ORCL › analyze
ORCL · Analyze from before
Date: 2026-09-10 · Command: /analyze-from-before · Price: $159.26 (post-earnings quote, immediately after the Q1 FY27 print; pre-print regular-session close $152.94, down 5.4% intraday on a broad tech selloff unrelated to Oracle; next-session press coverage reports a $160.38–165.14 range, opening ~$163.38 — the quote is still settling into the new session)
Baseline correction: The orchestrator named analyze-2026-07-09.md (BUY 7.0) as the prior report. kb.py find ORCL and a directory listing both surface a newer verdict-bearing report, analyze-from-before-2026-08-28.md (HOLD, conviction 6.5), which itself already superseded analyze-2026-07-23.md. Per protocol ("read the newest before re-deriving"), this pass is tested against 2026-08-28, not 07-09. This is the same correction the 08-28 report made against 07-09 — the orchestrator's cached pointer is now two reports behind.
Event since baseline: the print is out, and it beat. Oracle reported fiscal Q1 FY2027 after the close on 2026-09-10 — the exact event the 08-28 report's recheck was set for. This is the first genuine new fundamental data point since 2026-06-10 (the FY26 print), ending a nearly six-week window (07-23 → 08-28 → 09-10) in which two prior passes each concluded "nothing has changed, only price/sentiment moved."
What this updates — event list since 2026-08-28
- Q1 FY2027 results (2026-09-10, after close): Revenue $19.35B (+30% YoY, above the +27–29% guide and the ~$19.1B/+28% consensus). Non-GAAP EPS $1.92 (+30% YoY) vs. ~$1.74 consensus (+10.3% beat). GAAP EPS $1.56 (net income $4.68B) vs. $1.01 a year ago. OCI (cloud infrastructure) revenue +121% YoY to $7.4B — accelerating, not decelerating, from Q4 FY26's +93%. Total cloud (SaaS+IaaS) revenue +62% to $11.6B — squarely inside the ~58–64% band this report was tasked to check. Cloud Apps (SaaS) +10% to $4.2B, unchanged growth character.
- RPO $664B, up $209B YoY (+46%) and up $26B sequentially from the 08-28 baseline's $638B — beat StreetAccount consensus of $630.6B. Sequential dollar growth decelerated sharply ($85B added in Q4 FY26 → $26B added in Q1 FY27) even as the absolute beat continued. More than $30B in new AI cloud contracts booked in the quarter.
- Guidance raised, not just held: full-year FY27 non-GAAP EPS raised to $8.10 (from $8.05–8.06); revenue reaffirmed at "a minimum of $90B." Q2 FY27 guide: EPS $1.85–1.93, revenue +30–34%. Capex guidance unchanged: $90–95B gross, net cash outlay not to exceed $70B — no negative surprise on the funding story, and management explicitly reiterated spending will be non-linear (back-loaded) through the year.
- No credit-rating action. S&P remains BBB- (stable outlook per the 07-09 action), Moody's negative outlook unchanged. Zero-notch buffer to junk persists, unresolved.
- Litigation progressed on schedule, not resolved. Following the amended complaint (lead plaintiffs Sparinvest S.A. / SEB Funds AB, filed 2026-07-14), defendants' motion to dismiss is due 2026-09-16; plaintiff's opposition due 2026-11-17.
- No insider transactions recorded in the delta window (Aug 28–Sep 10) — extends the "no exodus, but also no buying at any level" pattern from 08-28.
- Analyst target dispersion widened into the print, in both directions. Several desks cut targets ahead of the print despite holding Buy ratings — TD Cowen $300→$240, Jefferies $320→$290, Scotiabank $241→$215 (this last one downgraded target the day before earnings) — while Guggenheim ($400, Buy) and Citizens ($285, Outperform) held firm and Morgan Stanley's Equal-Weight nudged up slightly ($207→$210). Post-print, at least one raise appeared (Bernstein/Moerdler $319→$325, Outperform).
