MW Oracle's not-so-secret weapon that's helped it quickly take on the cloud giants

By Britney Nguyen

Larry Ellison, now Oracle's CTO, has played a major role in helping the company win AI cloud deals and pivot beyond legacy software

Oracle projects that its Oracle Cloud Infrastructure business will reach $144 billion in revenue by the end of the decade.

When Oracle Corp. co-founder Larry Ellison said a year ago that he envisioned the company's data-center footprint ballooning to 2,000 facilities - from 162 - investors perhaps should have sensed that a big shift was about to take place.

To Kevin Cook, a senior stock strategist at Zacks Investment Research, the comments were a clue that Oracle $(ORCL)$ was getting serious about expanding well beyond database software. The company had been operating its Oracle Cloud Infrastructure (OCI) business since 2016, but over the last year it's made a huge push to take on the hyperscalers that have dominated the cloud-computing market for years - and taken advantage of soaring demand for artificial-intelligence infrastructure.

"We can sort of call that the 'Oracle prophecy,'" Cook told MarketWatch.

And at the heart of the efforts is Ellison himself. At 81, he's still relentlessly involved in the company he founded back in 1977 - speaking on earnings calls, taking the stage at events and using his Silicon Valley fame to help get deals done.

Ellison, now Oracle's chief technology officer and chairman, has been "a master salesman" for the company, according to Ben Reitzes, head of technology research at Melius Research, and his connections and respect in the industry have helped Oracle be part of the large deals defining the current AI era.

Now investors are set to see the fruits of these efforts, as Oracle's latest projections suggest big growth ahead. The company announced this week that it expects OCI to reach $144 billion in revenue by the end of the decade, versus just $10.2 billion for its most recent fiscal year.

Oracle's estimates imply the potential for a roughly 70% compound annual growth rate over five years, Jefferies analyst Brent Thill noted in a postearnings note. The company's shares ended this past week 25.5% higher, booking their best weekly performance since 1999.

Underpinning the revenue projections are four multibillion-dollar contracts signed in the latest quarter, which sent Oracle's backlog swelling to $455 billion worth of contracted revenue yet to be recognized. The key step, of course, is translating those obligations into actual revenue - but Cook said that Ellison, who's been in leadership roles at Oracle since founding the business nearly half a century ago, wouldn't share numbers unless he and other executives really believed they could meet those revenue targets.

To some analysts, Oracle's big, expensive bet on the cloud is a shrewd attempt at diversification as the traditional software industry gears up for a potentially existential battle against AI offerings. The model, perhaps, is Microsoft Corp. $(MSFT)$, which has turned its cloud business into such a powerhouse that its trademark Windows and Office offerings are now sideshows in the minds of investors.

But the push also comes at a cost: Oracle just projected $35 billion in capital expenditures for the current fiscal year, up from its $25 billion estimate given just three months back. That comes as the return on AI capital spending "remains uncertain today," Bank of America analyst Brad Sills wrote this week. And the company has been winning a lot of business with OpenAI, which is still trying to make the economics of generative AI work.

Following the data

Oracle is perhaps best known for its on-premesis software, which includes database tools and enterprise resource-planning systems. The creation of OCI in 2016 allowed the company to satisfy customers' increasing need to do something new with the massive amounts of data that were being collected, Cook said.

The foray was yet another example of Oracle's ability to stay at the cutting edge while other technology lions of the past failed to keep pace with the latest trends. Even when Oracle was growing in the single digits in the early 2000s, the company was releasing "unique and innovative technology," Guggenheim Partners senior managing director John DiFucci told MarketWatch.

For instance, Oracle introduced its real application clusters, or RAC, in 2001, which allowed multiple servers to access one database at the same time - making it possible to scale up to run mission-critical workloads with better performance and at a lower price. And RAC can be used beyond databases, including with cloud infrastructure.

After the launch of OCI, Oracle focused on moving its customers and their enterprise workloads to the cloud, and started offering more cloud-based services, Cook said, especially as corporations needed to access data securely and remotely.

"You could trust Oracle to host all that stuff," Cook said. "That's really where Oracle became a trusted partner."

Oracle's technology took time to gain traction, DiFucci said - but when the COVID-19 pandemic prompted a huge shift to remote work and Zoom Communications Inc. (ZM) needed more computing capacity, Oracle's platform could support those needs.

