Oracle AI Boom Costs $95B

Oracle’s AI boom is now colliding with a simple market question: who pays for all this power, concrete, and debt?

Quick Take

Oracle said fiscal 2026 capital spending reached about $55.7 billion, above earlier guidance and well above what many investors expected.^[1]^[2]
The company said fiscal 2027 capital spending could rise as high as $95 billion, and some of that may be repaid by customers.^[2]^[3]
Oracle also expects to raise about $40 billion in 2027 through debt and equity financing.^[2]^[3]
Investors are split between seeing real AI demand and seeing a risky cash burn that may take years to pay off.^[4]^[5]

Why the Stock Dropped

Oracle shares fell after the company’s spending plans came into sharper focus, not because investors suddenly stopped caring about AI. Reuters reporting and other coverage say the company’s forecast for capital spending in fiscal 2027 came in above Wall Street estimates, while Oracle also signaled it may need more financing.^[4]^[1] That mix shook traders who had been focused on growth, not on the size of the bill.

The market reaction shows how fast enthusiasm can turn when a company moves from hype to hard numbers. Oracle is still being rewarded for large cloud demand, but investors are now weighing that demand against rising borrowing needs, negative free cash flow, and the risk that returns arrive too slowly.^[2]^[5] For many readers, that is the same old problem in a new AI costume: big promises up front, uncertain payback later.

The Spending Bill Behind the AI Push

Oracle’s latest figures show the scale of the buildout clearly. One report says the company spent about $55.7 billion on data centers in fiscal 2026, beating its own target by $5.7 billion.^[1] Another report says Oracle expects to raise nearly $40 billion in 2027 through a mix of debt and equity, while fiscal 2027 capital spending could climb as high as $95 billion.^[2] Those are huge numbers even by big tech standards.

The key detail is that Oracle says some of the future spending may be repaid by customers.^[2]^[3] That matters because it supports the bull case that real contracts are driving the buildout, not just speculation. Still, the repayment language also shows why investors are uneasy. The company is spending first and waiting later for part of the money to come back, which puts pressure on cash flow, financing terms, and execution.^[1]^[2]

Backlog Is Big, But It Is Not Cash Yet

Oracle’s supporters point to its massive backlog as proof that demand is real. One report says remaining performance obligations reached $638 billion, up from $553 billion three months earlier.^[1] That is a powerful number, and it explains why many analysts still see Oracle as a major AI infrastructure player. But backlog is not the same as revenue, and it does not erase the need to fund the buildout before the money arrives.

That gap is the heart of the debate. Reuters-linked coverage says Oracle expects only part of that backlog to turn into near-term revenue, while the rest may take years.^[1]^[2] Critics say that creates a dangerous mismatch between spending and payback. Supporters say the contracts prove Oracle is building for real demand. Both views can be true at once, which is why the stock is under so much pressure.

What This Says About the Bigger Market Mood

Oracle’s stumble fits a wider fear that the AI trade may be running ahead of itself. Marketplace coverage says investors are worried the company is spending more than it makes, while other reporting says Oracle’s results revived talk of an AI bubble.^[5]^[6] That broader mood matters because once one big company gets hit for heavy spending, the same fears spread fast across the sector.

Oracle showed the new AI market problem:

Revenue can beat.
Cloud can grow.
Backlog can explode.
And investors can still worry if capex and debt rise too fast.
AI demand is bullish.
AI funding risk is real.
Oracle said fiscal 2027 capex could reach up to $95B, above estimates,…

— FIRErify | Financial Independence Tracker (@Firerifyapp) June 11, 2026

For ordinary investors, the lesson is blunt. AI demand may be real, but demand alone does not pay the bills. Oracle now has to prove that its giant contracts can cover giant costs without forcing too much debt onto the balance sheet.^[1]^[2]^[4] If it can do that, the spending may look smart in hindsight. If it cannot, this will be remembered as another case of Wall Street financing the future before the future was ready.