Morgan Stanley has raised important considerations regarding Oracle Corporation's (NYSE: ORCL) investment obligations connected to its expanding backlog of customer contracts. Over the past year, Oracle's backlog of contracted commitments has surged by $426 billion, reaching a current total of $523 billion in remaining performance obligations. This sizable backlog offers Oracle notable revenue visibility extending approximately five years.
However, Morgan Stanley draws attention to the fact that nearly 60% of this backlog—over $300 billion—is linked to customers securing infrastructure specifically for AI compute applications, including major partnerships like OpenAI. This focus on AI-related infrastructure implies considerable expenditure, given the capital-intensive nature of such assets.
The investment bank estimates that building a gigawatt of greenfield AI compute infrastructure could require as much as $35 billion in capital investment. Oracle aims to expand its AI compute capacity beyond 10 gigawatts by the close of its fiscal year 2030, a substantial undertaking by any measure.
This scale of expansion translates into an anticipated $356 billion in cumulative cash capital expenditures dedicated to IT equipment, such as GPUs essential for powering AI workloads. Additionally, Oracle could incur finance lease liabilities worth approximately $185 billion stemming from data center 'powered shells' necessary for this infrastructure growth.
Reflecting these capital requirements, Morgan Stanley has revised its base case earnings per share (EPS) forecasts for Oracle. The projections now stand at $8.51 for fiscal year 2028 and $10.02 for fiscal year 2030, figures notably lower than Oracle's own targets of $10.65 and $21.00 and below consensus estimates of around $10.68 and $19.67.
In parallel developments, Morgan Stanley has adjusted its price target for Oracle down to $213, expressing reservations about the profitability impact of the company's GPU-as-a-Service operations. Conversely, Guggenheim maintains a Buy rating for Oracle with a $400 price target, designating it as its 'Best Idea' within the software sector. KeyBanc Capital Markets also continues to hold an Overweight rating with a $300 price target, distinguishing between Oracle's traditional core business and its growing infrastructure division.
Additional corporate developments include Oracle's anticipated transaction to acquire TikTok's U.S. operations, alongside Silverlake and other American investors. Meanwhile, Oracle faces legal pressure as bondholders have initiated a proposed class action lawsuit, criticizing the company for allegedly not disclosing their intent to issue more debt to finance AI infrastructure expansion.
These dynamics illustrate a complex environment for Oracle, balancing extensive growth opportunities in AI infrastructure against the significant capital demands and regulatory scrutiny that accompany them.