The Quarter Was Good. The Stock Disagreed.

Something did not add up Tuesday night.

Oracle reported Q4 FY2026 results after the close on June 10. Revenue came in at $19.2 billion, up 21% year-over-year, ahead of the $19.1 billion estimate. Non-GAAP EPS came in at $2.11 against a $1.96 consensus. Cloud infrastructure revenue grew 93% year-over-year. The remaining performance obligation backlog, which is basically contracted future revenue, hit $638 billion – up 363% from a year ago and up $85 billion in a single quarter. Full-year FY2026 revenue landed at $67.4 billion, up 17%. By any honest read, that is not a company in trouble.

ORCL fell 7% to 11% overnight anyway.

So what actually happened.

The number that broke the mood was not in the earnings. It was in the guidance. Oracle announced it plans to raise $40 billion in new capital during FY2027 – a combination of debt and a $20 billion open-market equity offering. Layer on top of that the $70 billion in guided net capex for the year plus another $20 to $25 billion in component prepayments, and you get a company that could be deploying somewhere close to $95 billion in capital in a year when it expects to generate roughly $90 billion in revenue. That math is not a typo. Oracle may spend more than it takes in, and it told the market that openly. Some investors respected the transparency. A lot of them sold first and asked questions later.

To be fair, this is not new behavior.

In FY2026, Oracle raised $43 billion in debt and $5 billion in equity. Capex came in at $55.7 billion, overshooting their own $50 billion projection. The pattern here is that Oracle keeps raising its own spending targets and then beating them to the upside. Whether that is a feature or a bug depends entirely on whether the revenue follows. So far it has. The question is whether it continues at this scale.

There is a detail from the earnings call worth sitting with. CEO Clay Magouyrk said Oracle plans to bring nearly one gigawatt of computing capacity online this quarter alone. One quarter. That is roughly equivalent to what the company deployed across the entire fiscal year in FY2026. The acceleration in build rate is not incremental anymore – it is a different order of magnitude, and it is reshaping what Oracle actually is as a business.

The backlog is real, but how it is built matters. Most of the RPO growth in Q3 and Q4 came from large AI deals where customers either prepaid Oracle for GPU procurement or handed over hardware directly. That model has generated $75 billion in cumulative customer-funded investment, which meaningfully reduces Oracle’s out-of-pocket burden for new data center builds. Bank of America noted that more than 50% of the RPO traces back to a single counterparty: OpenAI. That is not diversified demand. That is one relationship carrying a lot of weight, and concentration risk at that scale is worth pricing in before you get too comfortable with the backlog headline.

Both things can be true. The backlog is real and the concentration is a risk.

At around $192 per share post-selloff, ORCL is trading at roughly 26 times FY2027 non-GAAP EPS guidance of $8.05. Revenue guidance for the year is $90 billion. The multiple is not obviously cheap. But for a company compounding cloud infrastructure revenue at 90%-plus with $638 billion sitting in contracted backlog, 26 times is not obviously expensive either. The balance sheet is the honest concern. Debt-to-equity is near 366%. That number is fine in a growth environment where revenue conversion stays on track. It becomes a different conversation if rates stay elevated and backlog-to-revenue timing slips.

What I am watching is not the RPO headline. It is how much of that backlog shows up as recognized revenue each quarter. That conversion rate is the only number that actually resolves the bull-versus-worry debate on this stock. Everything else is speculation layered on top of a very large backlog figure.

If you own it, the business has not broken. The financing structure is aggressive but it has been aggressive before and the revenue came through. If you are looking at a new position, the $175 to $180 range is worth having on your radar – that is where the valuation argument gets meaningfully more compelling. The Q1 FY2027 report around September 10 is the first real checkpoint.

The harder question – the one I do not think anyone has a clean answer to yet – is whether the market will keep rewarding this kind of capital intensity the way it has rewarded it at Microsoft and Amazon. Oracle is a different company with a different history, and the benefit of the doubt is not automatic.

Worth a look before September.