Are we in a bubble? Probably. But it's really hard to tell when the bubble's going to pop. It was obvious we were in a dot-com bubble in 1997, and yet the NASDAQ, I think, tripled. So it's very hard to time the market. And I think what you want to do is just always be in the market, always continue to save money and invest in the market, and be diversified in low-cost index funds such that you don't have to try and figure out if Walmart or Anthropic is the better value. The honest answer is I don't know. That's the bad news. The good news is, nor does anybody else. Yeah. Is it a bubble, or are these companies going to go up because of the incredible TAM? I think the answer is yes.
In today's office hours, we discuss the AI boom, career planning in the age of automation, and balancing parenthood with professional success. Question number one. Our first question comes from Steve, who emailed us. He asks, "I'm a simpleton in investing who listens to Prof G regularly. I know both are different sectors, but can someone explain how Walmart, with revenues of approximately $713 billion and operating profit of around $30 billion, can have a lower valuation than Anthropic, which has revenue of approximately $70 billion and forecast a profit of around $2.2 billion? Is that evidence of a bubble, or are investors simply pricing in much higher future growth and profitability for Anthropic?"
So typically, a stock price is meant to be the asset value. So just the land underneath the retail, or the intellectual property it owns, plus the present value of growth opportunity or cash flow. So all of these assets should be able to create an asset or a series of assets that create revenue that is greater than the inputs to produce that revenue. That's profits. And then your growth, or the multiple you get on those earnings, the P/E ratio, is a function usually of how fast that is growing or how stable and durable those cash flows are. And the reason why SpaceX and Anthropic and OpenAI are going out at anywhere from 30 or 40 times revenues to 100 times revenues is that people look at the addressable market and think, "Wow, if this company continues to grow in fairly short order, it could be the most valuable company in the world." And right now, if you are buying in any of these three companies, you're betting on a non-zero probability that this could in fact be the most valuable company in the world, because these companies are going out at anywhere between $1 and $2 trillion. The most valuable company in the world, I think, is around five or 5.5 — I can't remember if it's Nvidia or Apple. So you do end up with companies that are low growth, not exciting, not in software, don't have the margins of technology. You know, the addressable market for Walmart is pretty big. I think it has a 9 or 11% share of retail. It's probably never going to get above 20% because consumers, at some point, like differentiation and want to shop at different places, or regional friction, whatever it might be. But let's talk about TAM. Anthropic is 5 or 10x-ing every year. So people look at that and think, wow, that could be an enormous company at some point. And because it is in fact software, at some point when it hits break-even, a lot of those additional revenues aren't going to flow to the bottom line. So what's the takeaway here? Let's look at the numbers. Fiscal 2025 net sales of $675 billion at Walmart, adjusted operating income of $30 billion, operating margin of 4.4%, and the market cap as of late was $1 trillion.
So that's roughly one and a half times revenue, or 34 times operating income, which is actually pretty rich for a retailer. Walmart's operating margin has climbed from 3.3% in fiscal 2023 to 4.3%.
Walmart is just a different business. Walmart operates on the notion that with their scale and their operational excellence, they can pass the savings on to the consumer. So there are only three things that matter, or lines that matter, around shareholder value. There's the top line, which is perceived value. The next line, the middle line, is the price you charge. And then the line below that is the cost to deliver that service. And all shareholder value is a function of the ratio, relative distance between those lines. So Walmart is in the business of constantly pushing down the cost line. And then once you do that, some wonderful things happen. If you can either pass along the savings and pull down the price line that you're charging consumers, which is what Walmart does, and then the gap between the price you charge consumers and perceived value broadens, and the value proposition goes up, and you should expand share. And that's been Walmart's strategy the whole time — is that if they cut costs on operations such that they can ship a pallet of ginger ale to Arkansas more efficiently than someone else, they immediately pass those savings on to the consumer, and the value proposition of Walmart goes up. There's an old saying that starting to shop at Walmart is like getting a promotion in terms of the change in the quality of life, because you can now afford to buy imported beer versus domestic beer, if that's what you want to do. Other companies, such as Tiffany or Hermès, their job — they're in the job of pushing the perceived value line up as much as possible through innovation and scarcity and branding. And then that's wonderful, because then the gap between the perceived value and the price they charge goes up, which should expand share. But typically what they do is, as they push the perceived value line up, they raise the price.
