A $300B Deal and Supercomputers: Oracle Wants a Place at the Top of AI Market

TLDR:

• Oracle’s $300B megadeal with OpenAI ties the future of both companies together: Oracle builds the infrastructure, OpenAI commits to filling it.

• This partnership could elevate Oracle into the “Big Four” of cloud, alongside AWS, Azure, and Google.

• Risks are high on both sides, but the scale of the bet means new opportunities for engineers, developers, consultants, and partners who will be needed to build and sustain OCI’s mega-projects.

---

Oracle is a giant that doesn’t usually make headlines with flashy moves. But last week it stunned Wall Street and the entire tech industry, setting a couple of records in the process. 

According to MarketWatch, no company worth over $500 billion had ever gained more than 25% in a single day. Oracle’s stock, however, soared more than 40% – the biggest single-day surge in its history since the 1990s. The ripple effects didn’t stop there: co-founder Larry Ellison, who owns 41% of the company’s shares, briefly overtook Elon Musk as the world’s richest person, adding $100 billion to his fortune in just one hour. 

The spark behind all this market frenzy was a $300 billion megadeal with OpenAI – a move that may also reshuffle the ranks of companies leading the AI race so far. 

Both players bet big, leaving jaws on the floor with a deal of unprecedented scale that will only start in 2027, as Oracle needs time to build out the massive infrastructure required. In many ways, the two companies have tied their futures together: Oracle commits the capital expenditures (CapEx), with a colossal computing infrastructure to power OpenAI’s next-generation models, providing the hardware, energy, and networks needed to scale. 

 OpenAI, in turn, commits to the huge payments behind the contract, to advancing ever more powerful models, and (perhaps most importantly) to keeping ChatGPT as a market leader so that demand continues to grow. Otherwise, billions of dollars’ worth of compute capacity could sit idle. 

In this article, we’ll dig into what this means for Oracle’s strategy as it tries to pull off a transformation worthy of the history books – aiming to go from underdog to alpha dog in the AI era. 

Oracle’s push to become an aggressive player 

For years, Oracle carried the reputation of a “legacy vendor.” Not because it lacked determination, resources, or brilliant minds who could see things differently. Its niche and focus were simply elsewhere: reliability. 

While AWS, Microsoft, and Google chased headlines with flashy launches year after year, Oracle built its brand (quietly but firmly) on reliability, compliance, and security. Its core customers were banks, telcos, retailers, and governments: environments where uptime, governance, and data protection matter more than hype. 

Oracle’s entry into cloud came later (with its Gen2 infrastructure in 2016) but that was consistent with its pattern: move deliberately, avoid missteps, and prove performance in regulated, high-security settings. Then came AI through enterprise products like Autonomous Database – again, more about solving real business problems than launching shiny new tools. 

But in recent months, something has clearly shifted in the strategy of the Austin-based company. A clear signal of this more aggressive path came with TikTok. Oracle already managed U.S. user data for TikTok under Project Texas, routing activity through its servers in Texas to address security concerns. More recently, news broke that Oracle is set to join a consortium (with Silver Lake and a16z) that could control ~80% of TikTok’s U.S. operations — managing user data and building separate infrastructure under a new U.S.-based entity. 

This, too, is a bold bet – but also a clear message to the market: we have the infrastructure to handle massive user bases and absurd volumes of data. And yes, we’re ready for ChatGPT as well. 

What $300B Buys You in the AI Era 

The headline number still feels hard to process: $300 billion over five years.  

What OpenAI is actually buying is raw compute power on Oracle Cloud Infrastructure (OCI) – and at a scale that has never been seen before. Reports put it at 4.5 to 5 gigawatts of capacity, the equivalent of running about two million Nvidia GPUs in dedicated clusters. Just the hardware alone could cost close to $100 billion at today’s prices (before counting data centers, energy, and cooling). 

What Oracle is building is a network of supercomputers that, strategically, will function as a supercloud: a distributed infrastructure of mega data centers designed to satisfy the hunger of the next-generation models OpenAI brings to market. 

And because building something like this takes years, the deal doesn’t kick in until 2027. By then, Oracle and OpenAI expect to have new mega data centers online under Project Stargate — a $500 billion joint plan with SoftBank to construct unprecedented facilities on U.S. soil. 

For Oracle, it’s a bet the size of a small nation’s GDP; for OpenAI, it’s a lifeline to ensure the compute power keeps up with its ambition to build ever more powerful models edging closer to AGI. 

What Each Side Brings to the Table 

This deal it’s a set of deliverables and risks that lock both companies together: 

• Oracle’s side: build ~5 GW of new capacity by 2027 under Project Stargate; supply millions of GPUs, energy contracts, and high-speed networking. In short, the infrastructure backbone. 

