Skip to main content
Be The Investor
BTI Market Newsstocks
Back to News

Oracle’s $638 Billion Backlog Carries a Customer Concentration Risk

Oracle’s $638 Billion Backlog Carries a Customer Concentration Risk

Oracle’s enormous contracted backlog supports rapid cloud growth, but Steve Eisman’s estimate that roughly half relates to OpenAI highlights how much execution depends on one private customer.

Oracle’s $638 billion backlog looks like extraordinary revenue visibility. It also raises a less comfortable question: how much of that demand depends on one customer’s ability to keep funding a historic computing expansion?

Investor Steve Eisman said in an October 16 interview that roughly half of Oracle’s backlog is tied to OpenAI. That would equal about $319 billion. Oracle has not publicly broken out the customer’s share, so the figure should be treated as Eisman’s estimate, not as company disclosure.

Oracle’s reported numbers establish the scale. The company ended fiscal 2026 with $638 billion of remaining performance obligations, up from $553 billion one quarter earlier. Fiscal-year revenue was about $67.4 billion. In other words, contracted business now spans many years of current revenue, and the pace at which it becomes recognized sales matters as much as the headline amount.

Loading market context

The concentration question is sharper because Oracle is building capacity before receiving all of that revenue. The source cited negative free cash flow of $23.7 billion for fiscal 2026 and total debt of about $167 billion. Those figures mean the company is financing a substantial portion of the buildout while customers scale their own economics.

A large backlog is valuable only if contracts are enforceable, customers can pay, and infrastructure becomes available on schedule. Customer concentration increases the damage from renegotiation, delayed deployment or weaker demand, even if the total order book remains unchanged.

There is a bullish interpretation. OpenAI and other model developers need scarce computing capacity, Oracle’s cloud infrastructure revenue rose 93% in its latest reported quarter, and the backlog may support years of growth. The bear case is that one customer represents too much of the promised demand while Oracle bears capital costs now.

Investors should resist turning Eisman’s estimate into a settled fact. The better test is observable: watch how quickly remaining obligations convert into revenue, whether capital expenditure produces improving free cash flow, and whether Oracle discloses broader customer diversification. Until then, the backlog is both the company’s strongest asset and its most important concentration risk.