You sometimes have to leave the world of business to understand business. Right now, many of the biggest players in the hyper-competitive battle for AI dominance — including Microsoft, Google, Meta, Amazon, OpenAI, and Anthropic — are forcing us to question everything we know about how to value a stock. It’s turning us upside down. It’s making us unsure whether we would rather own more defensive stocks, such as Clorox or Procter & Gamble , at any price than high-growth megacaps. It has rekindled investors’ love affair with PepsiCo and Merck . It makes us pine for the now noncontroversial Johnson & Johnson since its talc litigation has become much less of an overhang on the stock, which wants to gallop to $160. Speaking of gallop, companies controlling the memory apparatus remain constrained, causing their stocks to race to new heights. However, with the exception of Micron, these companies are free riders: they simply aren’t spending enough to replenish the supply and alleviate the tightness, which makes us reluctant to own them. So let’s leave the world of publicly traded companies and go to the only other institution that’s analogous to the insular world that Big Tech CEOs now find themselves in. Doing math in the National Football League For many years, I have had the privilege of knowing the management of the Philadelphia Eagles, a terrific, well-run company that has won more than its fair share of games and championships, so many that I am actually not upset the team isn’t competing in the Super Bowl LX (60) in Santa Clara, Calif., this evening. But over my last three decades as a season ticket holder, there was an unfortunate stretch when a nearly rogue management team took over that didn’t understand the professional nature of the league. It approached the Eagles as if they were a college team, selecting and drafting players in an erratic way. Management didn’t understand the salary cap, an artificial free-cash-flow rule similar to the standards we set for public companies. That general manager wrecked the team. He was fired for doing so. A new GM comes in, a wonderful, tough man who instinctively understands so much — most of all, the salary cap. I was as excited as a child when I asked him, “Now that we are back, how long before we are very competitive?” I think he didn’t want to let me down, so he said three years. I was heartsick. Three years is a long time to get punished. How was that possible? The math was simple. Year 1, he said, we have to rebuild the offense, especially the offensive line. In Year 2, we fix the defense. Year 3? Add skill players. I was incredulous. How could it take that long? Because, he said, you can’t do it all at once. You have to take time to get it right, and three years is about as fast as it can be done. Of course, we know that some teams can get lucky. For example, it looks like the New England Patriots came right back. But the team took years to rebuild until you saw this year’s product, and as fans, we have no idea how long it really took since the Bill Belichick era (2000-2023) and when this team’s success began. But that turn’s about as quick as I have ever seen, and it’s an anomaly. Now, let’s go back to the world we try to live in. First, like the NFL, we are just fans. We only know what the company’s managers tell us, and they only like to tell us four times a year. There aren’t many press conferences. Lord knows I do my best to get the CEOs to talk, but in the end, I recognize how little power I have over any of these people, no matter how much I say or do. When they do talk, though, they often seem oblivious of their power and their words. Their words: think about that. With a few words, they can destroy tens of billions of dollars of shareholder value. Do they know it? I don’t think they do. They don’t want to obfuscate, but there is a sense of non-deliberate obfuscation, if there is such a thing. What these companies are all trying to do is build a new team within a team, one that has an offense, a defense, and special teams. But like the Eagles, it can’t be done all at once, in a quarter or even two or three quarters. It’s going to take as long as it does to build a championship team because all of the other companies are trying to do the same. However, who is willing to say, “We have to be realistic, it’s going to take three years. Not only that, we have to blow through what you thought was the salary cap to get where we have to go. We have to test our balance sheet, maybe even wreck it, and you need to trust us. If we don’t do it, the others will, and if we let them do it without us, we will be finished.” If not, Amazon can’t further develop Amazon Web Services, such a fabulous business. Google Cloud won’t get to $100 billion. Azure will fall behind Google Cloud. Meta doesn’t even have a cloud; it’s a one-trick pony, the product of incredible isolation from the fan base, which may make them an also-ran. Meanwhile, OpenAI and Anthropic don’t