Futures
Access hundreds of perpetual contracts
CFD
Gold
One platform for global traditional assets
Options
Hot
Trade European-style vanilla options
Unified Account
Maximize your capital efficiency
Demo Trading
Introduction to Futures Trading
Learn the basics of futures trading
Futures Events
Join events to earn rewards
Demo Trading
Use virtual funds to practice risk-free trading
CFD
U.S. stock CFD derivatives
US Stocks
Access real US stocks and ETFs
HK Stocks
Trade quality Hong Kong-listed stocks
Korean Stocks
SK Hynix
Real Korean stocks and top assets
Stock Futures
High leverage, 24/7 trading
Tokenized Stocks
Backed by real stock assets
IPO Access
Unlock full access to global stock IPOs
GUSD
Mint GUSD for Treasury RWA yields
Stocks Activities
Trade Popular Stocks and Unlock Generous Airdrops
Launch
CandyDrop
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
IPO Access
Unlock full access to global stock IPOs
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
Promotions
AI
Gate AI
Your all-in-one conversational AI partner
Gate AI Bot
Use Gate AI directly in your social App
GateClaw
Gate Blue Lobster, ready to go
Gate for AI Agent
AI infrastructure, Gate MCP, Skills, and CLI
Gate Skills Hub
10K+ Skills
From office tasks to trading, the all-in-one skill hub makes AI even more useful.
Oracle Rarely Reveals That a Data Center “May Not Be Able to Recoup Costs,” Oracle’s Stock Plunges 40% in June
In its latest annual financial report, Oracle rarely self-disclosed that its massive expansion of data centers may not be profitable, and if major clients fail to pay or renew contracts, the company could be left with a large amount of idle, expensive assets.
(Previous context: Too heavy a price for the AI gamble? Microsoft reportedly lays off 5,500 people in its third major裁员 within a year, with Xbox hardest hit) (Background supplement: Build an AI customer service in 2 minutes! xAI launches no-code Voice Agent Builder, beating OpenAI in voice benchmarks)
Table of Contents
Toggle
Oracle's capital expenditures for fiscal year 2026 (ending this May) surged to $55.7 billion, more than double the $21.2 billion from the previous year, and for fiscal 2027 it plans to spend another $90 billion to $95 billion. Almost all the money is flowing into data centers to supply computing demand for clients like OpenAI. But Oracle itself rarely admits in its latest annual report that this bet may not pay off.
This warning appears in Oracle's annual report disclosure. Oracle detailed every possible way its data center investments could go wrong, with a bluntness rarely seen among tech giants.
Self-disclosed risk: laying out the nightmare scenario
Oracle warns that data center construction costs may overrun and timelines may be delayed due to supply chain bottlenecks, government restrictions on data center development, or third-party vendors failing to complete on time. Simply put, any hiccup in building infrastructure like data centers can cause costs to spiral out of control.
More troublesome is what happens after completion. Oracle admits that major clients may not pay or choose not to renew contracts. If that happens, Oracle would be left with extremely expensive assets that "may not be able to be re-leased, repurposed, or transferred on acceptable terms."
In plain English: the data center is built, but the client leaves. Oracle can only watch as money gets stuck in concrete, steel, and chips. Capital expenditures—the visible big money spent on data centers and chips—turn into impaired fixed assets if rental income doesn't materialize.
Of course, Oracle emphasizes these are only potential worst-case scenarios, not high-probability events. But this level of detailed warning reflects growing Wall Street anxiety about the burn rate of the AI frenzy.
OpenAI, the elephant in the room not named
Oracle did not directly write the word "OpenAI" in its report, but the entire warning points almost unmistakably to that entity.
According to the Stargate agreement, Oracle is building large-scale data centers across the United States, providing computing power to OpenAI through OCI (Oracle's cloud computing business). The total value of this contract is $300 billion, the largest single order Oracle has ever received, and the largest share among the six giants' combined $850 billion in undrawn data center lease agreements.
In other words, the viability of this entire plan hinges on whether OpenAI pays its bills.
Oracle wrote in its filing: "Some of our customers may be highly leveraged and bear their own operational and regulatory risks. Even if our credit review and analysis mechanisms function properly, we may still encounter non-payment and non-performance risks when dealing with these customers."
Market rumors had already surfaced that the Stargate expansion was once stalled due to disagreements among the three parties over control of the data centers.
Six giants all doubling down, Oracle bears the heaviest burden alone
Wall Street's doubts about AI investment returns were already reflected in stock prices in June. Six companies, including Oracle, Microsoft, and Meta, collectively committed $850 billion to yet-to-be-built data center leases. But all six saw their stock prices fall in June, with Oracle dropping over 40% in a single month, the worst performer.
Moody's warned that Oracle's debt growth has outpaced its earnings, with leverage approaching four times EBITDA. Compared to Microsoft and Meta, which primarily fund data center construction with their own free cash flow, Oracle is the only one among the six giants that heavily relies on debt to finance AI infrastructure. It carries over $100 billion in debt on its books, and its free cash flow has turned negative.
However, for most tech companies, the risk of overspending is far outweighed by the risk of missing the AI wave. Oracle itself wrote in its report: "To grow our OCI business, we must continue to increase our computing capacity and invest substantial capital and operating expenditures." This contradiction may be the most honest portrayal of the current AI infrastructure race: everyone knows the bet may not pay off, but no one dares to leave the table first.