Don’t Fear The Bend: NVIDIA’s Growth In Context

Why toning down guidance makes sense–and why NVIDIA still outperforms.

KEY TAKEAWAYS

• NVIDIA delivered ~$46.7B revenue and ~$26.4B net income with YoY growth still extraordinary

• A minor <1% data center “miss” vs mean estimate doesn’t change the thesis

• AI is an always-upgrading infrastructure cycle that sustains recurring chip demand

• Growth naturally bends from hockey stick to S-curve long before capacity is reached

• Primary risks are regulatory, execution, and tech–but the moat remains wide

MY HOT TAKES

• Deceleration chatter is a distraction from undeniable cash-flow proof

• Consensus models missed by inches–investors risk overreacting by miles

• We’re early in AI infrastructure–capacity fears are premature

• Moat = silicon + software + ecosystem + supply

• This is not hype arbitrage–it’s fundamentals with velocity

• You can quote me: “Deceleration isn’t doom—it’s physics.”
  
  

Beauty in the eyes of the beholder. Most of us know that New York City’s Brooklyn Bridge is not for sale, though a chap named George C. Parker managed to sell it to a surprising number of gullible investors in the 1920s. The lure of owning that iconic brick-laden bridge was strong enough to get enough folks to take the risk, but the takeaway was still: if it seems too good to be true, it probably isn’t. That is why in business we are taught the famous adage of being weary of anyone trying to sell you a bridge. Parker, for his part in contributing to the body of sage business advice, was ultimately sent to Sing Sing prison for fraud. But his story lives on in many forms.

You would be hard pressed to find an econ or finance professor who does not teach their students to doubt any growth path that looks like a “hockey stick.” I can recall many projects that I worked on in my early days of finance where our forecasts, after creating highly complex models, yielded a hockey stick-like growth path. It was not the goal but rather the result of honest forecasting. Because that hockey stick was so ugly and “unbelievable” we would purposely tone down estimates to make the projects’ growths look more "believable," despite knowing that our initial results were accurate. There is no harm in underpromising, and companies have made it fashionable to do so. Toning down growth prospects to make growth look more nature-like is the way to do it. No straight lines and eventually slowing acceleration.

All that aside, it is common sense that one can only sell so many widgets before demand eventually tails off. Think ice picks. At one time in New York city, almost every home needed to own this basic but necessary tool. Once there was one in every home, the manufacturer came up with a marketing campaign to convince homeowners to own more than one. It worked, but sales growth slowed. The result was what we refer to as a “logistic curve,” or, more commonly, an “S-curve.” Sales start out looking exponential, racing upward like a hockey stick. But over time the pace slows and the curve levels off, settling toward a natural limit, called a carrying capacity. 

Enter Artificial Intelligence and the sun of its solar system NVIDIA. The market leader and poster child of chip solutions which is at the core of the AI revolution has a market cap of over $4 trillion, making it larger than the entire economies of countries like Germany or Japan measured by annual GDP. Its market value is greater than the combined worth of every company in the UK’s FTSE 100. It's that big. How did it get there? Check out this chart then follow me to the close.

This chart shows NVIDIA’s quarterly revenues since 2016. The right-hand side of the chart is the obvious reason for the company’s meteoric stock gains. Folks, these are actual revenues, not projected. Chip solutions are at the core of all AI, and NVIDIA is the market-leading producer of them. Not because the CEO is cool or because of any slick marketing campaign, but because the company makes the best, fastest, and most reliable products available. Demand for the product is so high that the company is taking orders for many quarters in the future. The company's biggest challenge is producing enough chips to keep up with demand without any quality failures, and it has done so. 

Now look back at the chart and see if you can forecast where those bars will be in 4 quarters hence. How about 4 years hence. First, do you recognize that this looks like the "hockey stick” I was talking about. Knowing that, would you simply draw a straight line, continuing that existing trajectory? Actually, if you are a visual person–as I am–you might notice that the growth is not actually straight and has diminished slightly quarter-over-quarter. Now that may not be reassuring to most people, but it shows that even a behemoth like NVIDIA is somewhat subject to the laws of nature. Still, if you look at year-over-year growth, it’s almost mind blowing. So, back to the original question: how long can that year-over-year growth go on unfettered? Obviously not forever, but does that even matter given the opportunity that lies ahead for AI?

