甲骨文

Bloomberg reported on Sept. 24 that Oracle sent Stack Infrastructure, a Blue Owl subsidiary, a force majeure notice to protect itself from payment obligations if the Project Jupiter data center in New Mexico is delayed. Oracle shares fell 5% in early trading.

Why it matters: If the project is delayed, we are confident that Oracle can protect its financial interests against the data center developer. However, given the missing capacity, revenue from Oracle Cloud Infrastructure customers like OpenAI may decrease, putting Oracle's long-term revenue goals in question. Project Jupiter's capacity of 2.45 gigawatts means that Oracle would potentially need to push back the timeline for $25 billion or more of its revenue, which represents a sizable part of our OCI revenue forecasts of $86 billion and $136 billion for fiscal 2028 and 2029, respectively. A potential delay could also pose a challenge for Oracle in maintaining its investment-grade credit rating. Oracle banks on cash inflows from artificial intelligence data centers to keep its leverage in check. A higher cost of borrowing inevitably squeezes the margin of new data center projects.

The bottom line: We maintain our $220 fair value estimate for narrow-moat Oracle as Project Jupiter's delay is only a possibility at this stage. Our base case assumes that Oracle can keep up with its delivery pace and fulfill its long-term growth targets. The shares currently look undervalued. We think the market's reaction does not fully represent the downside of an actual delay that would hurt Oracle's long-term revenue guidance. The stock could see another selloff if data center delays start to affect Oracle's performance at the company level.

Big picture: Pure cloud service providers like Oracle and CoreWeave are vulnerable to project delays because they rely on third-party data center operators to build the power shell. In November 2025, a data center delivery delay for CoreWeave sent the stock down 16%.

Our fair value estimate for Oracle is $220 per share, which implies a fiscal 2027 enterprise value/sales multiple of 8 times and an adjusted price/earnings multiple of 26 times. Following a period of rapid growth, Oracle's forward adjusted price/earnings should gradually step down to 16 times by fiscal 2030. We expect Oracle’s annual revenue growth to accelerate to an average of 31% over the next five years as the adoption of Oracle Cloud Infrastructure, or OCI, and Oracle Cloud Applications, or OCA, continues to tick up. The ramp-up of AI data centers is a major driver of Oracle’s top-line growth. Cloud should become Oracle’s key growth driver and accounts for around 85% of the company’s revenue by fiscal 2030. Meanwhile, we expect a five-year CAGR of 62% for OCI and 8% for OCA. Total cloud revenue (OCI plus OCA) is expected to grow around tenfold over the next eight years.

We acknowledge the high degree of uncertainty baked into our forecast, including long-term demand from key customers like OpenAI and Oracle's ability to secure the necessary resources for its data center build-out. As cloud services become the mainstream choice for customers, we model a low-single-digit revenue decline for both the license and hardware segments over the next decade. The services segment should fare better with low-single-digit revenue growth as customers continue to rely on Oracle for data infrastructure consulting services. Our fair value depends on AI demand remaining healthy relative to supply, allowing Oracle to pursue its data center build outs, and operating its data centers at full capacity.

Over the next five years, we expect Oracle's cloud infrastructure buildout to significantly drive down its GAAP operating margin, from 31% in fiscal 2026 to 28% in fiscal 2029. After that, with economies of scale, Oracle's operating margin should gradually climb back up to 33% by fiscal 2035. We also think that Oracle’s research and development expense ratio will remain in low double digits as the competition among database and enterprise cloud products intensifies.

We think Oracle has a narrow economic moat supported by high switching costs. Database systems and Oracle’s other enterprise software are critical to the day-to-day operation of modern enterprises. Companies tend to stay with the same vendor for years, or even decades, for core systems to ensure optimal business continuity. In addition, cloud infrastructure revenue is also sticky due to the high cost and risk of data ingress and egress. These features should keep Oracle’s return on invested capital above its cost of capital over the next 10 years, as it is a key player in these areas.

