By August of last year, Nvidia’s stock was substantially overvalued by most metrics. Now? The problem has nearly tripled. Nvidia’s market cap had risen threefold to more than $1.2 trillion since the start of 2023, to notch the biggest short-term valuation jump in the history of capital markets, a rampage that drove its P/E multiple, based on the past four quarters of GAAP earnings, to well over 100.

Back in those early days last August, Fortune’s analysis argued that to deliver its investors decent returns over the next seven years, Nvidia would need to raise its earnings at an annual rate of over 22%, a number that looked virtually impossible, given an inevitable shrinkage in its gigantic margins, and the onset of rivals bent on invading the most lucrative, and fastest-growing market in the tech universe. David Trainer, founder and CEO of research firm New Constructs and arguably the best mind on Wall Street for assessing companies’ worth based on fundamentals, confirmed my view, stating: “Nvidia’s valuation is ridiculous. It’s facing the same curse as Tesla. But when Tesla got profitable, loads of competitors entered the EV space, cutting margins and slowing sales. The same will happen with Nvidia.”

Nvidia’s 2024 explosion dwarfs last year’s meteoric rise

From July of 2023, a month before that story ran, to early January of 2024, Nvidia’s shares plateaued, hovering mainly between $46 and $49. Since profits kept rising, its multiple dropped below 60, still a huge number, but hinting that if the earnings trajectory remained on course, and the valuation fell or even flatlined, the feats required to reward shareholders could shrink from the virtually impossible to the highly challenging but plausible.

Blockbuster earnings for its Q4 of FY 2024 (its fiscal year ended Jan. 31), announced on Feb. 21, ignited the first stage for a moonshot that’s far outstripped the lightning run in the first half of last year. On May 22, CEO Jensen Huang added the booster rocket, unveiling Q1 profits of $14.9 billion, a 21% increase over the awesome numbers reported just three months earlier.

Huang further cheered Wall Street by announcing a 10-for-one stock split for June, and a 150% increase in the dividend. Nvidia’s next-gen GPU chip, Huang declared, will start generating revenues later this year, and “is going to be terrific,” its introduction marking “the next wave of growth.” Analysts garlanded the results by gushing that that there’s “no match for Nvidia’s products,” and that buying its shares amounts to a “no-brainer,” lauding its Q1 performance as “perfect.” Typical of the overwhelmingly rosy reception: an assertion that the only sure things in life are “death, taxes, and Nvidia beats on earnings.”

From early January through midday on June 12, Nvidia’s value famously jumped by 160% to $3.10 trillion, as its shares surged from $48 to $125, adding an incredible $1.90 trillion in market cap, more than twice the dollar liftoff over the first six months last year. In other words, it was the second near-tripling in valuation in only 18 months. The stock’s astounding run added 210% of the total valuation of Berkshire Hathaway. In 2024, Nvidia has leapfrogged Alphabet and Amazon to stand as the world’s third most valuable company, trailing only Apple ($3.30 trillion) and Microsoft ($3.25 trillion).

The rub: Nvidia’s stock sorcery greatly raises the bar for its performance going forward

Nvidia’s never-before-witnessed stock sprint poses a big problem for its current investors in deciding whether to exit and bank the windfall, and folks and funds who ponder purchasing its shares at these prices. We don’t know how Nvidia will perform in the future, but today’s valuation sets in stone the metrics it must meet to reward shareholders. And the hurdles just got far, far bigger.

In both my analysis last August, and in the current model, I posit that Nvidia must garner returns of 10%, at minimum over the next seven years, to enrich shareholders, since it’s a high-risk bet, especially following the astronomical gains over the past few months. We’ll assume that almost all of the future increases will come in capital gains. Despite the big dividend increase, the yield remains minuscule. Nvidia has also made substantial stock repurchases of $27 billion in the past nine quarters, but they’ve done nothing to raise its earnings per share or its price. “For all the big buybacks, the share count remains unchanged,” says accounting expert Jack Ciesielski. “Nvidia’s issuing a lot of shares in its stock plan for employees, and these repurchases are just offsets, so that it’s treading water on the share count.”

