Oracle also moved to secure the silicon supply chain for AI. Unlike Amazon, Google, or Microsoft, Oracle isn’t designing its own AI chips – it has no equivalent of AWS’s Trainium or Google’s TPUs. Instead, Oracle chose to double down on partnerships with GPU vendors:

• Oracle expanded its longstanding alliance with Nvidia, the market leader in AI chips. Oracle offers bare-metal Nvidia GPU instances in the cloud (no virtualization overhead), giving customers dedicated access to top-end chips like the A100 and H100 trefis.com. In fact, Oracle was among the first clouds to deploy Nvidia’s A100s in 2020 semianalysis.com and it now advertises OCI Superclusters that can scale to over 16,000 Nvidia H100 GPUs in one cluster for giant AI training jobs oracle.com. This deep Nvidia collaboration ensures Oracle clients like OpenAI or Meta can get the absolute highest-performance clusters (with fast networking and storage) for training AI models. It’s a mutually beneficial relationship – Oracle’s blowout forecast implied orders for vast numbers of Nvidia GPUs, lifting Nvidia’s stock as investors realized Oracle would be a major GPU buyer in coming years trefis.com trefis.com.
• At the same time, Oracle is not putting all its eggs in one basket. It struck a deal with AMD to offer AMD’s new MI300 series AI accelerators on OCI. In June, Oracle and AMD announced the upcoming Instinct MI300X GPU (CDNA 4 architecture) will be available on Oracle Cloud, promising >2× better price-performance than the prior gen trefis.com. Oracle said it plans to deploy up to 131,072 of these AMD GPUs across its data centers trefis.com. By embracing AMD’s top-tier chips (in addition to Nvidia’s), Oracle both diversifies its supply and gains leverage in pricing negotiations. As Trefis noted, “the move reduces single-supplier risk and should boost Oracle’s bargaining power with Nvidia”, helping keep costs in check as it scales trefis.com. The market recognized this dynamic – after Oracle’s earnings, AMD’s stock climbed on the expectation that it will “capture a meaningful share of Oracle’s AI spend” alongside Nvidia trefis.com trefis.com.

Oracle’s ability to scale infrastructure quickly – thanks to these supply-chain maneuvers – is a core reason it landed customers like OpenAI. Indeed, Microsoft’s Azure (OpenAI’s primary cloud) has faced its own capacity constraints; Microsoft execs admitted they “can’t build infrastructure fast enough” to meet AI demand and expect supply tightness into mid-2026 revolgy.com revolgy.com. Sensing an opportunity, OpenAI opted to diversify by adding Oracle (and reportedly Google) as cloud providers reuters.com. Andy Jassy, Amazon’s CEO, similarly said AWS has “more demand than we have capacity at this point” in AI, as the company scrambles to add data centers and power ciodive.com ciodive.com. This industry-wide crunch – “the single biggest constraint is power”, Jassy noted ciodive.com – left a gap that Oracle seized. By pre-building massive GPU farms and signing long-term leases ahead of demand, Oracle became the one cloud that could immediately slot in a huge AI workload.

In short, Oracle’s stunning $455 billion backlog represents the payoff of an AI capacity land-grab. The company spent big and moved fast to ensure it had the GPUs, data centers, and partners ready for the AI boom. Now, with Oracle Cloud “capacity-constrained but supply-coming online,” Safra Catz says the challenge is simply delivering on these contracts trefis.com trefis.com. Oracle even raised its planned capital expenditures to $35 billion for FY2026 – a 65% jump – to keep building data centers and procuring hardware for this pipeline trefis.com trefis.com. The message is clear: Oracle is “all-in” on AI infrastructure, determined to not let supply limits curtail its historic run. As one Morgan Stanley analyst put it, Oracle’s fundamental business is now shifting to that of a “GPU datacenter operator” crn.com – a far cry from the Oracle of old.