The delta ledger — every 2026-08-28 claim re-tested
(Numbering follows the 08-28 report's own ledger for continuity.)
| # | Baseline claim | Status | Fresh evidence |
|---|---|---|---|
| 6 | RPO $638B (+363% YoY) | 🔁 REFRESHED | New: $664B, +$209B YoY (+46%), +$26B sequential — beat consensus ($630.6B). Named watch item, not a break: sequential dollar growth decelerated from +$85B (Q4 FY26) to +$26B (Q1 FY27). One quarter of lumpier mega-deal timing is the more likely read (bookings this size aren't signed every quarter) than demand cooling, but this is the first data point suggesting the backlog's growth rate may be normalizing off its 2025–26 highs. Watch the next print before calling a trend. |
| 11 | Analyst consensus ~Buy, mean PT $244 (compressed but static) | 📉 DRIFTED | The "Street hasn't chased the rally" pattern intensified into outright target cuts at three desks (TD Cowen −20%, Jefferies −9%, Scotiabank −11%) in the two weeks before the print, even as ratings held Buy — professional money grew more cautious on the multiple even as the headline growth narrative stayed intact. Post-print, at least one raise (Bernstein $319→$325) suggests the beat partially repaired that caution, but there is no clean post-print consensus figure yet (Yahoo's aggregate feed had not refreshed at data-pull time). Net: valuation anxiety among sophisticated desks is real and growing, a genuine tension against the fundamentals beat. |
| 2 | FY27 FCF likely deepens toward −$45B to −$50B | ⏳ UNTESTED — first partial data, doesn't settle it | Q1 FY27: reported capex ~$28B gross, ~$18B net of prepayments/financing; reported FCF ≈ −$5.4B for the quarter (secondary-source figures — Yahoo's quarterly cashflow feed has not yet ingested the Aug-31 quarter, so this could not be independently reconciled against a primary filing this pass). Implied OCF for the quarter (FCF + net capex) ≈ $12–13B, well above Q1 FY26's $8.1B. This leans mildly supportive of a better full-year outcome than the baseline's fear (annualizing loosely: ~$45–50B OCF against ≤$70B net capex ≈ −$20B to −$28B FY27 FCF, not −$45 to −$50B) — but Q1 is historically Oracle's lightest OCF quarter (25% of FY26's full-year OCF fell in Q1), management explicitly flagged non-linear spending, and the underlying numbers are inferred, not filed. Marked UNTESTED rather than REFRESHED because one quarter against an explicitly-warned-nonlinear ramp cannot confirm or deny a full-year figure — the real test is Q2 and Q3. |
| 3 | Total debt $156.2B, net debt $98.3B, net debt/EBITDA ~3.1x | ✅ CARRIED (FY26 figure; Q1 FY27 balance sheet ⏳ UNTESTED) | The FY26 year-end figure stands unchanged (no new annual print). The interesting question — did leverage move during Q1 FY27 — could not be answered this pass: Yahoo's quarterly balance-sheet feed still ends at 2026-05-31 (Q4 FY26); the Aug-31 quarter has not propagated. Needs the 10-Q. |
| 4 | ROIC 12.1%, declining from 14.0% | ✅ CARRIED | No new annual print since the 08-28 pass's independent roic.ai re-confirmation (11.14% FY26 vs. 13.72% FY25, methodology gap from Yahoo's calc but same direction). Unchanged this pass. |
| 5 | Gross margin 65.8%, down from 70.5% | ✅ CARRIED | Same annual print; Q1-specific segment margin detail was not independently re-pulled this pass (sub-point ⏳ UNTESTED). |
| 1 | FY26 FCF −$23.7B, "the crux" | ✅ CARRIED | fin.py reconciles again to −$23.69B (OCF $31.98B − capex $55.66B). Static historical fact; unaffected by this quarter's print. |
| 7 | OpenAI/Stargate concentration is the central risk; BYO/prepaid mix de-risks the capital model | ✅ CARRIED | No restructuring, default, or scope change found. Construction (Abilene fully operational at 1.2GW; new sites in Michigan, Wisconsin, Wyoming, Pennsylvania) continues on the previously-known trajectory. Unresolved, unchanged, still the #1 named structural risk. |
