the AI to read financial statements. Then we got it to read footnotes. Then we put this data through advanced modeling. We had an ontology. We had a purpose for that data. We wanted it to give us superior insights about the profitability of businesses. Because we believe ultimately profitability drives valuation. A lot of noise in the interim. Allbirds is a great example. Oracle's a great example. There's a lot of noise in the interim, but ultimately the market is a weighing machine, not just a voting machine. Hat tip to Warren Buffett. That's his quote. AI will evolve like our AI evolved. First, we taught it financial statements and footnotes, advanced modeling, and then from that, We created Signal Core Earnings. The Core Earnings Leaders Index beats the S&P in the short term and the long term. That's novel alpha. And all that's automatic. We just send a data feed off to Bloomberg and they recalibrate and calculate the index every day based on that. And then that is what Google spent a couple million bucks to build an AI around. And now we have what I would call real AI. And AI, you can go in, you can say, which are the best stocks in the tech sector? Which companies are most likely to miss earnings next quarter? And you get reliable answers based on research that's generated alpha in the stock market. That's real data. So next step is let's talk about the future of the investing business for business school students. So how AI is going to affect the investing industry. First, we have to understand how much money is in legacy mutual fund ETF, legacy funds management. And these numbers, I'll admit to you, they're a couple years old, but they make— you get the point. If they're— if anything, they're bigger now. $103 trillion in AUM. And I've got this built as a citadel on purpose because the mutual fund industry Has all this money, $103 trillion in AUM, $244 billion in revenue, $83 billion in EBIT. I'd say these numbers are probably 20%, 30% higher, even though 75% of all funds underperform every year. 96% underperform anytime over 1 year. So, 25% get lucky. 99% of the 25% are just outperforming out of luck. Very few people outperform more than 1 year. Super few, less than 4%. And yet legacy firms still hold all of this money. How have they protected this money? In the beginning of the mutual fund, you couldn't even— you couldn't really invest in multiple stocks. It was too cumbersome. You couldn't even invest. You had to actually call your broker. Hey, will you buy me some of these shares? And then it was super expensive to trade. So, a lot of people couldn't do it for that reason. It required a lot of money just to buy one stock. You had to buy stocks in minimum 100 shares lots. What I'm saying is not only did you have to pay high transaction costs, but you couldn't buy anything less than 100 shares at a time. How many people could buy 100 shares of 100 stocks and have a diversified portfolio? Nobody. So they forced you in to buy mutual funds if you wanted to have any kind of intelligent strategy outside of just picking stocks on your own by calling up your broker. And so, that's how the mutual fund industry built their business and protected this huge amount of profits, all the while underserving people. I think it's one of the greatest grifts of all time. Billions and billions of dollars in fees for underperformance. The world's changing. So, we've seen things change in a big way in the last few years, right? You've got online brokers got really made it easy. to trade. They took the cost of trading to zero. Internet and social media make it easy to scale your business. You don't have to go through a mutual fund. You can get to an ETF. Fractional shares. You can own— you can create an ETF on your own or be highly diversified with very small amounts of money. The only thing in recent years that's kept this mutual fund industry alive and Wall Street alive in many ways is what we call, we nerds call information asymmetry. All that means is that there are people who know more than they take advantage of it. And I submit that with new constructs, even that is gone. So, why don't you just create your own ETF? Go into our system, pick a bunch of stocks, choose how you want to weight them. Click a button, print an order form, send it to your broker. What does that mean about legacy portfolio managers? Probably won't be many of them, right? When 96% underperform every 2 years, what do we need these people around for? Collecting excess fees, right? Getting paid to underperform. Where else does that happen? Where else do people get paid to underperform? Nowhere. We've taken down the walls. And so, I think not only can individuals, I think, manage their own money, but my thinking is that you're going to have influencers get into the business, right? If you want to raise a lot of assets and create an ETF, you don't put some guy you never meet behind a big fancy office on Wall Street or in Boston that's