“Truly Historic”: Market Reacts to Oracle’s AI-Fueled Guidance

When Oracle revealed its Q1 FY26 results and forward guidance, it sent shockwaves through Wall Street. On the earnings call, Safra Catz delivered a forecast that seemed almost unbelievable: thanks to the AI deals in backlog, Oracle would accelerate to 77% growth in cloud infrastructure revenue this fiscal year (to about $18 billion), then target $32 billion in FY27, $73 billion in FY28, $114 billion in FY29, and $144 billion by FY2030 investor.oracle.com. For context, Oracle’s total revenue last year was around $50–60 billion; they are now essentially projecting to triple revenue within ~4 years, entirely driven by cloud AI. Most of this, Catz emphasized, is already under contract in the RPO numbers investor.oracle.com investor.oracle.com.

Seasoned analysts on the call openly struggled to process these figures. “Even I’m sort of blown away by what this looks like going forward,” admitted John DiFucci of Guggenheim, calling it a potential “career event” for anyone covering tech investopedia.com. Deutsche Bank’s Brad Zelnick declared, “There’s no better evidence of a seismic shift happening in computing than these results… the backlog growth is truly historic.” investopedia.com. Others described Oracle’s quarter as “momentous” and the AI opportunity as a once-in-a-generation inflection investopedia.com investopedia.com. In research notes, Morgan Stanley dubbed Oracle’s $455 billion bookings number “the biggest we’ve ever seen in software”, remarking that none of the usual concerns (Oracle had missed revenue consensus by ~$100M, for example) “matter” in light of this new trajectory crn.com crn.com. The Futuriom analyst team summed it up: “We have signed significant cloud contracts with the who’s who of AI, including OpenAI, xAI, Meta, NVIDIA…” Safra Catz told them – an absolutely shocking turn for a company once considered an also-ran in cloud futuriom.com.

Investors reacted with euphoria. Oracle’s stock price exploded on the news, rising as much as 43% at one point and closing up ~37% on Sept 6, 2025 reuters.com reuters.com. It was Oracle’s best single-day gain in over 30 years reuters.com, catapulting the shares to all-time highs around $345. With that jump, Oracle – a company often absent from discussions of elite tech stocks – suddenly flirted with a $1 trillion market valuation reuters.com reuters.com. Analysts noted Oracle might soon join the exclusive “trillion-dollar club” alongside Apple, Microsoft, Saudi Aramco, etc., if its stock strength holds reuters.com reuters.com.

The rally in Oracle did more than enrich Oracle’s shareholders; it sent a ripple effect across the AI and cloud sector. Shares of Nvidia and Broadcom, key suppliers of AI chips and networking gear, leaped 4–9% on the day investopedia.com. Chipmakers AMD and Arm also saw a boost investopedia.com. “The rising tide of robust AI spending was lifting plenty of boats,” Investopedia noted, as Oracle’s blockbuster results validated the entire AI ecosystem’s growth story investopedia.com. One exception: Oracle’s major cloud rivals felt the heat – Amazon’s stock actually dipped ~3% and Meta’s ~2% that day investopedia.com, as investors perhaps rotated into Oracle or worried that AWS and others had lost some big deals. Even so, the broad theme was that Oracle’s news confirmed “the AI spending bonanza has ample room to run”, boosting confidence across tech markets investopedia.com.

For Larry Ellison, Oracle’s co-founder and chief technology officer, the stock surge resulted in a personal windfall of epic proportions. Ellison owns about 42% of Oracle’s shares reuters.com, so the nearly 40% jump in ORCL translated to roughly a $100 billion gain in his net worth overnight. By Forbes’ real-time tracker, Ellison’s fortune hit approximately $392 billion reuters.com. Bloomberg’s Billionaires Index went even further, estimating Larry’s wealth at $393 billion, which for the first time edged him past Elon Musk as the richest person on Earth crn.com. (For perspective, Musk’s net worth was around $385–440 billion depending on Tesla’s stock; the methodologies differ, but Ellison was suddenly in the same league, within a hair’s breadth of #1 reuters.com reuters.com.) Bloomberg noted that Ellison’s one-day gain of $101 billion was the largest ever recorded on its index crn.com. This is a staggering sum – more market value created in a day than the entire market cap of Intel or PayPal. It speaks to how dramatically Oracle’s narrative has shifted that Larry Ellison, at 81 years old, briefly became the world’s wealthiest man on the back of AI optimism.