Effectively, what you're doing in these companies is they're trying to push the perceived value up dramatically. They're also raising so much capital that they can offer these products at an incredible value. So supposedly Claude Pro or Claude Max, it costs $200 a month, but supposedly right now it costs them $5,000 a month for the compute and the chips to offer you that service. So they're trying to just capture a ton of share. These companies are growing incredibly fast, and their total addressable market seems almost infinite, which results in just kind of what I feel would be insane valuations. Let's talk about Anthropic. So again, the annual revenue trajectory was $87 million in January 2024, and it's $1 billion by 2024, and $9 billion by the end of 2025, and $30 billion by April 2026. So as you can see, this company — I mean, Anthropic really is a miracle of American innovation. It was started five years ago, and if it was in Europe, it'd be one of the five most valuable companies. The current run rate is about $40 billion ARR, according to people with knowledge of the company's financials, and the valuation was $380 billion post-money at Series G in February. And now it's in talks to raise money at a $900 billion pre-money. Internal documents project a $14 billion loss for 2026. No positive free cash flow until 2028 at the earliest. By the way, I think that's a couple years earlier than the projected positive free cash flow date for OpenAI. So at a $900 billion valuation and $45 billion in ARR, that's roughly 20 times forward revenues compared to Walmart's one and a half. They're just different businesses with different expectations. What's the answer here? You know, which is a better value? The answer is yes. And that is, I think, you want to invest in everything. I don't think you want to try and talk yourself into the illusion that you can find the needle in the haystack. Buy the whole haystack, dollar-cost average in low-cost index funds, so you don't pay fees or a lot of fees. So I do think there is a real serious bubble question. The bear case: Anthropic has taken in roughly $72 billion in funding and does not have free cash flow projected until 2028. And AI — you know, at some point AI is going to stop growing at the same pace. I think also, I think there's some existential threat from open-weight LLMs out of China. Regulatory concerns. I think the next big populist movement isn't immigration or affordability. It's AI regulation, which has fans on both sides of the aisle. So, are we in a bubble? Probably. But it's really hard to tell when the bubble's going to pop. It was obvious we were in a dot-com bubble in 1997, and yet the NASDAQ, I think, tripled. So it's very hard to time the market, and I think what you want to do is just always be in the market, always continue to save money and invest in the market, and be diversified in low-cost index funds such that you don't have to try and figure out if Walmart or Anthropic is the better value. The honest answer is I don't know. That's the bad news. The good news is, nor does anybody else. Yeah. Is it a bubble, or are these companies going to go up because of the incredible TAM? I think the answer is yes.
Support for the show comes from Cohere. As AI advances, one thing matters more than ever: staying in control. Most AI comes with strings attached, like sharing your data and infrastructure, or compromising your independence. With Cohere, you don't have to give up control to gain capability. You can have both. Cohere started seven years ago with a bold mission: scale enterprise intelligence without sacrificing sovereignty. With their solution, AI that leaves you in charge of your data and infrastructure, allowing you to maintain your competitive edge. It's designed for enterprises, governments, and organizations that refuse to give up control in pursuit of innovation. From frontier models to out-of-the-box tools, Cohere gives you everything you need to build and deploy secure, scalable AI that drives growth and delivers results. Their platform is designed around your business, with models that can be customized and trained on your proprietary data, so the solutions are uniquely yours. And control doesn't stop at customization. Cohere is built with multi-layer protections, access, and industry-certified security standards that help safeguard your data while supporting privacy, compliance, and other requirements. As AI continues to transform every industry, Cohere gives you the power to put it to work today on your terms. Take charge of your AI with Cohere. To learn more, visit cohere.com/vox.
Question number two.