After introducing “Generation 2 Cloud” in 2018, Oracle has seen an acceleration in the growth of its Cloud Infrastructure, or OCI, and Cloud Applications, or OCA, businesses. In our view, Oracle’s market share in cloud infrastructure should remain below 10% in the near term, but OCI’s rapid revenue scaling could lead to the fastest revenue growth in the company’s recent history. We believe OCI is an important piece that supports Oracle's overall moat rating of narrow due to its strong switching costs. Once a customer gets on board, they tend to stay with the same cloud provider due to the complexity, cost, and risk of data migration across different clouds. Also, a multicloud strategy, where companies use two or more public cloud providers, is becoming more common across enterprises. Given the limited number of cloud providers that are at scale, we have a hard time identifying a scenario that necessitates a complete retreat from OCI for enterprise customers that already signed up for it.

Although Oracle’s aggressive data center capacity ramp-up should, in theory, lead to lower unit cost of compute due to economies of scale, we do not see any signs that OCI’s AI-oriented GPU clusters have an economic moat as a stand-alone business unit. However, we recognize that OCI’s AI capacity should enable many cross-sell and up-sell opportunities at the company level, which should reinforce Oracle’s switching costs. Existing Oracle customers can leverage Oracle’s AI infrastructure to train and host machine learning models using data already in their Oracle systems, and new Oracle customers coming to OCI for GPU capacity can also become Oracle database or enterprise software users down the road.

We see strong switching costs with Oracle Database, too. Database is hard to replace because these are systems of record that facilitate the exchange of data across different enterprise software and AI systems. If a company would like to change its database system provider, the process of “rewiring” how it stores and accesses data can involve significant monetary and time-related costs that often exceed its internal IT capabilities. Oracle database serves the world’s largest enterprises. For these customers, database migration projects will take at least a few years with careful planning. Moreover, switching database systems can lead to significant operational risks if the company would like to keep business as usual during the transition. Overall, we believe the number of existing Oracle Database customers that have the capability of migrating away their databases is small, and they need a very compelling reason to justify the significant cost and potential disruption that could happen during the migration process.

Similar to Oracle Database, we see strong switching costs across Oracle’s applications lineup. The company has built a robust array of enterprise software through acquisitions, including enterprise resource planning, or ERP, supply chain management, or SCM, and human capital management, or HCM, platforms. It also has cloud-based enterprise software targeting small and medium-sized businesses called NetSuite. Implementing a new software package typically requires significant time and financial resources from the customer. Employees onboarding new enterprise software also face a learning curve, which can sometimes lead to lost productivity. There is not much that AI can help with to accelerate the process, given the magnitude and strong deterministic nature of workflows that Oracle systems are handling. Therefore, we usually see long upgrade cycles for enterprise software, especially among companies in traditional industries that lack digital transformation capabilities.

Besides switching costs, we also see traces of intangible assets in Oracle. Ever since pioneering the commercialization of relational databases in the 1980s, Oracle has faced strong competitive pressure from other closed- and open-source relational database products like Microsoft SQL Server, IBM Db2, and PostgreSQL. More recently, the rise of NoSQL databases, such as MongoDB, and data warehouses, such as Snowflake, has made the database market even more competitive. However, Oracle remains one of the top three database vendors due to its reliability, scalability, and security. Enterprises nowadays rely significantly on data to be the backbone of their business and fuel that powers various AI initiatives. Therefore, how such data is stored and accessed becomes one of the highest-stake decisions for chief information officers. Although Oracle Database products usually come with a higher price tag, many enterprises are still willing to pay the premium in exchange for its reliability and performance.

According to DB-Engines, Oracle has been the most popular database management system for over 12 years in a row based on a score calculated using Google search trends, technical discussions on Stack Overflow, and engineer profiles on LinkedIn. We believe the database engineering community shares the consensus that Oracle is equipped to handle the world’s most important data workflows that require the highest level of availability and accuracy.

Scaling of Oracle Cloud Infrastructure has helped retain customers, port workloads to the cloud, and create new cloud service revenue, all of which should continue in the coming years.

Oracle Database should be able to keep its market leadership as customers continue to depend on its industry-leading performances in terms of stability and security.

Oracle Cloud Infrastructure was built with flexibility and ease-of-use in mind, which could bring a significant base of first-time Oracle users to the company, strengthening top-line results.

Demand for generative AI can undershoot expectations, potentially causing Oracle to miss its long-term revenue and earnings goals.

Oracle needs to overcome multiple supply chain challenges to deliver the contracted data center capacity. Delays in the supply chain can negatively affect Oracle’s revenue growth.

Oracle’s balance sheet is among the most leveraged within our software coverage, which could limit operational flexibility and future acquisition opportunities.