Ellison himself has never been shy about grand visions, and on the earnings call he was in rare form. He crowed about Oracle’s “unprecedented” opportunity in AI and outlined how the company will differentiate. One highlight was Oracle’s plan to launch the “Oracle AI Database” – a new cloud service that embeds generative AI directly into the Oracle Database platform investor.oracle.com. “Next month at Oracle AI World, we will introduce a new Cloud Infrastructure service called the ‘Oracle AI Database’ that enables our customers to use the large language model of their choice – including Google’s Gemini, OpenAI’s ChatGPT, xAI’s Grok, etc. – directly on top of the Oracle Database,” Ellison explained investor.oracle.com. This will let Oracle’s tens of thousands of database customers “easily access and analyze all their existing data” using AI models, effectively bridging the gap between data stored in Oracle software and insights from AI investor.oracle.com. Ellison portrayed this as revolutionary, saying it will “instantly unlock the value in their data” and drive “dramatically increasing cloud demand and consumption over the next several years. AI changes everything,” he declared emphatically investor.oracle.com.

That catchphrase – “AI changes everything” – has become something of a mantra for Oracle’s leadership. It encapsulates why investors bid Oracle shares up 45% year-to-date (outperforming even the likes of Nvidia and Microsoft) reuters.com. Oracle is no longer being valued as a slow-growth legacy software vendor; it’s being valued as an AI-era growth stock. As KeyBanc analysts put it, the skepticism about Oracle’s cloud prospects has been “put to bed” now that the growth is “real and contractually identifiable” in backlog form crn.com. And importantly, many of these AI contracts span 5–10+ year terms (longer than typical 3–5 year software deals), essentially locking in Oracle’s status as a primary infrastructure provider for the AI generation crn.com crn.com.

The market’s about-face on Oracle can be summed up by a striking quote from Melius Research: “In the once mighty software realm, Microsoft and Oracle really aren’t software companies anymore – they are AI cloud infrastructure stocks that happen to sell software,” the analysts wrote crn.com. By the same token, “Google may soon be a cloud infrastructure stock that happens to sell Search,” they quipped crn.com. In other words, Wall Street now sees Oracle through an AI lens. The company’s valuation and narrative are tied to data centers, silicon, and AI workloads – not just databases and ERP apps. This is perhaps the biggest validation Oracle’s cloud pivot could have hoped for.

Cloud Wars: Oracle vs AWS, Azure, Google in the AI Era

Oracle’s resurgence comes as the “cloud wars” enter a new phase centered on AI. How does Oracle stack up against the industry’s giants on key metrics and strategy? Below is a comparative snapshot:

¹ Microsoft does not break out quarterly Azure revenue publicly; $21.5B is an analyst estimate based on Intelligent Cloud segment growth and the disclosed $75B annual Azure run-rate reuters.com reuters.com.
² Microsoft’s $368B performance obligations include Azure and other long-term contracts (like Office 365) fierce-network.com. Azure-specific backlog is not fully disclosed, but CFO Amy Hood said record spending is “correlated to contracted, on-the-books business” demand reuters.com.
³ Google’s cloud segment (including GCP and Workspace) was $13.6B in Q2 2025 with 32% growth revolgy.com. We estimate core GCP infrastructure/platform services at ~$8B. Google doesn’t disclose cloud RPO publicly; $106B is an industry estimate fierce-network.com.