going to manage money for you. You're going to have people out there that are going to convince you that, hey, they've got insights over and above fundamental analytics, right? Because that stuff is going to be something that They can't do, but who knows? Maybe Kim Kardashian knows more about the makeup industry. Maybe Mark Cuban knows more about something in startups. A-Rod about athletics. Doesn't matter who it is. So one of the things I might do real fast here is show you how this do-it-yourself thing can actually work. Let me show you what I'm talking about. You would go to the new construct screeners and you would say, I want to find all the best stocks, very attractive, and I want to get Strong economic earnings, and I want gap earnings distortion score that's good, and I want return on capital over 15%. Give me all those stocks, and we're gonna here we go. We got a pretty big list. Let's say I want a free cash flow yield greater than 10%, and you know what? Let's keep that list. It's pretty big. Maybe we want to be this diversified. Let's remove all the suspended rating stocks. There we go. Now we're going to put this into our portfolio. I'm purposely not showing these stocks, by the way. We're going to add this new portfolio. We're going to say our Do-It-Yourself Fund Example. So it's— now we have a portfolio. Here it is. This is a great portfolio. Obviously, we haven't done a ton of diligence, but we've been pretty fast here, right? Now we click on Do-It-Yourself ETF, and here are all those stocks. Now, the only thing we need now is to determine how much money. Do you want it to be $10,000 or do you want it to be $1,000? It doesn't matter. We got fractional shares. We can do this for any amount of money you want. And then you choose your weighting methodology. Do you want to weight equal cap or market cap like everybody else? Maybe you want to weight based on core earnings. And let's say, all right, you want this to be— we'll do a round number— $21,000 at $1. And then here are the shares you need to execute on that. And this is how the stocks should be divided up based on core earnings. You could do it on other things like free cash flow yield. You could create a custom weighting. Typing starts out, assumes equal in the beginning. Point is, you can do this any way you want. You can also upload a list of tickers. You could go say, hey, I want to download the tickers for the QQQ and then pull them in here and say, I want the QQQ, but I want to weight— instead of waiting on market cap, I want to weight those stocks based on Return on invested capital, or I want to do it on anything smarter. You get the idea. And this is what we do with one of our Bloomberg ETFs. It's the Enhanced S&P 500, weight stocks based on core earnings. Then all you got to do next— there it is. Now you go execute. That's something you can share with whoever you want, but you get the idea. This, I believe, changes the investing industry Forever. And I didn't go into this much detail because the HBS students, they just got one class in terms of how AI is going to affect investing. You got the whole course. And I'll be sending out diplomas later today. Just kidding. But I hope that this was helpful in terms of laying out a thesis on exactly how and why the investing industry can and should evolve and specifics on exactly how that could happen. I hope it was helpful. If you guys don't have any other questions, we are right—

I did have one quick question, David. Someone asked, why do we suspend stocks?

Yeah, we suspend stocks because sometimes financial statements aren't reliable. We suspend stocks because they announced they're going out of business, like Alberts, right? That was sort of what alerted us to this. We suspend stocks because sometimes their auditors tell us their financial statements aren't reliable, like Supermicro. Before it dropped 50% last year. We suspend stocks when they don't file their 10-Ks. We actually have a whole section of our site where we explain— if you want to pull that analyst notes section, you could share that with folks, Kyle. We have explanations on why we suspend stocks. But look, at the end of the day, not all data is always great. So not only do we have the best data and we go through the filings, we do all that work, but we make sure that data is still reliable. And when we're not sure, we suspend. And that's part of why I had the partners we've had. That's why Harvard Business School leans on us to talk to people about AI and to teach people how to do real fundamental analysis. Why Google, Bloomberg, and some of the largest asset managers in the world— Goldman Sachs Asset Management, Allspring— I can go on, clients of ours in the years. Yeah, good question. If that's it, we will sign off. Kyle, were you able to get that link?

Yes, I threw that in the chat there. Great.

Okay. Thanks, everybody. We hope you have a great weekend, and we will see you next time. Bye.

Thank you.