• OpenAI’s side: commit ~$60B a year in spend; ramp usage as larger models roll out; align its custom Broadcom chips with Oracle’s new data centers. In short, the demand engine. 

The trade-off is simple but high stakes: 

• If Oracle slips on build or power, OpenAI stalls. 

• If OpenAI can’t monetize fast enough, Oracle’s $455B backlog risks evaporating. 

Too big to fail? the risks behind the deal 

“Too big to fail” has long been a Wall Street mantra – a way of justifying bailouts or market bets on institutions so large that their collapse would shake the entire system. Oracle’s $300B cloud pact with OpenAI now invites the same question. Both companies are tying their futures together: Oracle depends on OpenAI’s ability to pay, and OpenAI depends on Oracle’s ability to deliver flawless scale. Oracle’s record backlog now depends heavily on a single client, while OpenAI is committing to a cloud bill six times its current revenue. 

But context matters. Oracle is hardly a speculative player. It has built its reputation for decades on reliability, compliance, and execution, serving the U.S government, banks, and global enterprises. Safra Catz, the company’s CEO, recently highlighted that Oracle had signed four multi-billion-dollar contracts, pushing outstanding cloud revenue to $455B — up 359% year over year. 

OpenAI, for its part, may be younger, but it is the clear market leader in generative AI – and it isn’t betting solo. Microsoft remains its core backer, with more than $13B invested since 2019 and deep product integration across Office, Bing, and Azure. 

To secure raw compute, OpenAI signed a $12B, five-year contract with CoreWeave and even took an equity stake in the company. At the same time, it is working with Broadcom on a $10B plan to co-design custom AI chips by 2026–2027, a move aimed at reducing reliance on Nvidia. And through the $500B Project Stargate with Oracle, SoftBank, Nvidia, and Arm, OpenAI is anchoring one of the most ambitious data centers buildouts ever attempted on the country. 

In few words, while OpenAI’s history is relatively short, it has built a strong exoskeleton of investors and partners that solidifies its ability to meet future commitments. 

A sky-high burn rate: will ads become the next revenue model? 

Perhaps the greatest challenge Sam Altman’s company faces is generating consistent profits. By mid-2025, OpenAI reported around $10B in annual recurring revenue (ARR), up sharply from ~$5.5B the year before. But its burn rate is still massive. Analysts estimate it spent nearly $8B in 2025 on operations and infrastructure alone. More importantly, its long-term cash burn forecast has surged – OpenAI now projects to burn roughly $115B through 2029. 

This reality means OpenAI will probably need to expand its revenue streams beyond its current mix of ChatGPT Plus and Enterprise subscriptions, API usage, and enterprise licensing. 

One possibility? Ads. According to The Verge, OpenAI is considering ways to diversify income, and bringing ads to ChatGPT is on the table. Nick Turley, Head of ChatGPT, said he’s “humble enough not to rule it out categorically,” but emphasized they would need to be “very thoughtful and tasteful.” He added that while ChatGPT itself might not become an “ads-y” product, OpenAI could develop other offerings where advertising plays a role — keeping the option open while focusing on the fast-growing subscription model. 

Sam Altman himself has voiced ambivalence. At Harvard Business School he called ads in AI “uniquely unsettling” and a “last resort,” but later softened his stance on OpenAI’s own podcast, saying he’s “not totally against it.” Meanwhile, rivals like xAI’s Grok are already moving to include ads in responses. 

Beyond ads, OpenAI is also testing other monetization angles, such as “Commerce in ChatGPT,” an initiative to earn a cut from purchases triggered by product recommendations. Turley has stressed that the integrity of responses must be preserved — ChatGPT should recommend products independently, without being skewed by affiliate revenue. 

We’ll see where the ball drops. 

Where’s Oracle Targeting to Become One of the Irreplaceables in the AI Market 

Oracle is not a newcomer when it comes to powering AI companies with raw compute. Back in 2023, it sealed a partnership with Cohere, where CEO Aidan Gomez was outspokenly optimistic: “OCI provides the best supercomputing on the face of the planet, plus the database that dramatically improves accuracy and provides an intuitive way to search and understand our models.” 

By late 2024, Oracle struck another deal—this time with Meta, set to kick off in Q3 2025. Larry Ellison framed it as a breakthrough moment: “Oracle continues to win large AI training workloads because we’re faster and less expensive than the other infrastructure clouds.” 

And by mid-2025, Oracle doubled down again, signing additional agreements with both Google and xAI, further stacking its roster of marquee AI customers. 

Even with OpenAI, Oracle is no stranger. The partnership traces back to the summer of 2024, when OpenAI first began running workloads on OCI. By early 2025, the AI giant made it official—diversifying its infrastructure and moving away from relying solely on Microsoft Azure as its exclusive cloud provider. 