As the table suggests, Oracle’s cloud business is still smaller in absolute revenue – Oracle’s quarterly cloud revenue ($7.2B) is about one-fourth of AWS’s and one-third of Azure’s. Oracle’s share of the global cloud infrastructure market was roughly 3–4% in mid-2025 (vs. AWS ~30%, Azure ~20%, Google ~13%) ciodive.com. However, Oracle is now growing faster than its larger rivals and has amassed a bigger forward pipeline (RPO) than anyone. At ~$455B, Oracle’s backlog of cloud orders exceeds even Microsoft’s total (which was $368B across all cloud/software as of June) fierce-network.com. It’s over 2× the size of AWS’s $195B backlog fierce-network.com. This is a stunning flip of the script – historically AWS enjoyed the largest enterprise commitment pipeline, but Oracle has leaped ahead by locking down multi-year AI deals.

Growth Trends: AWS’s growth had decelerated into the teens, though it ticked up to 17.5% last quarter as AI demand picked up ciodive.com. Microsoft’s Azure re-accelerated to nearly 40% growth with help from AI workloads (its highest growth in many quarters) reuters.com. Google Cloud’s growth (~32%) also surpassed AWS. Oracle’s overall cloud growth was 28% last quarter investor.oracle.com, but specifically OCI (infrastructure) jumped 55% investor.oracle.com. And looking forward, Oracle’s projected ~77% growth in FY26 OCI revenue far outstrips what Amazon or Microsoft are guiding. In effect, Oracle is trying to join the leaders and potentially create a “Big Four” of cloud providers nasdaq.com. “This outlook suggests that the ‘Big Three’ may soon become the Big Four,” observed Motley Fool’s analysts nasdaq.com, given Oracle’s eye-popping forecast.

Backlog & Bookings: The backlog story is a key differentiator. Cloud providers don’t traditionally disclose “RPO” for IaaS, but under new accounting, both AWS and Azure have given some signals. AWS’s ~$195B backlog (as of Q2 2025) grew 25% YoY, reflecting strong long-term commitments tied to AI ciodive.com. Amazon CEO Andy Jassy said AWS’s pipeline is healthy but fulfillment is gated by capacity constraints ciodive.com. Microsoft hasn’t broken out an Azure-only backlog, but CFO Amy Hood explicitly said their record $30B capex spend for the Sep quarter is backed by “contracted, on-the-books” demand that Microsoft needs to deliver reuters.com. This implies Azure’s backlog is also ballooning (in the earnings call, Satya Nadella mentioned Azure AI services usage is growing 2× and they have huge Azure OpenAI Service commitments). Google’s cloud backlog (~$106B) was the smallest of the bunch fierce-network.com, but Google too noted higher bookings related to AI (and subsequently announced big deals like the Meta partnership). Still, Oracle’s $455B eclipses them all. Oracle essentially pulled ahead by being first to sign 10-year, $100B-scale deals in AI. KeyBanc analysts noted Oracle added about $300B of backlog in just three months, roughly five years’ worth of its current annual revenue added in one quarter crn.com crn.com. “The question of whether customers will come is now put to bed,” KeyBanc wrote – the demand is undeniably there crn.com.

Capital Investment: A fierce capacity build-out is underway at all providers, as they race to build data centers and install tens of thousands of GPUs. Amazon is by far the biggest spender in absolute terms – in the first half of 2025 alone, Amazon poured $56 billion into capex (with the “majority” aimed at AWS and AI) ciodive.com. It plans around $100 billion this year, an unprecedented sum even for Amazon ciodive.com. This includes doubling its data center footprint in many regions and buying large quantities of Nvidia H100s (and also developing its own AI chips for cost control). Microsoft, not to be outdone, stunned investors by guiding to $30 billion in capex for just the July–Sept 2025 quarter reuters.com – its largest ever quarterly spend, and above what even Amazon spends per quarter. This suggested Microsoft will spend $90B+ annually at the new run-rate. “Microsoft’s higher-than-expected capex forecast – its largest ever – puts it on track to potentially outspend rivals over the next year,” Reuters noted reuters.com. Google hasn’t given one number, but it said it would “spend more on data centers” for AI and indeed its Q2 capex was up significantly year-over-year reuters.com. Industry data from Synergy Research shows the Big 3 (AWS, Azure, GCP) spent a combined $78–87 billion on capex in Q2 2025 alone revolgy.com revolgy.com – a staggering figure that underscores how critical AI is to their future.