The platform gained more attention because it excelled at deploying high-performance computing workloads and network-intensive applications such as Zoom video calls, and Oracle's architecture gave it distinct advantages on cost and performance to its competitors, Derrick Wood, a managing director at TD Cowen, told MarketWatch.

How Oracle stands out

Oracle has various perceived advantages in the cloud, including when it comes to pricing and technological architecture. But Ellison himself might just be one of its biggest selling points.

Cook thinks Ellison's iconic status in Silicon Valley gives the company a leg up, especially when compared to Amazon.com Inc.'s (AMZN) Amazon Web Services, Alphabet Inc.'s $(GOOGL)$ $(GOOG)$ Google Cloud and Microsoft's Azure, where "there's not really one person that is the go-to trusted guy that you've always known." Oracle has that with Ellison.

The company has also been able to differentiate itself strategically from the hyperscalers, which often ink multiyear, multimillion-dollar contracts with enterprises for cloud services. When a hyperscaler knows it can reach a certain capacity threshold for a data center in a certain region, it will build one, DiFucci said. The hyperscalers then expect demand to come for the rest of that capacity - a bet that has so far paid off.

Oracle, on the other hand, can scale down to a smaller footprint that serves as a dedicated private cloud for a company at an efficient cost, Wood said, while the hyperscalers build football-field-sized data centers that are architected for very large footprints.

"Oracle architected for very modular footprints, and [it] enabled them to go into these new types of opportunities beyond the kind of general cloud businesses that Azure and AWS were serving," Wood said.

On the company's earnings call earlier this week, Ellison mentioned an OCI instance called "butterfly," which can scale every feature and function of Oracle Cloud into three racks of servers.

"We can give you a private version of the Oracle Cloud with every security feature, every function, everything we do for $6 million," Ellison said. "I think the cost for the other hyperscalers is more than 100 times that."

Customers that have commitments to other cloud providers can also move those workloads to run on Oracle's cloud, DiFucci said. It also works in favor for both Oracle and its competitors because Oracle gets paid by the hyperscaler, which then doesn't have to run certain workloads.

The most demanding workloads right now are for AI training, and whichever cloud vendor can provide higher performance at a lower cost will win there, DiFucci said. To him, that's Oracle.

"They've become the de facto standard for AI training," DiFucci said, adding that the company's ability to run more intense workloads at a lower price than competitors should be attractive to companies that also need to run traditional workloads.

Now, demand is shifting from training to inferencing, or actually running AI models. On the company's earnings call, Ellison said he expects the market for AI inferencing "will be much, much larger than the AI-training market."

Wood thinks that Oracle's infrastructure and expertise in the world of AI training will lend itself well to the AI-inferencing boom.

Yet Oracle's recently disclosed deals raise questions on how it will afford to meet all of the demand for its services.

The cost of growth

In a note after the earnings release, Reitzes said he thinks Oracle's growth "comes at a price," as its free cash flow is in the negative and will take a hit as it funds its cloud efforts. But he doesn't see this as an issue for Oracle right now.

D.A. Davidson analyst Gil Luria also noted concerns for Oracle after its earnings report - writing that, at best, the company's cloud is operating at a single-digit operating margin. That compares with the 20.7% operating margin Google Cloud reported in July for its second-quarter. He added that Oracle could be taking a loss from all the computing it is dishing out.

The opportunity "presents a drastic shift" for Oracle, Luria said, noting that the company has historically operated its legacy business at a 50% operating margin. Luria said the pivot "could hinder earnings growth to single digits, even if revenue growth reaches the midteens."

What comes next for Oracle?

What does the future look like for Oracle? Perhaps Microsoft offers a model.

MW Oracle's not-so-secret weapon that's helped it quickly take on the cloud giants

By Britney Nguyen

Larry Ellison, now Oracle's CTO, has played a major role in helping the company win AI cloud deals and pivot beyond legacy software

Oracle projects that its Oracle Cloud Infrastructure business will reach $144 billion in revenue by the end of the decade.

When Oracle Corp. co-founder Larry Ellison said a year ago that he envisioned the company's data-center footprint ballooning to 2,000 facilities - from 162 - investors perhaps should have sensed that a big shift was about to take place.