Yesterday night, NVIDIA announced its Q2 earnings. It reported revenue of about $46.7 billion, up 56% from last year, with net income soaring to $26.4 billion. Earnings per share beat expectations, driven by record data center sales and strong demand for its new Blackwell AI chips. More specifically, the company beat analysts’ estimates across the board but for 1 notable exception. The mean quarterly Data Center Revenue estimate for the quarter was $41.293 billion, but NVIDIA only managed to book $41.096 billion–a miss by -0.48%. That’s right by less than 1%. Now, I don’t know about you, but for me… well, I am OK with it. Remember, what is typically reported as “estimates” is an average of what 39 analysts are expecting. In other words, 39 smart analysts struggled with that line exercise I described before. Based on what they hear from the company, some of them tailed their line more aggressively than others. In the end, Bloomberg simply took an average of all those eager finance bros and shockingly, NVIDIA came up short–by less than 1%. The range of estimates, by the way, was pretty wide with a low coming in at $39.4 billion and a high of $47.5 billion. So, yeah, it would have been awesome if NVIDIA’s report exceeded the high estimate, but alas it only missed the mean consensus by less than 1%. I am hoping that you are picking up on my tongue-in-cheek description here. NVIDIA had a great quarter, hands down.

AI is still in its early days. Its core infrastructure is still being developed. That core infrastructure is not like a sports arena–built and updated every 25 years. No, once built, there is constant demand to make it faster and more efficient. To do that more chips–faster more efficient ones–are required. Think about your smartphone or computer. It felt so fast when you got it, but over time as the applications got more complicated, the old hardware became quickly outdated. This analogy can be applied to AI infrastructure, which is only at the beginning of its many-years-hence growth cycle. At the core of that is NVIDIA and any of its competitors. There are some, but no one comes even close–NVIDIA is a clear winner with a large headstart and mote.

So what can go wrong? We have established that it is a leader in a rapidly growing market. Well, regulations can hamper growth. We have witnessed that with the company’s H20 chip sales to China being restricted. Despite workarounds being proposed by the administration, the restrictions will ultimately impact the company’s earnings. But wait, the company removed all H20 sales from its models, and it still has the healthy prospects we love. If the company finds a resolution, it is only upside–a put. Also, bear in mind, the company’s competitors will be limited by the same restrictions. Folks, there are really only 2 things that can go wrong with this scenario: a management slip up or a technological failure.

Management will have to continue delivering on the high expectations of its customers not only on delivery of existing orders but on the evolution of the technology. To date, the current products appear to be operating as promised. Big market: ✅, early stages of growth ✅, market leader: ✅, competent management: ✅, high barriers to competition: ✅, technology performance: ✅. Challenges: regulatory. My friends, I think that it is safe to say that a solid investment thesis on NVIDIA still holds. This is no bridge for sale! 

Will it last forever? Of course not, we all know that carrying capacity is out there in the future, but we are not there yet–nowhere near. You wanna sit and wait for it? Be my guest, look around the globe for stocks which offer similar, credible upside opportunities and call me when you find some.

YESTERDAY’S MARKETS

Stocks gained yesterday as optimistic investors awaited NVIDIA's earnings announcement. The optimism was enough to drive the S&P 500 to a fresh high. And to think, just a week ago, the media was talking about an AI Winter. 

NEXT UP

• Annualized Quarterly GDP / Second Estimate (Q2) is expected to be 3.1%, an upward revision from 3.0%, the initial estimate.

• Initial Jobless Claims (August 23rd) is expected to come in at 230k, down slightly from last week’s 235k claims.

• Pending Home Sales (July) may have slipped by -0.2% after declining by -0.8% in June.

• Fed Governor Christopher Wallace will speak today.

• Important earnings today: Victoria’s Secret, Dick’s Sporting Goods, Dollar General, Best Buy, Bath & Body Works, the Gap, Petco, Webull, Affirm, Marvell, Autodesk, Ulta Beauty, and Dell Technologies.

DOWNLOAD MY DAILY CHARTBOOK HER
E 📈