| 8 | Dual-CEO governance (Magouyrk + Sicilia), governance risk score 9/10 | ✅ CARRIED | Confirmed via fresh get_stock_info officer roster — both still listed as Co-CEOs. Unchanged. |
| 9 | Active securities-fraud class action, amended complaint filed 2026-07-14 | 🔁 REFRESHED | Progressed on the normal litigation calendar: motion to dismiss due 2026-09-16 (six days from this report), plaintiff's opposition due 2026-11-17. Alive, not dismissed, not settled — an actionable near-term date now exists (the MTD ruling) that the baseline did not have. |
| 10 | Zero-notch buffer to junk (S&P BBB-, 2026-07-09; Moody's negative outlook) | ✅ CARRIED | No further rating action found since 07-09. The buffer remains exactly as tight as the 08-28 pass described it. This did not improve on the back of a beat-and-raise quarter — worth noting, since a quarter this strong is exactly the kind of print that could prompt a rating agency to reconsider, and none has (yet). |
| 12 | Entry zone $110–115 fired (week of 07-27) and passed | ✅ CARRIED | No new information changes this. The window closed at $114.50 and has not reopened. |
| 13 | No insider-exodus red flag; no insider buying at the trough either | ✅ CARRIED | Insider-transactions feed shows zero new transactions of any kind in the Aug 28–Sep 10 window — extends the absence-of-signal finding in both directions. Still not a red flag, still not a golden flag. |
| 14 | Vendor forward P/E is the wrong fiscal year (methodology finding, reused from 07-23) | ✅ CARRIED, reconfirmed a third time | Current feed: epsCurrentYear 8.06 / priceEpsCurrentYear 18.97 (FY27, matches the company's own guide) vs. epsForward 10.967 / forwardPE 13.95 (one fiscal year further out, FY28). A live second instance of the sibling trap appeared this pass: epsCurrentYear still reads 8.06 even though Oracle's own press release just raised full-year guidance to $8.10 — the vendor field has not caught up to today's guidance raise. See [[pitfall-vendor-forward-eps-is-stale-on-the-day-of-a-guidance-cut]], previously documented only for guidance cuts; this is the same staleness mechanism on a raise. This report uses the company's own $8.10 figure, not the stale vendor field. |
🆕 NEW — no baseline counterpart: - The Q1 FY27 beat itself (row set out in "Event list" above) — the first real fundamentals data point in six weeks of price-only re-rating. - FY27 EPS guide raised to $8.10 (from $8.05–8.06) — small in magnitude (+0.5–0.6%) but directionally important: it confirms this quarter's move is still mostly a re-rating (multiple expansion), not an earnings-estimate re-rating, even though real growth proof arrived. - RPO sequential-growth deceleration (#6 above) as a named watch item. - Analyst target dispersion widening into the print (#11 above) as a named tension.
How the close calls were decided
Claim #2 (FY27 FCF trajectory). Two forces pull opposite ways: Q1's implied OCF (~$12–13B, well above Q1 FY26's $8.1B) and unchanged capex guidance both lean toward a less severe FCF deficit than the baseline's −$45B to −$50B fear. Against that: Oracle's own FY26 pattern shows Q1 is structurally the lightest OCF quarter of the year (25% of the annual total), management explicitly warned spending would be non-linear, and the Q1 figures used here are secondary-source estimates, not yet reconciled against a filed 10-Q. Decided UNTESTED, not REFRESHED, because a single quarter against an explicitly-flagged non-linear ramp is a data point, not a trend — but it is now on record as not the bad kind of quarter, which matters for how much weight the FCF fear should still carry in conviction.