All of this points to a broader strategy. Oracle seems to be moving to position itself as one of the irreplaceables in the AI stack, much like Nvidia is with chips. The $300B OpenAI deal is the anchor, but it is not the whole play.  

The logic is simple: if Oracle can deliver nearly 5 GW of compute for OpenAI by 2027, it proves not just scale but reliability. Pulling off a project of that magnitude builds confidence in Oracle’s ability to deliver, no matter the scale, the complexity, or the stakes involved. 

Industry analysts have noted that this could elevate Oracle into a “Big Four” alongside AWS, Azure, and Google. Matt Kimball of Moor Insights said OpenAI choosing Oracle “should bring some confidence” to other enterprise customers, validating that OCI can handle the most demanding clients on earth. He also highlighted that Oracle has been methodical in its data center expansion, ensuring consistency across regions so it can scale up for OpenAI without degrading service for others.  

Oracle’s messaging is that everyone benefits from this AI investment. Tens of billions poured into GPUs and high-speed networks do not only serve OpenAI or Meta. They now will have the perfect proof point: if the most advanced AI workloads in the world can run on OCI, then so can yours. 

And beyond infrastructure, industry watchers suggest there may be another advantage: Oracle could gain priority access to OpenAI’s next-generation models, embedding them into its enterprise products and extending the value of the partnership beyond compute. 

The Ripple Effects: Layoffs and New Opportunities 

In recent weeks, Oracle has gone through turbulence of its own, cutting more than 3,000 jobs worldwide in 2025. These layoffs arrive in the broader context of the post-ZIRP era, where rising interest rates have pushed Big Tech companies from hyper-growth to efficiency mode. Meta, Google, Amazon, and now Oracle have all had to make drastic adjustments once easy capital and moonshot projects became relics of the zero-interest rate years. 

Some analysts argue that this is part of a restructuring meant to free up resources and finance Oracle’s AI ambitions. 

Whatever the reason, one big question looms. Given the enormous scale of these projects, who will be the people configuring and managing all of this inside OCI? 

Because at the same time, new opportunities could open up in areas like cloud infrastructure, AI operations, security and compliance, and enterprise integrations — roles that will be essential to turn Oracle’s $300B bet into reality. 

For engineers, developers, and consultants inside Oracle or working with OCI customers: 

• Infrastructure & Cloud Engineers: to design, configure, and maintain multi-gigawatt data centers. 

• AI/ML Ops Specialists: to manage large-scale training clusters, optimize GPU usage, and ensure efficiency. 

• Security & Compliance Experts: to guarantee governance and data protection at massive scale. 

• Enterprise Integration Developers & Consultants: to connect OCI with ERP, CRM, and AI-driven workflows across industries. 

For partners and the broader ecosystem: 

• Systems Integrators & Consulting Firms: to help enterprises adopt Oracle’s AI cloud capabilities. 

• Talent Providers & Outsourcing Partners: to supply specialized engineers and developers at scale. 

• Independent Software Vendors (ISVs): to build AI-native tools and apps optimized for OCI. 

• Energy & Infrastructure Partners: to meet the unprecedented power and cooling demands of 5 GW facilities. 

If you build it, they will come 

Someone has to build the future. Someone has to wire the servers, configure the clusters, and keep the engines of AI running. Over the next few years, that responsibility could open new opportunities for Oracle professionals and anyone working around OCI. 

Of course, the risks are obvious. But this is capitalism: risk first, results later. Oracle is tying a massive part of its future to a single customer. OpenAI is committing to cloud bills that are multiples of its current revenue. If either side stumbles, the fallout could be enormous. Yet precisely because the stakes are so high, both companies are fully invested in making it work. 

And that is where the opportunity lies. Oracle, with this aggressive strategy, is not just trying to be a reliable vendor. It wants to be one of the irreplaceables. If Oracle can deliver 5 GW of compute for OpenAI by 2027, it will send a powerful message: if the world’s most advanced AI workloads can run on OCI, then so can yours. 

This mega-project will not only demand capital—it will demand talent. Engineers, developers, consultants, and partners will all be needed to configure, optimize, and sustain a cloud infrastructure of this magnitude. 

For the broader Oracle ecosystem, that could mean good news in the medium to long term. Today, the skies may look cloudy, but things could get better for the talented people working around OCI. 

At Inclusion Cloud, as Oracle partners, we are ready for that future. We help companies accelerate their projects with elite, certified resources. Through inMOVE™, our AI-powered recruiting engine, we identify true talent in an era where AI makes it harder than ever to tell if a candidate is skilled or simply leaning on a model to fake it.  

Q&A: