Oracle's (ORCL.US) ambition received positive reviews from analysts, but customer concentration risks are still receiving attention

The Zhitong Finance App learned that Oracle (ORCL.US) recently held an “Analyst Day” event and announced ambitious long-term financial goals at the conference. The driving force behind this is the continued rise in demand for artificial intelligence (AI) applications and infrastructure. This plan has won generally positive reviews from analysts. Goldman Sachs and Piper Sandler (Piper Sandler) have all recently released relevant reports and raised target prices.

On “Analyst Day,” Oracle announced its long-term financial targets for fiscal year 2030 (FY30): total annual revenue is expected to reach US$225 billion, corresponding to a five-year compound annual growth rate (CAGR) of 31%; the diluted earnings per share (EPS) target is $21, with a five-year CAGR of 28%. Among them, the Oracle Cloud Infrastructure (OCI) business, which is the core growth engine, raised its revenue target from US$144 billion to US$166 billion, accounting for nearly 75% of its total revenue, an increase of more than 15%.

Oracle Co-CEO Clay Magouyrk (Clay Magouyrk) also revealed details of the gross margin of the AI data center GPU computing power sales business:

“For example, let's say a customer wants to buy a GPU accelerator of about 1 gigawatt, what would happen?” Maguyoke said, “If the customer signs a 6-year contract, the total contract value (TCV) will reach 60 billion US dollars based on 10 billion US dollars per year.”

He further disassembled the cost structure: “I split the relevant costs into two parts. The first part, which I call 'site, data center, and power costs', clearly includes building facilities, actual power generation costs, and the labor costs required to run this part of the business—this part of the cost ultimately accounts for about 35% of the total cost of delivering this service.”

Maguyoke added, “The second part is the cost of computing, networking, and storage, that is, the cost of all the equipment we deploy within the data center, which accounts for about 65% of the total cost... but what I need to be clear is that the 35% gross profit margin I'm talking about doesn't just cover every year after the business is in operation. This 35% gross margin actually includes upfront start-up costs.”

“Obviously, you can imagine that if we use the first year of business launch as the dividing point, we will incur expenses before starting, but we have not received the corresponding revenue at this point. Moreover, this will happen at multiple project sites — we are not just building one such data center, but are proceeding with the construction of 10 or 20 projects at the same time.”

Many big banks are “singing a lot”, and analysts still suggest risks

Piper Sandler (Piper Sandler) analysts Hannah Rudoff (Hannah Rudoff) and J.R. Herrera (J.R. Herrera) stated in an investor report released last Friday: “Oracle's high-profile analyst day event revealed updated long-term goals, including reaching $225 billion in revenue for fiscal year 2030 (FY30) — which means a five-year compound annual growth rate (CAGR) of 31%, higher than our previous estimate.”

The report further explained: “At the core of supporting this goal is the strong growth forecast for the cloud infrastructure business: the five-year operating cash flow (OCI) compound annual growth rate of the business is expected to reach 75%, driving its fiscal year 2030 revenue scale to US$166 billion (up from the previous forecast of US$144 billion).”

Based on this judgment, Piper Sandler reiterated the “overweight” rating of Oracle and raised the target price from $330 to $380.

In addition to Piper Sandler, Guggenheim also raised Oracle's price target from $375 to $400; T.D. Cowen also raised its target price from $375 to $400.

However, analyst Kevin Anthony D. Arroyo (Kevin Anthony D. Arroyo) pointed out that there are still risks for Oracle to achieve these lofty goals. Specifically, the company's business relies too much on OpenAI — about two-thirds of Oracle's nearly $500 billion undelivered orders came from the startup led by Sam Altman.

Arroyo emphasized, “This poses a significant risk of customer concentration for Oracle, especially considering that OpenAI is still unprofitable.”

He further explained in an analysis report released last Thursday: “Under the agreement, OpenAI will reach an 'infrastructure leasing' cooperation model with Oracle Cloud Infrastructure (OCI). For OpenAI, this means spending around $60 billion a year; if calculated as a five-year agreement, the total amount would amount to $300 billion, and the agreement would come into effect in 2027. But the question is, how will OpenAI pay for this, or even maintain this collaboration?”

Goldman Sachs gave a “neutral” rating and raised the target price to $320

Goldman Sachs released a research report on Oracle on October 17, reaffirming the “neutral” rating while raising the 12-month price target from $310 to $320. As of the close of trading on October 16, 2025, Oracle's stock price was $313, which means that the potential increase corresponding to the target price is only 2.2%. The report points out that Oracle's long-term growth target is attractive, but implementation uncertainty and financial risk cause the overall risk return to balance, so it maintains a neutral rating.

On Analyst Day, Oracle made a positive outlook on future prospects. In the short term, Goldman Sachs also clearly predicts Oracle's financial performance for the 2025-2028 fiscal year: total revenue will gradually grow from $57.399 billion in fiscal year 2025 to US$115.583 billion in fiscal year 2028; diluted EPS will increase from $6.03 to $10.05, but gross margin will decline year by year, from 72.0% to 58.1%, reflecting cost pressure during business expansion.

Goldman Sachs said that the OCI business is the key to supporting Oracle's long-term goals, and analysis is carried out from the three aspects of growth drivers, customer development, and segmentation performance. In terms of growth, the explosion in AI infrastructure demand is the core — OCI can deploy large-scale AI GPU clusters more efficiently with the Acceleron network architecture (equipped with a dedicated Ethernet network card), while reducing costs to meet the market's demand for AI training and inference.

In terms of customer development, Oracle achieved a breakthrough in the second quarter of fiscal year 2025 (F2Q), adding 4 new non-OpenAI customers (including Meta), and signed a total infrastructure business contract value (TCV) of 65 billion US dollars, effectively mitigating the previous dependence on a single customer.

In terms of segment performance, OCI's four major sectors all achieved high growth: AI Infrastructure had the highest year-on-year growth rate of 117%, with a gross profit margin of 30-40%; Distributed Cloud grew 77% year over year, with a gross profit margin of 40-60%; Cloud Natives grew 49% year over year, with a gross profit margin of 40-60%; and the enterprise business grew 33% year over year, with a gross profit margin of 65-80%.

In addition, Goldman Sachs also revealed the cost structure of typical OCI contracts (such as a 6-year AI infrastructure order with a TCV of 60 billion US dollars): land, data center, and electricity costs account for 35% of total costs (including initial start-up costs), 65% of computing, network, and storage costs, and an overall gross margin of about 35%.

In addition to OCI, Goldman Sachs believes that Oracle's AI, agent (agent) ecosystem, and SaaS full-stack strategy also constitute an important growth complement. In the field of AI, Oracle has deployed more than 600 agents (including more than 400 in Fusion ERP and more than 200 in vertical industries). If agents developed by customers and partners are included, the actual scale is even larger; at the implementation level, 2,400 customers have used industry AI agents within 18 months, and the deployment cycle is only a few days without professional service or training.

In the SaaS field, Oracle implements a “database+data platform+OCI” full-stack architecture. The annual recurring revenue (ARR) of customers using this architecture is 150 times higher than that of customers using only a single product; the company plans to expand the Universal Credits (Universal Credit) model from OCI to the application layer to further lower the customer AI application threshold and promote business penetration.

The report also highlights the four core risks Oracle faces, which are the key reasons for maintaining a neutral rating. The first is execution risk. After FY2027, there are uncertainties about the data center construction cycle, GPU procurement volume and pricing, and energy costs, which may affect the pace of business delivery.

The second is customer concentration risk. Of Oracle's nearly $500 billion undelivered orders, about 2/3 come from OpenAI, which is not yet profitable. Its five-year $300 billion infrastructure lease agreement with OpenAI (effective from 2027) poses potential payment capacity and sustainability risks.

Third, financial pressure. Due to high capital expenditure (CapEx), free cash flow is expected to be negative for the 2026-2028 fiscal year, which will slow EPS growth and the release of operating leverage. Fourth, there is the risk of market competition. Oracle's share of the database market has continued to be lost in the past ten years, and the IaaS business still ranks fourth in market share, lagging behind leading vendors such as AWS, Azure, and GCP.