Last night, chipmaker NVIDIA announced quarterly earnings which came in slightly below Wall Street’s very lofty expectations.
The company beat earnings and sales expectations for the quarter, but fell short in a couple of areas, specifically Q2 data center revenue, which missed by a hair. The company also did not raise guidance (something Wall Street has viewed as routine), instead saying that it expects Q3 revenue to fall broadly in line with
street estimates of around $54 billion USD, raising concerns that the AI spending boom may be decelerating.
Jensen Huang struck an upbeat tone on the earnings call that followed, saying the “AI race is on” and citing tremendous demand for its next generation of chips. Notably, the company removed any China revenue from its forecast, citing uncertainty, and said that with China revenue, they would have exceeded expectations, with analysts citing anywhere from $2-5 billion in additional quarterly revenue. As this became clearer, the stock recovered after market before opening up a few basis points today.
The stock is now down less than 1% as we publish this newsletter.
Quarter-to-quarter, NVIDIA finds itself in somewhat of a bind: the market has come to expect stellar performance from the company, for good reason. NVIDIA has beaten expectations by an average of $1 billion in each of the last few quarters, and yet despite consistently exceeding estimates, the stock has typically traded
down more often than
up. Alas, this time was no different.
Zooming out, what can we glean from these results? As I first wrote
last year, NVIDIA has become the hardware supplier to the AI economy and a leading catalyst of one of the biggest infrastructure buildouts in human history. These results tell us that NVIDIA is still the king – for now. And with a $4.4 trillion market cap, NVIDIA remains the largest public company in the world. It is larger than all of Germany’s public companies combined. It represents 8% of the S&P500 and 10% of the NASDAQ indices. Indeed, U.S. stock indices are at or near all time highs, driven by the performance of NVIDIA and other AI-related stocks.
So if NVIDIA sneezes, the market will catch a cold. For now, the patient appears healthy enough.
The AI buildout is transforming whole industries, reshaping the power grid, fueling a hiring bonanza with eye-watering payouts, raising the fortunes of communities benefiting from the data center boom (and provoking the ire of some residents), changing the makeup of the labour market, spilling over to the geopolitical fight between the U.S. and China over tech supremacy, and driving economic growth in ways that few could have predicted less than three years ago, when Chat-GPT burst onto the scene.
For example, after decades of low-to-no growth, power demand
surged by 4.3% last year (more than double the annual average rate of the past decade) and power utilities are responding to demand. Dominion will
spend $50 billion to expand Virginia’s grid. And American Electric Power is courting hyperscalers across Texas and Ohio. In Canada, Hydro One and Brookfield are trying to capture rising demand, with Brookfield
already signing a 10.5 gigawatt renewable contract with Microsoft. The rise of AI, like the data boom before it, necessitates the build-out of
sovereign warehouses for sensitive data. Canada increasingly understands this, which means more demand for compute and storage, and ultimately power locally. Other countries – with less abundant natural resources and clean power- will follow-suit with unpredictable impacts on the grid and the climate.
Private equity players are now some of the biggest investors in the sector. Blackstone
acquired AirTrunk for $16 billion and
launched a $25 billion joint venture with PPL to build data-center and power infrastructure. KKR and ECP
created a $50 billion platform for AI campuses and later a $25 billion
JV with ADQ. DigitalBridge and Silver Lake
backed Vantage Data Centers with $9.2 billion, while Apollo
bought into Stream Data Centers. Pimco and Blue Owl
arranged $29 billion in financing for Meta’s Louisiana buildout. As I noted
earlier this year, there is a rush to driven to own the infrastructure of the AI era.
If you’re getting desensitized to these big numbers - $25 billion here…$50 billion there - well here’s an even
bigger one: Consulting firm
McKinsey estimates that:
“By 2030, data centers are projected to require
$6.7 trillion
worldwide to keep pace with the demand for compute power. Data centers equipped to handle AI processing loads are projected to require $5.2 trillion in capital expenditures, while those powering traditional IT applications are projected to require $1.5 trillion in capital expenditures. Overall, that’s nearly $7 trillion in capital outlays needed by 2030—a staggering number by any measure.
This kind of spending is bound to impact and reshape the economy. According to the recent GDP figures,
AI data center buildout contributed more to economic growth than the entire U.S. consumer. But some fear that AI spending is more of a sugar rush than a thanksgiving feast - construction adds a short-term jolt to GDP and jobs, but once operational, datacenters require relatively few staff.
Some skeptics are warning of a bubble. Tech now accounts for 34 percent of the S&P 500 and valuations for some leaders are nearing dot-com levels. And the success of the market is really the success of the AI trade. Everything hinges on it. As Derek Thompson wrote recently, “in the last two years,
about 60 percent of the stock market’s growth has come from AI-related companies, such as Microsoft, Nvidia, and Meta. Without the AI boom, stock market returns would be putrid.”
NVIDIA’s results show that demand remains high and capital continues to flow into the AI infrastructure for this new digital age. As
we have said before, the AI future is already here. And with NVIDIA supplying the world its cutting edge GPUs, it is now more distributed than ever.
August 28, 2025