Claim #6 (RPO). The absolute number beat consensus and grew further; the rate of sequential growth fell by two-thirds. Decided REFRESHED (not DRIFTED) because the level still supports the thesis and beat expectations — but the deceleration is named explicitly as the thing to watch, since a second consecutive quarter of shrinking sequential adds would be a genuine structural signal, not noise.
Claim #11 (analyst targets). Pre-print cuts at three desks vs. a post-print raise at one is a real tension, not a resolved one — the cuts happened on a stale information set (before the beat), and Yahoo's consensus feed had not refreshed to reflect the print at data-pull time. Decided DRIFTED rather than CARRIED specifically because the direction of professional sentiment on the multiple (independent of the beat) had been souring, which is itself informative context for how richly this stock is now priced.
Thesis persistence and conviction delta
Structural claims (7, 8, 14): 3/3 CARRIED — 100%. Trend claims (1, 2, 4, 5, 6): 4/5 CARRIED or REFRESHED, 1 UNTESTED — 80%. Combined Structural + Trend persistence: 7/8 = 87.5%. State claims (3, 9, 10, 13): 4/4 CARRIED or REFRESHED — 100% (leverage figure itself UNTESTED at the Q1 level, but the claim as stated at FY26 stands). Price claims (11, 12): 1 DRIFTED, 1 CARRIED.
This is the healthy version of high persistence: unlike the 08-28 pass — where the business was unchanged and only price moved (a pure re-rating on sentiment) — this pass shows the business delivering against the thesis (OCI accelerating past its own break-trigger direction, RPO beating consensus, capex guidance held, no new credit or litigation deterioration) while price moved only modestly in response (+5–9% net over two weeks, most of it in the post-print pop, after an unrelated −5.4% intraday drop the same day). Fundamentals validated roughly in proportion to the price move, rather than price running ahead of proof as it did between 07-23 and 08-28.
Conviction: 6.5 → 7.0. Driving rows: #6 (RPO beat + absolute growth continuing), the NEW Q1-beat and raised-guide rows (OCI re-accelerating past the named break trigger, not toward it), #7 and #10 (the two biggest named risks — OpenAI concentration and the credit buffer — did not get worse, and a beat-and-raise quarter is exactly the kind of print that could have started to repair the credit story, even though it didn't). Held back from a larger move or an upgrade to BUY by: #11 (valuation richer, and professional desks grew more cautious on the multiple, not less, into the print), #2 remaining UNTESTED (the single largest quantitative risk in the file — FY27 FCF — is not yet resolved by one light quarter), and #10/#7 being unresolved rather than improved (a strong quarter that doesn't touch the two structural risks that matter most doesn't earn a full risk-premium re-rating).
Updated verdict — HOLD, conviction [7.0]/10
Valuation, re-derived (EPS $8.10 FY27, company's own current-fiscal-year guide — not the vendor forwardPE field, which is one year off per claim #14):
| Scenario | Multiple (current-FY basis) | Implied price |
|---|---|---|
| Bear | 10–11x | $81–89 |
| Base | 14–16x | $113–130 |
| Bull (near-term, now backed by a delivered quarter, not just sentiment) | 18–21x | $146–170 |
Fair-value range: $90–170. At $159–165 (the settling post-print quote), the stock sits near the top of the bull-case bucket, essentially unchanged in position from the 08-28 pass's read at $151.94 — the modest price move and the modest EPS raise moved together, leaving the multiple only slightly richer (~18.85x → ~19.7–20.2x on the same current-FY basis) even after a real beat. Graham IV (√(22.5×5.52×13.04)) ≈ $40 — still not a useful anchor for this name (negative tangible book, capex-heavy). Bogle: ~1.2% yield + the raised EPS growth path, contingent entirely on the multiple holding near 20x rather than reverting to the 14–16x base case — the same ~20–25% air pocket as before.
