ORCL Research only — not a buy/sell tip

Does AI backlog turn into cash — or stay a dilution story?

The one idea

Oracle is the cleanest near-term test of whether AI-infra demand converts into cash — or stays a backlog story funded by dilution. Contracted demand looks enormous; free-cash conversion does not.

Bull — why it could work

FACT

FY2026 RPO $638B (+363% YoY); Q4 cloud infra +93%; company guided Q1 FY27 cloud growth 58–64% and FY27 ~$90B revenue / non-GAAP EPS $8.05 (Oracle IR Jun 10, 2026).

FACT

~$75B prepaid/customer-supplied GPUs in large AI contracts reduces some Oracle-funded capex (company IR).

INFER

A high-end print that shows RPO converting and a credible FCF bridge could re-rate the “growth vs cash” debate.

Bear — what breaks it

FACT

FY2026 FCF −$23.7B; plans ~$40B FY27 financing incl. ATM equity (company IR).

FACT

Interest expense rising (Q4 $1.44B vs $0.98B YoY in company tables).

INFER

Miss vs cloud guide or soft cash commentary re-prices dilution/credit risk ahead of growth.

When it matters

Q1 FY2027 results Thu Sep 10, 2026 after the close; call/webcast 4:00 p.m. CT (Oracle IR / PR Newswire Sep 2).

Watch: Watch on the call: cloud vs guide, RPO, cash flow / financing / ATM usage.

What would change our mind

Toward bear — cloud below guide + larger equity raise / credit path worsens.
Toward bull — high-end cloud + clear peak-cash-burn-behind bridge without more dilution.

Coming to terms — words that make this stick

RPO: Remaining performance obligations — contracted future revenue still to be recognized.
FCF: Free cash flow — cash from operations minus capex.
ATM equity: At-the-market stock sales that can dilute existing holders.

FACT = company/official. INFER = judgment. Source: ORCL-2026-09-07.md