Oracle and SMCI both reported strong AI results and record-level backlogs, yet their stock prices collapsed. Shouldn't strong results lead to higher stock prices?
2026-06-11
Strong AI Results — So Why Did the Stock Fall?
Oracle disclosed record-breaking figures in its FY2026 Q4 results on June 10: cloud infrastructure (OCI) revenue +93% and remaining performance obligations (RPO) of $638B. Yet the stock fell -7% in after-hours trading. SMCI announced a $39B AI server backlog while simultaneously disclosing a $7B equity raise — and plunged -19.7%.
On the surface, this looks like a contradiction. In reality, the question investors are asking has fundamentally changed.
Why Investors No Longer Simply Look at 'How Much Was Sold'
In the early days of the 2023 AI boom, a cloud revenue increase of +30% alone was enough to push a stock higher. The market's logic was simple: "AI demand is exploding → infrastructure providers' revenue grows → that's good." Now the market demands the next question: "How long will it take to recoup that revenue?"
Oracle spent $55.7B in capex in FY2026 alone (exceeding initial guidance of $50B). In Q4 alone, it deployed $8.5B, pushing free cash flow to -$362M. No matter how strong the cloud revenue growth rate, if the payback period exceeds market expectations, the near-term stock price falls (Seeking Alpha, 2026-06-10).
Microsoft faces the same issue. Against an FY2026 AI-related revenue target of $25B, estimated capex runs from $97.7B to $150B — meaning it is pre-spending 4–6 years of capex upfront (Seeking Alpha, 2026-06-10).
SMCI's Case: Why Did a Backlog Announcement Lead to a Stock Crash?
SMCI explained that it "needs $7B to fill the $39B backlog," but the market saw something different. Specifically, the dilution was structured as: immediate new shares of $1.25B + an ATM program of $2.0B (starting Q3) + $3.75B in conditionally convertible preferred shares due 2029. The dilution to per-share value was certain, while the timeline for converting the $39B backlog into revenue is 2027 or later. Revenue is in the future; dilution is now — this asymmetry created the -19.7% drop (TechTimes, 2026-06-10; Cryptopolitan, 2026-06-10).
The 'AI Capex Headwind' Is Selective
Not all AI infrastructure companies face the same pressure. The key distinction is how capex is funded and how quickly it is monetized.
| Type | Example | Market Reaction |
|---|---|---|
| Self-funded via cash flow + rapid monetization | NVDA, AVGO | Relative outperformance |
| Heavy capex + negative free cash flow | ORCL, MSFT | Sustained pressure |
| Capex funded by equity dilution | SMCI | Most direct hit |
There is also the structural difference that NVIDIA is a beneficiary of AI infrastructure (the one selling), while Oracle and SMCI are the ones buying AI infrastructure.
What Should Investors Do?
The logic of "buy it because it's AI-related" is no longer sufficient. The checklist has changed.
- Is free cash flow (FCF) positive? If capex overruns push FCF negative — as with Oracle — dividend and buyback capacity shrinks and the stock's price floor weakens.
- When does RPO/backlog convert to revenue? Oracle's RPO of $638B is impressive, but if the conversion period is 3–5 years, there is no reason for it to be immediately reflected in today's stock price.
- How is the capital raised? Companies filling capex gaps with equity issuance (the SMCI approach) take longer to recover their stock price.
In the short term, the core message from today's market is that sectors with no capex burden — XLE (+1.50%), XLP (+1.69%) — showed relative outperformance. AI infrastructure investment is an attractive long-term theme, but a safer entry timing is after FCF recovery is confirmed.