To Kevin Cook, a senior stock strategist at Zacks Investment Research, the comments were a clue that Oracle (ORCL) was getting serious about expanding well beyond database software. The company had been operating its Oracle Cloud Infrastructure (OCI) business since 2016, but over the last year it's made a huge push to take on the hyperscalers that have dominated the cloud-computing market for years - and taken advantage of soaring demand for artificial-intelligence infrastructure.

"We can sort of call that the 'Oracle prophecy,'" Cook told MarketWatch.

And at the heart of the efforts is Ellison himself. At 81, he's still relentlessly involved in the company he founded back in 1977 - speaking on earnings calls, taking the stage at events and using his Silicon Valley fame to help get deals done.

Ellison, now Oracle's chief technology officer and chairman, has been "a master salesman" for the company, according to Ben Reitzes, head of technology research at Melius Research, and his connections and respect in the industry have helped Oracle be part of the large deals defining the current AI era.

Now investors are set to see the fruits of these efforts, as Oracle's latest projections suggest big growth ahead. The company announced this week that it expects OCI to reach $144 billion in revenue by the end of the decade, versus just $10.2 billion for its most recent fiscal year.

Oracle's estimates imply the potential for a roughly 70% compound annual growth rate over five years, Jefferies analyst Brent Thill noted in a postearnings note. The company's shares ended this past week 25.5% higher, booking their best weekly performance since 1999.

Underpinning the revenue projections are four multibillion-dollar contracts signed in the latest quarter, which sent Oracle's backlog swelling to $455 billion worth of contracted revenue yet to be recognized. The key step, of course, is translating those obligations into actual revenue - but Cook said that Ellison, who's been in leadership roles at Oracle since founding the business nearly half a century ago, wouldn't share numbers unless he and other executives really believed they could meet those revenue targets.

To some analysts, Oracle's big, expensive bet on the cloud is a shrewd attempt at diversification as the traditional software industry gears up for a potentially existential battle against AI offerings. The model, perhaps, is Microsoft Corp. (MSFT), which has turned its cloud business into such a powerhouse that its trademark Windows and Office offerings are now sideshows in the minds of investors.

But the push also comes at a cost: Oracle just projected $35 billion in capital expenditures for the current fiscal year, up from its $25 billion estimate given just three months back. That comes as the return on AI capital spending "remains uncertain today," Bank of America analyst Brad Sills wrote this week. And the company has been winning a lot of business with OpenAI, which is still trying to make the economics of generative AI work.

Following the data

Oracle is perhaps best known for its on-premesis software, which includes database tools and enterprise resource-planning systems. The creation of OCI in 2016 allowed the company to satisfy customers' increasing need to do something new with the massive amounts of data that were being collected, Cook said.

The foray was yet another example of Oracle's ability to stay at the cutting edge while other technology lions of the past failed to keep pace with the latest trends. Even when Oracle was growing in the single digits in the early 2000s, the company was releasing "unique and innovative technology," Guggenheim Partners senior managing director John DiFucci told MarketWatch.

For instance, Oracle introduced its real application clusters, or RAC, in 2001, which allowed multiple servers to access one database at the same time - making it possible to scale up to run mission-critical workloads with better performance and at a lower price. And RAC can be used beyond databases, including with cloud infrastructure.

After the launch of OCI, Oracle focused on moving its customers and their enterprise workloads to the cloud, and started offering more cloud-based services, Cook said, especially as corporations needed to access data securely and remotely.

"You could trust Oracle to host all that stuff," Cook said. "That's really where Oracle became a trusted partner."

Oracle's technology took time to gain traction, DiFucci said - but when the COVID-19 pandemic prompted a huge shift to remote work and Zoom Communications Inc. (ZM) needed more computing capacity, Oracle's platform could support those needs.

The platform gained more attention because it excelled at deploying high-performance computing workloads and network-intensive applications such as Zoom video calls, and Oracle's architecture gave it distinct advantages on cost and performance to its competitors, Derrick Wood, a managing director at TD Cowen, told MarketWatch.

How Oracle stands out

Oracle has various perceived advantages in the cloud, including when it comes to pricing and technological architecture. But Ellison himself might just be one of its biggest selling points.

Cook thinks Ellison's iconic status in Silicon Valley gives the company a leg up, especially when compared to Amazon.com Inc.'s (AMZN) Amazon Web Services, Alphabet Inc.'s (GOOGL) (GOOG) Google Cloud and Microsoft's Azure, where "there's not really one person that is the go-to trusted guy that you've always known." Oracle has that with Ellison.