Position management: - Add zone: $110–129 (below 16x current-FY EPS). Unchanged — already fired once (week of 07-27) and has not reopened. - Do-nothing: $129–170 (current price sits near the top of this band). - Trim/reassess: above 21x current-FY EPS (≈$170 at the new $8.10 guide — the dollar level rises automatically with the EPS guide under the house convention; the multiple itself is unchanged from the 08-28 pass because neither the credit buffer nor the concentration risk improved enough to justify raising the ceiling).
Key risks, re-ranked: 1. 🚩 Credit — zero buffer to junk, unchanged by a beat-and-raise quarter. The single most telling non-event this pass: a quarter this strong did not move the rating conversation at all. 2. 🚩 OpenAI/Stargate concentration — unresolved, unchanged, still roughly half the RPO base. 3. 🚩 FY27 FCF trajectory — still UNTESTED. Q1 data leans supportive but is thin, inferred, and covers Oracle's seasonally lightest quarter. Q2 (due ~Dec 10) is the real test. 4. ⚠️ RPO sequential deceleration ($85B → $26B QoQ) — new, named, not yet a trend. 5. ⚠️ Valuation has richened on proof, not just hope, but offers no further margin of safety — professional-desk target cuts into the print show real caution on the multiple even as the fundamentals beat. 6. ⚠️ Litigation — MTD ruling due imminently (2026-09-16); a denial would extend the overhang meaningfully, a dismissal would remove it.
Upgrade conditions (what would move this to ACCUMULATE/BUY): a further rating-agency response to the improved OCI trajectory (an affirmation with a stable-not-negative Moody's outlook, or S&P daylight beyond zero notches); a second consecutive quarter confirming the lighter-than-feared FCF trajectory with primary-source (10-Q) figures; or a pullback into the $129 base-case band without a fundamental deterioration.
Downgrade triggers: OCI growth falling below 40% for two consecutive quarters (currently +121%, the furthest from firing it has ever been); a further credit downgrade into junk; an equity raise beyond the guided $20B ATM; an OpenAI/Stargate restructuring or default; RPO sequential adds continuing to shrink for two more quarters.
Bottom line: the business delivered on the promise this time, and the price barely had to move to accommodate it — the healthiest kind of quarter for a thesis, but not, on its own, cheap enough or de-risked enough on the two things that matter most (credit buffer, concentration) to earn more than a modest conviction bump. Hold what's held, do not chase, and treat Q2 FY27 (~Dec 10) as the test that actually resolves claim #2.
What this pass did NOT test
- Claim #2 (FY27 FCF full-year trajectory) — the single most consequential open question in the file. One light quarter, secondary-sourced, against an explicitly non-linear guided ramp. This is a claim that has now stayed open across three consecutive passes (07-23 flagged it, 08-28 called it UNTESTED, this pass finds a data point but still cannot close it) — worth flagging per protocol: the system keeps deferring the same question, and it will not resolve until at least Q2 or Q3 FY27 print with primary-source figures.
- Q1 FY27 balance sheet (debt, net debt, leverage ratios) — not yet in any structured feed pulled this pass (Yahoo's quarterly aggregator still ends at the 2026-05-31 quarter). Needs the 10-Q, expected in the following weeks.
- Q1 FY27 gross margin / segment-level detail — not independently re-pulled; the annual FY26 figure was carried forward without a fresh quarterly check.
- Post-print analyst consensus mean target — Yahoo's aggregate feed had not refreshed to reflect the 09-10 print at data-pull time; the $241 mean cited in the "Price" section predates the beat. Cited pre-print target moves (TD Cowen, Jefferies, Scotiabank) and the one confirmed post-print raise (Bernstein) are the only individually-sourced data points available.
- Exact current price — the quote was still settling between the immediate post-market print ($159.26) and next-session press-reported action (~$160–165) at the time this report was written; treat any single point figure in this range as approximate.