As of Q1 2024, Nvidia has garnered profits of $42.6 billion over the past four quarters. Hence, our model—incorporating only modest returns—demands that profits wax at 25% annually over a seven-year span. Of course, the biggest percentage gains would come in the earlier period. Still, my projections show that from mid-2030 to mid-2031, the profit increase over the previous 12 months would total around $30 billion. That’s 40% of the all-in figure for Microsoft and Alphabet over their past two fiscal years.

Achieving yearly gains of 10% doubles the stock price over our seven-year window. In the last story, when Nvidia’s market cap stood at $1.2 trillion, I reckoned that to ring the bell, Nvidia needed to raise profits to $80 billion over the chosen time frame. (I have since made a few changes to the calculations.) So what’s the required target today? Instead of reaching $2.4 trillion (double the then $1.2 trillion valuation), Nvidia must now swell its valuation by mid-2031 to $6.2 trillion (twice its current $3.10 trillion cap). Let’s posit that by then, its P/E falls to around the current Nasdaq average of 30, no small number historically, by the way, and a multiple that would forecast years of strong profit growth beyond 2031. So a bull’s-eye requires amassing net GAAP profits of just over $200 billion seven years hence (the $6.2 trillion projected valuation divided by a P/E of 30).

What are the chances Nvidia could clinch the $200 billion goal? Remote. Adjusted by the yearly inflation of 2.25% forecast by the 10-year Treasury break-even rate, $200 billion equates to $170 billion in today’s dollars. That’s 70% more than Apple earned in its past fiscal year, and double Microsoft’s profits over the past four quarters. A major problem: Nvidia is booking unsustainably fat margins. In FY 2024, it collected almost 50 cents in net profit for every dollar in sales, and in Q1, the take rose to 57 cents. By contrast, as of their last fiscal years, Alphabet is at 24%, Apple 25%, and Microsoft 35%, while super-profitable software giants Oracle, SAP, and Qualcomm respectively registered 17%, 20%, and 21%.

It’s the same story regarding revenues. A reasonable prediction lowers Nvidia’s margins to the still-fantastic Microsoft level of 35% by 2031. In that case, we’d zoom forward to $580 million in sales by the close of our seven-year window. That’s half again what Apple collects on its iPhones, laptops, and sundry products, and almost twice Alphabet’s revenues.

Put simply, the stratospheric performance of Nvidia stock jacks an already elevated bar so high that it’s nearly impossible for even this extraordinary pole-vaulter to clear.

Trainer provides some scary comparisons for the size Nvidia needs to achieve

I consulted Trainer for an update on his take on Nvidia, now that its valuation has far more than doubled since he rated it radically overpriced last year. Trainer deploys cash flow projections that show the heights earnings and revenues must hit to “discount back” totaling the enterprise’s market cap today. “I ran the numbers at $100 a share,” he says, noting that Nvidia’s price is now 20% higher. “To be worth that number, Nvidia would need revenues of $3 trillion, the size of the GDP of India, the most populous country in the world, 15 years from now. Its profits would need to hit $1.1 trillion, over 10 times where Apple is today.”

For Trainer, investors are awarding Nvidia such an inflated valuation “by extrapolating the competitive advantage it has today in perpetuity.” Margins the size of Nvidia‘s don’t last long, he says. “That’s analyst training 101,” he adds. “Anytime a competitor sees margins like that, they jump in and make big investments to grab part of your business. For now, Nvidia can do what no one else can do, but the best competitors will find a way. It will become a race to the bottom.” Never has an enterprise inspired such widespread faith that it can transform the entire world economy. But the market’s belief, and the crazy valuation it’s created, has rendered reaching a golden destination for shareholders a bridge too far.

Looking ahead, Nvidia may be destined for fabulous profitability, growth, and leadership in AI, and still prove an extremely poor investment.