even pretend to have salary caps. The constraints are all there if you read between the lines. Google has its own chips to save money. So does Amazon. They are so proud of that, of saving money, of not being hostage to Nvidia and its CEO Jensen Huang, that they believe they go to the playoffs faster than the Eagles did. But these hyperscalers have it all wrong. Investors don’t care if they build on top of Nvidia. As expensive as Nvidia is, it’s the best. We didn’t care that Microsoft built on top of Intel. We just didn’t want to spend money on hardware. Microsoft easily achieved a much larger market cap than Intel. To these public companies, Nvidia has become Moby Dick. We aren’t interested in investing in Captain Ahab. We own Nvidia, for heaven’s sake. OpenAI and Anthropic are built on Nvidia, and we don’t care anymore about Nvidia’s power more than Intel’s. So as we listen to Amazon drone on, and Alphabet pontificate on how great their chips are, we think, what the heck? Are you setting us back? Are you building an offense? Are you playing defense? And can you at least think about the salary cap? The result? Alphabet and Amazon both sell off after their earnings calls on Wednesday and Thursday. They don’t know the power of their words. It’s not that they think they are smarter than we are; they believe they have outsmarted Nvidia. So what. In the immortal words of Al Davis, longtime owner and general manager of the Las Vegas Raiders: “Just win, baby.” In the meantime, to switch sports for a moment, OpenAI is like a fractious racehorse. It’s rearing and snorting all the way to the gate, and the reports that it is fighting with Jensen only magnify it. The Wall Street Journal recently reported that negotiations between the companies were “on ice” after some at Nvidia expressed doubts about OpenAI’s business model. But Huang told me last week: “There’s no drama involved. Everything’s on track.” One winner, for the moment, is Anthropic, which has worked closely with a whole cohort of executives from all different sectors, trying to figure out how it can create living, breathing agentics that can on board and do IT — two functions that are the domain of ServiceNow , hence that stock’s vicious decline. Anthropic can create agents that perform marketing and perhaps rival Salesforce ‘s agents. We don’t know if Anthropic has a competitive product, but it doesn’t matter. It is now the golden child, and investors take its word even though it hasn’t said it can destroy Salesforce. I did a screen for these stocks that showed they trade at a lower multiple on earnings, not annual recurring revenue (ARR), not cash flow, but actual earnings, than Mondelez or Colgate . Maybe they should. Mondelez has Forever Oreos, and Colgate has a product that can whiten teeth in a week. With Anthropic’s business-to-business model (B2B), anything is magically possible. Offense or defense? Woe be to Meta Platforms , which is now flirting with Merck’s multiple, excluding its lucrative cancer drug Keytruda. Can we really own a company with just an offense (advertising) and no skill players? Do we really want to make up a valuation for WhatsApp? Anyway, why do investors have to do it? Why can’t Meta? Microsoft’s situation has become totally problematic. Here we have the greatest software company of all time, in an era when software is considered antiquated. Could Microsoft be the next Sun Microsystems? Novell? Oh heavens, Computer Associates? It gets worse: Microsoft has an ownership stake in the fractious horse OpenAI, the one we thought would be Secretariat, but now feels like Man o’ War. Still good, but not as great as we thought. Sadly, OpenAI plays us for fools. We do not know how much of Azure is OpenAI, and Microsoft seems loath to tell us, even though it is not an NFL team. Maybe Microsoft was the champ for so long that it forgot it needed to tell investors. We are just season ticket holders, I guess. It gets worse with Microsoft. The open secret is we don’t really like Microsoft, but our companies do because it’s cheap and it runs on inexpensive hardware. We tolerate Microsoft. Don’t you think those NFL players would rather have Apple products on the sidelines with an operating system they know? We haven’t cared all that much because we are lumpenproletariat in the face of Microsoft. But when Microsoft built the Potemkin village that is Copilot and charged for it, well, no thank you. The most painful moment of this earnings season, other than the salary cap dismissal I am about to get to, was Microsoft’s pride in selling 15 million paid Copilot seats, even though it has an installed base of 1.5 billion. I used to love CEO Satya Nadella’s narrative on the call, but this time it seemed like a campaign speech from a candidate without a platform or a portfolio. The usually soothing nature of Amy Hood, one of my favorite CFOs, couldn’t cut through the clutter. We don’t really know what Microsoft is spending, so we can’t figure out