Oracle, which has maintained steady growth with databases and enterprise software, has made its biggest bet since its founding. Following $18 billion (about 27.144 trillion won) in September last year, it also issued $25 billion (about 37.695 trillion won) in corporate bonds in February this year to invest in artificial intelligence (AI) data centers. Oracle said it raised funds "to build additional infrastructure capacity to meet the contracted demand of cloud infrastructure customers such as AMD, Meta, Nvidia, OpenAI, TikTok, and xAI."
But the market reaction is not all positive. International credit rating agency S&P on the 9th (local time) downgraded Oracle's credit rating one notch to BBB- from BBB. BBB- is the lowest tier among investment-grade ratings, and one more notch down would be speculative grade. What S&P was concerned about was not AI itself. It is that a company with high profitability and stable cash generation is shifting its business to the AI data center sector, which requires massive upfront investment, making its financial structure more sensitive to investment and business cycles than in the past.
The share price is reflecting these concerns. Oracle shares soared to an all-time high on AI growth expectations, hitting $328.33 intraday in September last year. The stock had more than doubled from around $160 at the start of last year. However, after the company announced a $18 billion corporate bond issuance in September last year, concerns grew over large-scale borrowing and financial burdens from expanding AI infrastructure investment, and the stock turned downward. After sliding to $136.48 on Feb. 5 this year, the stock has been choppy. In particular, in results for the fourth quarter of fiscal year 2026 (March to May) released on the 10th of last month (local time), revenue came in at $19.2 billion and adjusted operating profit at $8.6 billion, up 21% and 22%, respectively, from a year earlier, beating market expectations. Even so, the stock fell about 10% in after-hours trading. Although the revenue growth rate topped 20% for the first time since the third quarter of 2023, cloud revenue slightly missed market expectations, and concerns resurfaced about financial burdens from massive investment and borrowing to expand AI infrastructure.
Attention is on whether AI investment will become Oracle's trigger point or a decisive move for a latecomer's counterattack.
◇ Why the latecomer Oracle is placing a bet
Founded in 1977, Oracle grew into the world's largest enterprise databases company. Until the mid-2010s, licenses for databases and enterprise resource planning (ERP) were core, but with the launch of the second-generation cloud (Gen2 OCI) in 2018, it began a full-fledged transition into a cloud company. After reorganizing enterprise software into cloud-based subscription services, from 2023, when Generative AI spread, it shifted its focus from developing AI models to an AI infrastructure strategy centered on graphics processing unit (GPU) superclusters and data centers.
Oracle is currently the No. 4 player in the cloud market after Amazon Web Services (AWS), Microsoft Azure, and Google Cloud. According to Synergy Research Group, Oracle's share is about 2% to 3%, far behind the top three. Even so, as Oracle took on debt and jumped into the AI race, the market reaction seems to be more concern than optimism.
Oracle's large-scale investment also means that competition in the AI industry is shifting from technology-centric to capital-centric preemption of infrastructure. In traditional cloud businesses, it was common to add servers when customers arrived, but AI data centers are different. GPU procurement, power infrastructure, cooling facilities, and ultra-high-speed networks must be built even before customer contracts. Investing after customers arrive is already too late. In the AI era, even latecomers cannot enter the market without massive upfront investment.
Oracle CEO Safra Catz also said, "We will secure additional customers worth several billion dollars in the coming months, and remaining performance obligations (RPO) are expected to exceed $500 billion." Oracle Chairman Larry Ellison has explained the investment background by saying, "Because long-term contracts (RPO) have surged, we have to build data centers to fulfill those contracts."
◇ Will targeting gaps with an AI-only cloud work… "The direction is right, but the financial burden is heavy"
Oracle's strategy is not to build AI models like GPT or Gemini from the start. Instead, the strategy is to become a cloud infrastructure provider that trains and runs AI models, as well as general cloud. In general cloud, it competes with Microsoft and Google, but at the same time aims to be a provider of computing infrastructure used by AI companies such as OpenAI, xAI, and Meta, including those two.
This background aligns with the philosophy of Oracle Chairman Larry Ellison. He expects the AI industry will grow with a focus more on "inference" than on "training." Ellison recently said, "Oracle's core goal is to enable AI systems to safely access public and private high-value data and to perform inference without exposing the data," adding, "Most of the world's most valuable data is already inside Oracle databases. We are advancing the databases so AI can perform inference based on that data."
The markets Oracle is focusing on are not only general cloud but also AI-dedicated cloud. Rather than chasing head-on the general-purpose cloud market already dominated by AWS and Microsoft Azure, it has chosen a strategy that concentrates on GPU clusters and AI data centers.
Experts say Oracle's direction is right, but the financial burden is heavy. Rebecca Wettemann, CEO of IT market research firm Valoir, said, "Oracle's multicloud strategy and enterprise data capabilities are key drivers of AI growth." By contrast, S&P said, "The growing AI infrastructure business is diluting the company's previously stable business profile."