Oracle’s AI empire is running on Nvidia, but the numbers don’t add up

Oracle is fearless when it comes to hedging big bets.

A company is reinventing itself on a massive scale, often by spending a lot of money before competitors know what’s coming. Now, the market is paying close attention to its latest innovation: a big push into artificial intelligence using cutting-edge Nvidia chips.

oracle Partnership with Nvidia seems like a dream come true for Silicon Valley and Wall Street. Oracle has the ability to rent AI compute to customers like OpenAI, and Nvidia is expanding its reach into another part of the enterprise cloud.

All parties involved are working together to build the infrastructure for what Larry Ellison, Oracle’s chief technology officer, calls “the most transformative technology of our time.”

We’re not just buying chips, we’re building the world’s next supercomputer network, Ellison told investors in September. This is a generational shift, and Oracle intends to lead it.

But behind the big promises and multi-billion-dollar commitments, there’s a quieter story that’s not about the hype or the hardware, but the math.

As Oracle gets more serious about its AI goals, people are starting to wonder what the true cost of this future is, and whether even the fastest-growing cloud businesses can avoid the laws of financial gravity.

<em>The partnership between Oracle and Nvidia has caught attention, but its AI cloud margins may be lower than investors think.</em>Image source: Andrew Harnick/Getty Images” loading=”eager” height=”640″ width=”960″ class=”yf-1gfnohs loader”/></div>
</div><figcaption class=

It will never be cheap for Oracle to get into the AI business. Over the past year, the company has spent billions of dollars building GPU supersets using Nvidia hardware.

These same chips are fueling the generative AI boom at companies like OpenAI and Anthropic. This spending spree makes Oracle one of the most important companies in the race for AI infrastructure.

But the race is as expensive as it is fast. This week, a recent report said that Oracle’s AI cloud margins may be much lower than investors think, despite all the talk about hypergrowth.

According to internal numbers in this report, Nvidia made about $900 million last quarter, but only a small portion of that went to the bottom line.

In Oracle’s public financial statements, this is a rounding error. In fact, this shows how expensive it is to build an AI empire.

Even a technology veteran like Oracle finds that scale alone does not guarantee profitability. This is due to the high power needs of data centers, chip shortages, and pressure from customers to lower prices.

Oracle’s message right now is to be patient: growth comes first, then profits. However, in a market accustomed to quick rewards, patience may be the hardest thing to sell.

Stargate is Oracle’s big bet: a multi-site US build with OpenAI and SoftBank that will be able to handle gigawatts of AI capacity and lease it at scale. If Oracle does what it promises to do, the project will not only bring in more money; It would also shift the cost curve that has kept margins low.

The first lever is the scale. Larger groups mean better utilization, easier scheduling, and fewer hours of downtime. These are small changes that add up across thousands of GPUs.

The second lever is the mix. With capacity full, Oracle can sell not only raw computing power, but also software, storage, networking, and managed AI services that make more money.

There is also the ability to purchase items. Buying a lot of Nvidia generations at once can make them more available and reduce the price difference between the list price and the overall cost.

Over time, better integration between Oracle’s database, middleware, and AI tools can raise average revenue per customer without having to hire more people or acquire more powerful tools.

None of this happens by itself. Capex intensity, capacity limits, construction schedules, and issues in the supply chain can slow the ramp. The late slope keeps the margin needle stuck.

But if Stargate achieves its goals, Oracle will finally get what its AI story has been missing: the ability to use its resources more effectively.

Investors are putting Oracle’s AI story on the line to see how far it can grow without making money. The company’s backlog numbers are huge, its partnerships are impressive, and its goals are hard to ignore.

But Wall Street’s patience is wearing thin.

People who study the stock say the real question is not whether Oracle can sell its AI capability, but how long it will take to make money from it at acceptable margins.

Oracle relied on short-term hope: Cloud reservations were up 359% from a year ago, and executives promised usage rates would rise as new clusters came online. But investors have heard this story before, and not every technology company can turn scale into profits.

Some funds have begun moving money from AI stocks with high capital expenditures to software stocks with less debt. This dynamic puts Oracle in a difficult position: It must prove that AI revenues can grow faster than AI costs.

The risks are high. If Stargate succeeds, Oracle will be an essential part of the world’s AI infrastructure. If not, people may remember it as the company that helped build the future but never benefited from it.

This story was originally reported by The Street On October 12, 2025, he first appeared in Investment news and strategies to divide. Add TheStreet as Favorite source by clicking here.

https://media.zenfs.com/en/thestreet_881/1c7f21acdc1fbb231bdefead61806ea3

Source link

Sunday Night Football: How to watch the Lions and Chiefs game tonight

Meta and Fortnite are pushing AI video game characters into the mainstream

Leave a Comment Cancel reply