Key Points
Oracle delivered 850 megawatts of AI capacity, containing more than 300,000 GPUs, to customers in its fiscal first quarter alone.
Even after those additions, Oracle said GPU utilization across its fleet was 97.9%.
GPU capacity that came up for renewal was renewed or resold at a 20% premium, and most of those chips are at least 4 years old.
5 Stocks Our Experts Predict Could Double In the Next Year
By submitting your email, you'll also get a free pivot & flow membership. A free daily market overview. You can unsubscribe at any time.
- 10 stocks we like better than Nvidia ›
Oracle (NYSE:ORCL) spent Thursday evening’searnings callputting numbers on something Nvidia (NASDAQ:NVDA) investors usually have to take on faith. The software and cloud giant delivered 850 megawatts of artificial intelligence (AI) capacity to customers in its fiscal first quarter of 2027 (the period ended Aug. 31). That hardware contained more than 300,000 graphics processing units (GPUs). And Oracle’s GPU fleet ran at 97.9% utilization, meaning nearly every chip was busy serving customers.
Most of those chips are Nvidia’s. Oracle’s cluster design for its biggest AI deployments connects hundreds of thousands of Nvidia GPUs, and its largest planned builds target up to 800,000 of them.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a “Double Down” signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same “Total Conviction” signal is flashing for a company 1/100th the size of Nvidia. Continue »
So while Nvidia didnt reportthis week, a company installing its chips by the hundreds of thousands just showed what demand for them looks like.
Image source: Getty Images.
Deliveries nearly tripled, and the fleet stayed full
Underscoring how fast Oracle is building, the 850 megawatts it delivered was almost triple the prior quarter’s total — and 73% of everything it delivered in all of fiscal 2026. The company’s campus in Abilene, Texas, which serves OpenAI, received 131,000 GPUs in the quarter. And revenue followed the hardware: Cloud infrastructure revenue rose 121% year over year to $7.4 billion. That growth is an acceleration from 93% in the fiscal fourth quarter and 77% for fiscal 2026 as a whole.
To me, utilization is the number to hold onto. A fleet can run full because its owner stopped adding capacity. Oracle’s ran 97.9% full while capacity arrived at a record pace — demand absorbed the hardware as fast as Oracle could install it.
Even more, customers keep lining up ahead of the supply. Oracle closed more than $30 billion of new AI contracts in the quarter. Its remaining performance obligations (contracted revenue it hasn’t yet delivered) reached $664 billion, up $209 billion year over year.
Four-year-old chips are repricing higher
“Of all the GPUs that came up for renewal in Q1, that capacity was renewed or resold at a 20% premium to prior contracts,” Oracle co-CEO Clay Magouyrk said on Thursday’s call. “The majority of those GPUs are 4 years or older.”
After all, computing hardware is supposed to get cheaper as it ages. GPUs that old predate Nvidia’s Blackwell generation entirely, and with enough supply around, capacity like that would rent at a discount.
Instead, it repriced 20% higher. Prices tend to move that way when buyers can’t get enough new hardware.
The revenue opportunity Nvidia sees per gigawatt of capacity is climbing a ladder of its own. A gigawatt of its older Hopper systems meant about $18 billion of revenue opportunity for Nvidia, chief financial officer Colette Kress said on the company’s late-Augustearnings call Blackwell runs about $25 billion, and the new Vera Rubin platform is about $40 billion.
Customers paying a premium to keep 4-year-old chips arguably won’t balk at those step-ups.
Does one crowded fleet prove a shortage?
Not on its own. Oracle’s 97.9% covers one operator’s fleet, not the industry’s, and the company doesn’t split the figure by chipmaker.
However, there’s a second caveat. Oracle presells most of this capacity under long-term contracts (that $664 billion backlog). A full fleet partly reflects commitments customers made quarters ago.
But the seller’s disclosures point the same way. Nvidia’s revenue rose 106% year over year to $96.2 billion in its fiscal second quarter of 2027 (the period ended July 26). Data center revenue made up $89 billion of that.
Kress told analysts to expect about 70% revenue growth in fiscal 2028, with supply capping the figure. She also said Nvidia’s compute was running fully utilized across every cloud it serves.
A vendor’s assurance like that is easy to discount. Oracle just published the decimal.
The scale fits, too. At Kress’s figures, Nvidia’s $89 billion data center quarter works out to very roughly 3.5 gigawatts of systems. Oracle alone took delivery of 850 megawatts — about a fourth of that pace. The comparison is loose (megawatts include far more than chips, and Oracle buys from more than one vendor). But it shows how much of this build-out one customer can absorb.
So, does Oracle’s quarter prove a worldwide shortage? Proof is too strong a word for one company’s fleet. Ultimately, though, I think it’s the best customer-side evidence yet that what limits Nvidia is manufacturing capacity rather than appetite.
Nvidia shares trade around $220 as of this writing, about 8% below their 52-week high of $236.54, at about 28 times earnings. For a business whose 4-year-old products are repricing higher, I think that’s a fair price, and I’d buy the stock here. If a future quarter shows Oracle’s old GPUs repricing lower instead of higher, that will be an early sign the shortage is easing. This one showed the opposite.
Should you buy stock in Nvidia right now?
Before you buy stock in Nvidia, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Nvidia wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004… if you invested $1,000 at the time of our recommendation, you’d have $417,413!* Or when Nvidia made this list on April 15, 2005… if you invested $1,000 at the time of our recommendation, you’d have $1,341,294!*
Now, it’s worth noting Stock Advisor’s total average return is 950% — a market-crushing outperformance compared to 210% for the S&P 500. Don’t miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
*Stock Advisor returns as of September 12, 2026.
Daniel Sparks and his clients do not have positions in any of the stocks mentioned. The Motley Fool has positions in and recommends Nvidia and Oracle. The Motley Fool has a disclosure policy.