Nvidia stock gained $184 billion in a day, vaulting it past Tesla. But a ...

Investors had already incorporated a giant premium for that grand vision in Nvidia’s market cap. What triggered this the May 25 takeoff was a huge lift in guidance that investors took as proof that Nvidia is already beginning to ride A.I. to one of the fastest profit explosions ever. Kress forecast that revenues will jump by 52% to $11 billion in Q3, suggesting that Nvidia’s growth has hit an entirely new gear. That news sent shares skywards 24% by mid-afternoon on May 25. In a single session, Nvidia’s market cap swelled from $755 billion to $939 billion, or $184 billion. It’s now worth two-thirds more than Tesla. Only Apple and Amazon and Microsoft have ever posted higher dollar gains in a single day, and each by a “puny” extra $7 billion at $191 billion.

While the surge proved a rare windfall for anyone who already owns Nvidia’s stock, it makes it much harder for anyone buying the stock at these prices to get a decent return. What looked tough on May 24 looks virtually impossible on May 25. To understand why, let’s examine how much Nvidia must earn a decade hence to justify that current valuation knocking at the portals of the $1 trillion club.

Keep in mind that the gap between Nvidia and the trillion club it’s nearly joined is that the other members, Apple, Microsoft, Amazon and Google, have already achieved giant earnings that have pushed down their PE ratios. Apple, for example, sports a multiple of 29, and looks extremely rich. Even assuming Nvidia hits its Q2 goal of growing revenues over 50%, it will still be selling at a stratospheric PE approaching 80.

Let’s assume that investors will want a return of at least 10% a year to bet on Nvidia. It’s a risky stock, to put it mildly. The A.I. excitement must play out big time for Nvidia to even have a shot at the kind of success the market’s built into its price. Nvidia pays a tiny dividend, but for this analysis, we’ll assume it reinvests all profits and that all returns flow from capital gains, in other words, a rise in its stock price. If shares meet our target by waxing 10% a year through the spring of 2033, Nvidia will boast a valuation of $2.5 trillion.

What’s a reasonable PE ten years from now? We’ll say 20, which assumes that Nvidia’s profit growth will exceed its cost of capital, that it will reaming in a high-octane phase, well beyond our decade-long window. In that scenario, Nvidia would need to be earning $125 billion a year a decade from now. That’s one-third more than Apple, America’s biggest earner, booked in the past four quarters.

Getting there requires Nvidia to average 27% annual earnings growth over ten-year span. And that requirement assumes it won’t make any big additions to its share count, which could prove necessary to fund the new manufacturing facilities needed to drive its expansion in AI.

David Trainer, founder of investment research firm New Constructs, ran a model that showing that indeed, Nvidia needs to conquer a steep mountain, racing at top, unwavering speed, to reward investors. Trainer reckons that Nvidia’s a good buy providing it can raise its current operating profit after tax from the current 27% to almost 45%, and increase revenues by 20% a annually for the next twenty years.

How likely does Trainer rate that sorcerous performance? “You’ve heard of priced for perfection,” he told Fortune. “This is priced for fantasy.” Of course, we’ve seen several tech champions pull off something like the feat Nvidia much achieve. The problem: Nvidia’s already over 80% as expensive as Amazon and half-again as pricey as former trillion club member Meta. Nvidia can still be a highly successful company, but a poor investment because such Homeric expectations are already baked into its price. Best to take a deep breath, avoid getting drunk on the vision and get sober on the numbers. Otherwise, a wicked hangover awaits.

Subscribe to Well Adjusted, our newsletter full of simple strategies to work smarter and live better, from the Fortune Well team. Sign up today.