The company has also been able to differentiate itself strategically from the hyperscalers, which often ink multiyear, multimillion-dollar contracts with enterprises for cloud services. When a hyperscaler knows it can reach a certain capacity threshold for a data center in a certain region, it will build one, DiFucci said. The hyperscalers then expect demand to come for the rest of that capacity - a bet that has so far paid off.

Oracle, on the other hand, can scale down to a smaller footprint that serves as a dedicated private cloud for a company at an efficient cost, Wood said, while the hyperscalers build football-field-sized data centers that are architected for very large footprints.

"Oracle architected for very modular footprints, and [it] enabled them to go into these new types of opportunities beyond the kind of general cloud businesses that Azure and AWS were serving," Wood said.

On the company's earnings call earlier this week, Ellison mentioned an OCI instance called "butterfly," which can scale every feature and function of Oracle Cloud into three racks of servers.

"We can give you a private version of the Oracle Cloud with every security feature, every function, everything we do for $6 million," Ellison said. "I think the cost for the other hyperscalers is more than 100 times that."

Customers that have commitments to other cloud providers can also move those workloads to run on Oracle's cloud, DiFucci said. It also works in favor for both Oracle and its competitors because Oracle gets paid by the hyperscaler, which then doesn't have to run certain workloads.

The most demanding workloads right now are for AI training, and whichever cloud vendor can provide higher performance at a lower cost will win there, DiFucci said. To him, that's Oracle.

"They've become the de facto standard for AI training," DiFucci said, adding that the company's ability to run more intense workloads at a lower price than competitors should be attractive to companies that also need to run traditional workloads.

Now, demand is shifting from training to inferencing, or actually running AI models. On the company's earnings call, Ellison said he expects the market for AI inferencing "will be much, much larger than the AI-training market."

Wood thinks that Oracle's infrastructure and expertise in the world of AI training will lend itself well to the AI-inferencing boom.

Yet Oracle's recently disclosed deals raise questions on how it will afford to meet all of the demand for its services.

The cost of growth

In a note after the earnings release, Reitzes said he thinks Oracle's growth "comes at a price," as its free cash flow is in the negative and will take a hit as it funds its cloud efforts. But he doesn't see this as an issue for Oracle right now.

D.A. Davidson analyst Gil Luria also noted concerns for Oracle after its earnings report - writing that, at best, the company's cloud is operating at a single-digit operating margin. That compares with the 20.7% operating margin Google Cloud reported in July for its second-quarter. He added that Oracle could be taking a loss from all the computing it is dishing out.

The opportunity "presents a drastic shift" for Oracle, Luria said, noting that the company has historically operated its legacy business at a 50% operating margin. Luria said the pivot "could hinder earnings growth to single digits, even if revenue growth reaches the midteens."

What comes next for Oracle?

What does the future look like for Oracle? Perhaps Microsoft offers a model.

(MORE TO FOLLOW) Dow Jones Newswires

September 13, 2025 09:00 ET (13:00 GMT)

MW Oracle's not-so-secret weapon that's helped it -2-

Bernstein analyst Mark Moerdler has seen shades of Microsoft in Oracle's actions for years. Back in 2023, he wrote that Oracle's transition to becoming a cloud provider reminded him of Microsoft's shift in 2015. At the time, Oracle was moving its ERP system to the cloud, and Moerdler noted that it had a nascent growth opportunity in OCI Gen2 that was similar to what Microsoft's Azure went through.

Meanwhile, there's a broader shift happening, according to Melius's Reitzes, who has been writing for more than a year that software-as-a-service companies are facing increasing challenges from AI.

"The clear winners here are those who make the pipes and the clouds that offer tools to make the applications," Reitzes said in an April 2024 note. The following month, Reitzes cautioned that AI could become "a disruptor to traditional databases as unstructured data rises in both importance and usability," adding that Oracle wasn't immune from the threat.

But Microsoft and Oracle aren't just software companies - critically, they're "'software' companies with clouds," Reitzes said this August. Microsoft paved the way for that sort of diversification, and Oracle is quickly following suit. In his view, that sets the company up to better withstand AI disruption than Adobe Inc. $(ADBE)$, Salesforce Inc. (CRM) and Workday Inc. (WDAY), which are more singularly focused on software-as-a-service offerings.

-Britney Nguyen

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

September 13, 2025 09:00 ET (13:00 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.