if they are building an offense or a defense, and we don’t know if they know what a skill player looks like. Which brings us back to Alphabet and its focused approach versus Amazon and its confused, verging on the erratic, quarterly talk. Let’s start with Amazon. Going into the call, we predicted Amazon planned to spend around $150 billion on artificial intelligence. On the call, we learned it will be more like $200 billion . Amazon proceeded to shed just over $300 billion from its market cap, the most of any hyperscaler. And they deserved to lose every penny of it. Why? One, it was so jarring and unexpected. Two, it will have to borrow money, as the Amazon of old did. Three, management spent more time talking up their crackerjack semiconductors to make sure we knew they weren’t writing huge checks to their once-valued partner, Nvidia, which now feels like the enemy. And, four, it’s not clear what they get for the additional $50 billion in capex? Will they buy a natural gas company? Will they buy 10,000 Caterpillar engines that convert natural gas? Will they try to reopen decommissioned nuclear power plants, even though the point of decommissioning a plant is to render it unusable so it can’t be recommissioned? Is Amazon going to find hydro where it hasn’t been found before? Solar? Oh yeah, it’s not sunny all of the time, and then there is that night factor. And how about the actual data centers? Can you just give Vertiv a call? Is Corning on speed dial? Do we have enough memory? Can we just gloss over that, please? There’s more to the story. Until that conference call, we were all bullish on things like One Medical, Amazon’s terrific health-care service, and the possibility of Whole Foods going all-in as our nation’s grocery store with super-fast delivery. Now I am thinking Walmart is killing them. Maybe Amazon should just spend $70 billion and buy Target . We were falling in love with Amazon Prime. And, of course, that incredible, now turbo-charged Amazon Web Services. Wow, 24% growth! What’s the plan, Stan, with Amazon Leo, its rebranded satellite internet project, formerly known as Kuiper None of it mattered. None of it. I found myself thinking, now what? Maybe, like the denouement of the defrocked-era Eagles, we can get excited again in three years, once the rebuild is finished? Ah, but if that’s the case, do I have to stay a season ticket holder? I wasn’t born in Amazonville. I was born in Philadelphia. I am not a fan of Amazon’s stock if it is a loser. I am, alas, agnostic. Now, let’s consider the far better Alphabet, which said it would spend $175 to $185 billion on capex, up from the $115 to $120 billion we expected. Ugh. But wait. Amazon Web Services had a backlog of $244 billion. Alphabet had $240 billion. Advantage, Alphabet. Alphabet had a surprising 750 million average monthly users for its Gemini app, which was just a few years ago a laughingstock. YouTube is becoming the dominant video product of our time. Low cost, huge user base. An analyst said he spotted a flaw in the story, the slower ad dollars, but a patient Philipp Schindler, the company’s chief business officer, explained that the far more lucrative and sticky subscription revenue was replacing it. Wait, there’s more. Self-driving business Waymo is a smashing success, so far ahead of Tesla that it’s painful. Oh, and there’s little old Google Search, which still holds a near-monopoly in the global search market. 90% So what if they are over the salary cap? You don’t have to wait two years. They are competitive right now. The champs Now, let’s end with some irony. The one stock you don’t need the NFL to explain its stature? Apple . Many have been stymied by Apple’s multiple of 32 times forward earnings estimates. That’s pretty laughable. What has happened to Apple and its installed base of 2.5 billion active device users? It’s managed to get a free ride on hundreds of billions in spending by not having to build out its AI. Why bother? It’s got the best with Alphabet’s Gemini. Apple announced in January that it would partner with Google to power its artificial intelligence features, including a major Siri upgrade expected later this year. Meta is trying to do something on the cheap, which I do not know anything about. Microsoft is so caught up in Copilot and OpenAI, I have no idea what it is doing. Amazon may not be worth as much as Anthropic soon because one is focused and the other is helter-skelter. Alphabet is spending a little more than we thought, but that’s strictly to play offense. And Apple isn’t spending anything because it doesn’t need to. Who’s the winner? The one that was luckier than good, although it did take offering a superior hardware product to win. I’ll take it. I am, after all, an Apple season ticket holder and a fan. (See here for a full list of the stocks in Jim Cramer’s Charitable Trust.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust’s portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.