Oracle, while smaller, is also ramping its investment dramatically. Oracle spent about ~$2–3 billion per quarter on capex during 2024 (significantly up from prior years as it expanded OCI regions). Now, Oracle plans to spend $35 billion in fiscal 2026 on capital expenditures trefis.com – roughly double its recent run-rate. This includes building out another 37 cloud data centers to add to Oracle’s existing 34 sites, specifically to serve its new AI hyperscale customers reuters.com investor.oracle.com. In fact, Oracle is collaborating with Microsoft, Amazon, and Google in some cases to add Oracle Cloud racks inside those giants’ data centers (“Oracle Database@Azure” is one example) – a strategy to quickly gain capacity by piggybacking on larger players’ footprints reuters.com investor.oracle.com. Oracle’s willingness to spend aggressively, despite being smaller, shows how high the stakes are. As Microsoft’s CFO noted, these massive capex bets are crucial to meeting demand: “I feel very good that the spend we’re making is correlated to contracted business we need to deliver,” Amy Hood reassured investors reuters.com. Oracle’s Safra Catz could say the same – the $35B Oracle will spend is essentially pre-paid by the likes of OpenAI in the backlog.

AI Partnerships and Ecosystem: The competitive landscape is also defined by who partners with whom in AI:

• Microsoft’s big win was its early OpenAI investment and exclusive cloud tie-up. That partnership supercharged Azure’s AI credibility – by hosting ChatGPT, Azure attracted a wave of AI startups and enterprise customers wanting the same tech. However, that exclusivity is eroding: OpenAI is renegotiating terms with Microsoft amid investor pressure, and has started using other clouds (Google, Oracle) to augment capacity reuters.com. Microsoft is responding by broadening its own ecosystem – hosting Meta’s open-source Llama 2 models, partnering with Elon Musk’s xAI and others to ensure Azure isn’t dependent on just OpenAI reuters.com. Microsoft also offers its Copilot AI assistants across Office and developer tools, which could drive Azure backend usage. Its strategy is to be the AI platform for enterprise, leveraging its software dominance.
• Google’s approach leverages its strength in AI research (DeepMind, Google Brain) and custom silicon (TPUs). Google Cloud offers TPU v5 pods for AI training and has made a point of embracing open AI models: it provides access to models from Anthropic (Google is a major Anthropic investor), Meta’s Llama, and various open-source models on its Vertex AI platform. Google’s marquee partnership is the Meta deal – a bit ironic given they are consumer tech rivals. But as one commentator noted, “a $10B deal shows scale beats rivalry” in the cloud business datacenterdynamics.com cloudwars.com. Google winning Meta’s AI workload (reportedly involving 1.3 million GPUs worth of capacity over time infoworld.com) validated Google Cloud’s seriousness. Google is also reportedly in talks with OpenAI to host some services, which would’ve been unthinkable a couple years ago reuters.com. Clearly, in the AI era, yesterday’s competitors can be today’s clients.
• AWS has been a bit more closed in its model strategy – rather than strike one big exclusive deal, Amazon launched Bedrock, a service that offers various third-party models via API (including startups like AI21, Stability AI, and Anthropic’s Claude). AWS touts that it has the most complete set of AI building blocks, from custom chips (Trainium, Inferentia) to ML ops tools (SageMaker) to those partner models. But AWS did not scoop a headline-grabbing deal with OpenAI or Meta – those went to others. AWS CEO Andy Jassy instead emphasizes AWS’s overall size and reliability: “We have a meaningfully larger cloud business than others… ~ $123B run-rate, the second player is ~65% our size,” he said, subtly pointing to Azure ciodive.com. AWS’s argument is that their scale and operational maturity (security, enterprise sales, global reach) will win out, even if Oracle or others see short-term AI pops. It’s worth noting AWS is still extremely profitable (32.9% operating margin) and generates far more cash to reinvest than Oracle or Google Cloud revolgy.com revolgy.com. So AWS has resources to respond – indeed it just announced an expansion of its partnership with Nvidia to offer next-generation Nvidia GH200 “Grace Hopper” superchips on AWS, and is reportedly working on a more powerful Trainium2 chip for 2024 fintool.com zacks.com. The true test will be whether AWS can secure equally large long-term AI contracts or if some AI upstarts remain committed to multi-cloud (or even shift more toward Oracle for cost/performance advantages).
• Oracle’s partnerships we largely covered: Nvidia, AMD on the tech side; Cohere on the AI model side; and even unusual alliances with competitors (Oracle databases on Azure/AWS, and Oracle Cloud cooperatively deployed alongside Microsoft, Google, and even Alibaba in some regions). Oracle’s strategy seems to be interoperability and “playing nice” to insert itself wherever possible. For example, Oracle says its multi-cloud database service (where Oracle manages databases in other clouds) saw 15× growth year-on-year investor.oracle.com, albeit from a small base. This indicates that Oracle is finding ways to make money even from customers running on AWS, Azure, etc., by providing unique tech like Autonomous Database in those environments. Ellison is effectively using Oracle’s niche strengths (database, ERP apps) as a Trojan horse to cross-sell OCI for AI. And now with the AI Database idea, Oracle will try to leverage its huge installed base of enterprise data stored in Oracle systems – offering those customers AI capabilities co-located with their data so they don’t have to move to a rival.

