Oracle results are out — it beat the numbers across the board, so why did the stock fall?

In yesterday’s preview, I said, "This earnings release is the first test of whether the money poured into AI infrastructure is starting to come back as revenue and cash." The results are out. In one line, it comes down to this.

Revenue, EPS, and RPO all beat market expectations — but the stock fell.

After-hours, it at one point fell by more than 7% , then trimmed its losses as the earnings call went on, recovering to -1%(about 199dollars). This is a classic pattern of "the numbers are good, but the market isn’t buying it." We plugged the actual figures into the 3 checkpoints we had been watching and organized them below.

---

First, the earnings report card

Item	Market expectation (preview)	Actual announcement	Verdict
Revenue	about 191 billion dollars	192 billion dollars (+21% YoY)	Beat
Adjusted EPS	about 1.96 dollars	2.11 dollars	Beat
RPO	previous 5,530 billion dollars	6,380 billion dollars (+363% YoY, quarter over quarter +850 billion)	Surge
Cloud infrastructure growth	60~70% expected	+93%	Above expectations

On the surface, ①the RPO growth we had seen ② and revenue conversion passed both checkpoints cleanly. Even so, the share price fell. The culprit was the third checkpoint, debt and investment.

---

Scoring by checkpoint

① RPO, did the growth trend continue? → Pass

RPO, which was 5,530billion last quarter, came to 6,380billion this time, an increase of 1year on year. In one quarter, 363billion more was added. These numbers directly rebut the concern that 'AI contracts are cooling.'850,

However, what we pointed out in the preview, 'where the increase came from,' turned out to be a double-edged sword. Oracle said that most of the RPO increase came from large AI contracts, which at the same time also means high dependence on a small number of large customers.

②Will RPO turn into revenue quickly? → Passed, and visibility was also provided

Cloud infrastructure (OCI) revenue grew 93%, far above our estimate (60~70%). This is evidence that AI demand is being converted into actual revenue.

Specific figures on the pace of conversion were also disclosed. 6,380 billion dollars of RPO, 12% is expected to be recognized as revenue within the next 12 months, and an additional 34% within 13~36 months, and they said this share will keep speeding up. In the preview, we said that "the best outcome would be a specific conversion reference in the earnings call," and that is effectively what came out.

③Debt·investment vs. cash flow → This is where it got stuck

The part we emphasized in the preview as "the real center of gravity of this announcement" surfaced exactly as the weakness. There are three points.

First, the financing was not 'debt' but 'new shares.' Oracle said it would raise roughly 2027 in FY 400 billion dollars, and of that, 200 billion dollars is already announced stock issuance (ATM method). Instead of covering it with debt, it is issuing new shares, which dilutes existing shareholders. The direct trigger for the after-hours plunge of 7% was precisely this financing plan.

Second, capex came in above expectations and will rise further next year. In FY2026it spent about 560billion, and then projected FY2027net cash capex at 700billion. That is well above the "450to 500billion" range we saw in the preview. FY2026free cash flow was negative 237billion. Cash is going out faster than it is coming in.

Third, it signaled margin compression. Because of the timing before data center projects begin contributing materially to revenue, and the impact of the business mix, it said FY2027gross margin would decline. Growth is continuing, but profitability is temporarily worsening — the combination the market hates most.

---

Still, there were positive signals

For balance, let’s also note the good points.

• FY27some upward revision to guidance. It kept the revenue target at 900billion, but raised the adjusted EPS target to 8.05dollars. FY27total revenue growth rate was presented at +34% on a constant-currency basis, above the company’s long-term 5-year average growth rate.

• The prepayment structure eases capital burden. A significant share of large AI contracts is structured so that customers prepay for GPUs or supply them directly, and the scale amounts to 750hundred-million dollars. That means Oracle has less capital it needs to procure directly, which serves as a partial offset to debt concerns.

• Recovery from the drop. After-hours 7% plunge narrowed during the earnings call to -1%. The explanation on the prepayment structure and conversion visibility during the call may be a sign that some concerns were eased.

---

Summary: What diverged

At the preview, the question was, "Is it the promise (RPO), or is it the actual results?" The answer is this.

Both the promise and the results were good. But the 'cost' of realizing that promise was larger than the market expected.

Revenue and RPO beat did not outweigh the issues with funding, capex, and margins. On Wall Street, Oracle has become a key gauge for measuring “AI credit risk” amid the Big Tech borrowing race for AI infrastructure, and this reaction showed that character clearly. The market is now looking more closely at whether it can bear that growth, rather than how fast it is growing.

Bull case	Bear case
RPO 6,380billion, providing visibility into conversion	FY27 capex 700billion, surging
OCI +93%, accelerating revenue conversion	free cash flow -237billion
Prepayment structure 750billion, easing capital burden	400billion raised = concern over new share dilution
EPS guidance raised	FY27 gross margin decline forecast

---

What should investors watch next?

1. Tomorrow’s regular-session close. The share-price figures above are based on after-hours trading, which is highly volatile. Whether the decline widens in regular trading, or whether the stock rebounds with a “beat is a beat” response, will show the real direction.

2. The actual size and pace of new share issuance. How much dilution occurs, and how quickly it happens, will be the key variable holding back the stock.

3. Timing of the switch to surplus free cash flow. When capex peaks, RPO converts into revenue, and cash flow turns positive — that is the turning point in the debt narrative.

4. Customer diversification. Given the heavy reliance on large AI contracts, whether the customer base broadens determines the ‘quality’ of RPO.

Next quarter’s results will be 9in 10 announced.

This article is for informational purposes only and is not investment solicitation or advice. Investment decisions and responsibility rest with you. The stock price figures are based on after-hours trading immediately after the announcement and may differ in the regular session.