In summary, the competitive takeaway is that AI has reshuffled the deck: It created a new mega-growth opportunity that even a smaller player like Oracle could exploit by moving quickly. While AWS, Azure, and GCP remain much larger, Oracle has carved out a lucrative niche by being an “AI cloud specialist” willing to customize and commit resources to hungry AI players. As a result, Oracle’s growth rates now rival or exceed peers, and its stock has performed correspondingly. Yet, execution risks remain – Oracle must actually deliver the promised capacity on time, and the quality of these revenues (profit margins, etc.) is yet to be seen. Notably, some critics caution that Oracle’s revenue is still far below the implied backlog burn rate, meaning most of those contracts will play out over 5–10 years. If AI demand were to cool off or if clients re-negotiate, there could be turbulence. A Seeking Alpha commentator warned that the backlog might be a “distraction” if the AI “hype” moderates, noting Oracle’s overall revenue growth (12% last quarter) isn’t as dramatic as the backlog suggests seekingalpha.com seekingalpha.com. For now, however, the market is squarely focused on the long-term prize, and Oracle has firmly re-established itself in the conversation alongside the cloud hyperscalers.

Larry Ellison’s $100B Day and Oracle’s Public Perception Pivot

For decades, Larry Ellison was known as Silicon Valley’s flamboyant, fiercely competitive billionaire who built Oracle into a database powerhouse (often by deriding rivals and bucking tech trends – he infamously mocked “cloud computing” as a buzzword in the 2000s). Now, at an age when many executives retire to their yachts (Ellison indeed has a few), he is orchestrating one of the most dramatic second acts in tech history. Oracle’s AI-cloud triumph has turbocharged Ellison’s personal wealth and, arguably, his legacy.

Ellison’s net worth before the latest earnings stood around $280–290 billion – already making him a top-five richest person, thanks to Oracle’s steady rise in recent years and big 2023 stock gains. The post-earnings stock pop added roughly $100 billion to his fortune in one swoop reuters.com. This briefly placed him #1 on Bloomberg’s ranking at $393 billion crn.com. While day-to-day market moves may shuffle the exact ordering, it’s remarkable that Ellison has closed the gap with the likes of Elon Musk (whose wealth, largely tied to Tesla/SpaceX, was around $440B before falling slightly) reuters.com. Oracle’s value is now approaching that of Tesla, which is a testament to how investors now view Oracle as an “AI winner.”

The public narrative around Ellison has accordingly shifted from “legacy tech baron” to “AI kingmaker.” Media headlines noted that Ellison “tops Musk as world’s richest man after $101B gain” on the AI news bloomberg.com. Axios quipped that “Larry Ellison has surpassed Elon Musk as the world’s richest person after his stock in Oracle skyrocketed.” axios.com. For someone who long trailed peers like Bill Gates or Jeff Bezos on wealth lists, this was a personal milestone – albeit one driven by a single day’s stock spike, it symbolized Oracle’s new standing.

In interviews and public appearances, Ellison now strikes a tone of vindication. He has said that AI workloads are transforming the tech landscape and that Oracle is uniquely positioned because of its “high-performance cloud and autonomous database” combo. On the earnings call, he emphasized how Oracle’s technology can power AI for enterprises safely: “Our focus is to enable enterprises worldwide to create business value with AI, while keeping their data secure and private,” he said, referencing the Cohere partnership and Oracle’s AI-in-the-database approach oracle.com oracle.com. Ellison is effectively re-branding Oracle as a critical enabler of the AI age, not just an old software firm.

This pivot is also evident in Oracle’s marketing – for example, Oracle’s upcoming CloudWorld/AI World conference in October 2025 is heavily promoting Oracle’s AI capabilities, with Ellison likely to deliver keynote demos of AI services running on Oracle Cloud. The company has even begun using slogans like “Oracle: The Cloud for AI” in press materials. Such positioning would have sounded far-fetched a few years ago, when AWS and Azure dominated AI cloud mindshare, but now Oracle has real proof points.

Public perception of Ellison himself can be polarizing – he’s admired as a visionary by some, criticized by others for past aggressive business tactics. But the AI success has undoubtedly enhanced his image as a tech luminary who still has tricks up his sleeve. It’s worth noting Ellison is one of the few from the 1970s generation of tech founders (he founded Oracle in 1977) still at the helm and deeply involved in product strategy. Many on social media marveled at an octogenarian leading a cutting-edge AI charge. “Larry Ellison, 81, just pulled off the biggest coup in cloud computing,” one analyst tweeted, referring to the OpenAI deal. Ellison’s longtime bet on vertically integrated systems (software + hardware) seems to be paying off; Oracle’s experience tuning databases and hardware for performance is a selling point for AI loads that need every ounce of efficiency.

There’s also a storyline of redemption: Oracle was late to cloud computing initially. Ellison famously mocked cloud in 2008 (“Maybe I’m an idiot, but I have no idea what anyone is talking about. What’s the cloud?” he scoffed) and Oracle’s early cloud efforts faltered semianalysis.com semianalysis.com. It took until 2016 for Oracle to launch a truly competitive cloud (OCI Gen2) semianalysis.com, and even then Oracle remained a niche player for years. Many in the industry wrote Oracle off as an also-ran that could never catch up to AWS/Azure. Now, Oracle is being described as a “dark horse” that has suddenly outrun the field in the AI sprint. This unlikely comeback narrative – “from legacy to AI leviathan” – has not been lost on the public. Even comedians joked about Larry Ellison coming out of “villain retirement” to beat the younger billionaires at their own game.

Furthermore, Ellison’s personality – often viewed as combative and sales-driven – is now seen as an asset in striking mega-deals. One financial commentator mused that Oracle securing multi-billion commitments from the likes of OpenAI “has Larry’s fingerprints all over it,” as he likely negotiated directly with Sam Altman (OpenAI’s CEO) and others, leveraging relationships and perhaps offering more flexible terms than bigger rivals. Indeed, Oracle’s willingness to co-invest (e.g., possibly taking an equity stake in OpenAI’s cloud venture, similar to how it invested in TikTok’s U.S. operations in 2020) might have sweetened the pot, though details aren’t public. Ellison has a history of bold, sometimes unconventional deals – and here it appears to have netted Oracle the crown jewel of AI customers.

Of course, with great success comes greater scrutiny. Some analysts warn that Oracle must prove it can execute on margins and not overspend for growth. If Oracle’s costs to build all this AI capacity balloon, it could pressure profits. The balance of power with customers like OpenAI will also be something to watch – OpenAI now owes Oracle $300B in usage over years, but if AI economics change, will they utilize it fully? Ellison’s confident stance is that these AI companies absolutely will consume that and more, as AI services proliferate across industries. In his view, we’re at the start of a decades-long cycle where every enterprise demands massive compute for AI, and Oracle will get its fair share of that pie.

In the eyes of investors right now, Ellison has earned some benefit of the doubt. Oracle’s stock performance in 2023 (up ~55% as of mid-September) outpaced even high-fliers like Nvidia on a relative basis reuters.com. That indicates a re-rating of Oracle’s growth prospects. “Oracle has transformed into a veritable AI power player,” wrote one Bloomberg commentator, noting that its forward P/E multiple expanded as growth estimates were revised upward. Oracle even announced a dividend increase to $0.50, sharing some of the wealth with shareholders investor.oracle.com.

Ellison’s personal journey – from mocking “fashion-driven” cloud trends to now embracing “AI cloud” – illustrates the adaptability needed to survive in tech. As one investment firm (Melius Research) wryly noted, “Microsoft and Oracle aren’t really software companies anymore – they happen to sell software, but they are valued on AI cloud infrastructure” crn.com. Ellison would likely agree. In fact, he might argue Oracle always sold performance and integration (the hardware-software combo), and now AI is the ultimate performance-sensitive workload that rewards that approach.

One thing is certain: Larry Ellison is reveling in this moment. After spending much of the 2010s fending off the narrative that Oracle was a dinosaur, he now gets to play the industry visionary again, touting how Oracle’s technology will “revolutionize” business via AI. And the numbers, for now, back him up. Oracle’s unlikely run has proven that in the era of AI, even a “legacy” giant can reinvent itself and come out on top – with a little foresight, a lot of capital, and a founder who refuses to cede the spotlight.

Sources

• Oracle Investor Relations – Oracle FY2026 Q1 Financial Results (Press Release) investor.oracle.com investor.oracle.com
• Oracle Investor Relations – Oracle FY2025 Q1 Financial Results (Press Release Highlights) stocktitan.net stocktitan.net
• Reuters: Sriram, A. and Shah, A. “Oracle soars on AI cloud gains, Ellison closes in on Musk as world’s richest.” Sept. 10, 2025 reuters.com reuters.com
• Reuters: Babu, J. “OpenAI, Oracle sign $300 billion computing deal, WSJ reports.” Sept. 10, 2025 reuters.com
• CRN: Novinson, M. “Oracle Q1 Takeaways: …$455B Backlog, AI Replacing Humans, SaaS Questions.” Sept. 13, 2025 crn.com crn.com
• Investopedia: Laidley, C. “The AI Trade Picks Up Steam After Oracle’s ‘Truly Historic’ Quarter.” Sept. 10, 2025 investopedia.com investopedia.com
• CIO Dive: Wilkinson, L. “AWS Q2: generative AI helps cloud market near $100B, capacity race on.” Aug. 3, 2025 ciodive.com ciodive.com
• Reuters: Sophia, D.M. “Microsoft to spend record $30 billion this quarter as AI investments pay off.” July 30, 2025 reuters.com reuters.com
• Reuters: Ghosh, S. “Meta signs over $10 billion cloud deal with Google – source.” Aug. 22, 2025 reuters.com reuters.com
• Trefis (Analysis): “Oracle’s $455B Cloud Shock: Nvidia, AMD, Broadcom Are Big Winners.” Sept. 11, 2025 trefis.com trefis.com
• Oracle News: “Oracle to Deliver Generative AI services with Cohere.” Press release, June 13, 2023 oracle.com oracle.com
• Fierce Telecom: Goovaerts, D. “Cloud backlog nears $1T with AI in focus.” Aug. 18, 2025 fierce-network.com fierce-network.com