V-Score Card

V = 0.25(4) + 0.22(3) + 0.18(5) + 0.12(3) + 0.15(4) − 0.06(4)
  = 1.00 + 0.66 + 0.90 + 0.36 + 0.60 − 0.24
  = 3.28

Gate 1:  E = 3 > 1                  → PASS
Gate 2:  A = 3 > 1                  → PASS
         C + E + U = 4 + 3 + 5 = 12 → PASS (≥12)

V = 3.28 × 1 × 1 = 3.28

Verdict: EMBEDDED κ = (3.28 − 3.0)⁺ = 0.28 Bustamante fast screen: 1/3 — proprietary data (database), no regulatory mandate, no transaction embedding.

DimensionScoreWeightContributionEvidence
C — Compound Cognition40.251.0049yr cross-domain, 7 suites × 22+ verticals, 9 compounding layers, common data model, 1,000+ AI agents. #2+ in most domains. [10-K L460-493, Q3 L55]
E — Irreducible Infrastructure30.220.66"Substantially all" renew ≈$20B support. PL/SQL 12-24mo switching cost. NO NRR/GRR disclosed. Software support 0% CC = 3-5%/yr net attrition. [10-K L806, 10-Q L1934]
U — Ecosystem Breadth50.180.9029+ product categories, 14+ departments, 22+ verticals, healthcare + infra + developer. Broadest in scored universe. [10-K L442-448, L460-890]
A — Distribution & Discoverability30.120.36Multicloud DB 531% YoY, AI Agent Studio, marketplace +89%. No MCP/LangChain, proprietary framework risk. [Q3 L98, Q3 L91-94]
M — Ecosystem Gravity40.150.60$57.4B rev, #1 RDBMS, Java, ≈52K patents, $60B+ acquisitions. #2+ in ERP/HCM/SCM. [10-K L736, L781]
F — Ecosystem Friction4−0.06−0.24$5.2B services at 19% margin, 12-18mo implementations. Anti-friction: 2,000+ go-lives/Q, APEX, Autonomous DB. [10-K L3358, Q3 L65]

Regime Context

Trailing 15 weeks (2025-12-10 to 2026-03-27). 74 trading days.

r_ORCL = −0.016%/day + 0.457 × r_SPY + 1.084 × r_IGV + ε

α̂ (annualized)  = −4.1%    (t = −0.05)
σ_idio           = 40.9%
IR               = −0.101   (noise — regime, not stock)
R²               = 0.456
%Idio variance   = 55.9%    (below 75% target)

ρ_intra = 0.413 — 99th percentile of 2-year history. All-time high for this peer set. The enterprise SaaS sector is moving as a single block. Spearman ρ(V-Score, 15-week return) = 0.071 (p = 0.87) — the market is NOT discriminating on structural survival quality. SAP (V = 4.09, Fortress) is down 32.8%. ADBE (V = 2.78, At Risk) is down 31.8%. Identical treatment.

When ρ_intra → 1: ε_i → 0 for all names. IR → 0 for all names — not because alpha is absent, but because the measurement window contains no idiosyncratic signal. IR measures the regime. V(s) ⊥ r_sector(t).

Rolling ρ_intraValueRegime
6 months ago0.140Normal
3 months ago0.172Normal
1 month ago0.327Elevated
Current0.413Extreme (99th pctl)
2-year mean0.179
2-year 95th pctl0.334

δ: Structural Value vs Market Pricing

During indiscriminate selloff, V_market converges to peer average. The edge is δ = V − V̄.

TickerVκIR (15w)15w ReturnRSIδ
SAP4.091.09−0.80−32.8%11.4+0.70
NOW3.930.93−0.97−41.9%15.2+0.54
PANW3.560.56+0.41−24.6%27.3+0.17
PLTR3.330.33+1.17−21.3%37.4−0.06
ORCL3.280.28−0.10−36.8%39.3−0.11
CRM3.100.10−0.16−31.2%30.6−0.29
WDAY3.050.05−2.38−44.1%19.6−0.34
ADBE2.780.00−1.68−31.8%15.5−0.61

ORCL δ = −0.11. Below peer average V̄ = 3.39. The indiscriminate selloff is slightly generous to ORCL — it is being treated as structurally equivalent to names (SAP, NOW) that have 2-3x its conviction weight.

Dimension Deep Dive

C = 4 — Under Challenge

What's load-bearing: Database kernel (40yr query optimizer), ERP business process encoding across 22+ verticals, Healthcare/Cerner clinical workflows, the common data model connecting ERP → HCM → SCM → EPM.

What's not: Middleware/WebLogic (commodity, open alternatives), Java (open-source), OCI (commodity compute), AI/ML layer (1-2 years old, nothing crystallized yet). Four of nine claimed "compounding layers" are not crystallized cognition.

Frontier model re-derivation timeline: Standard ERP workflows (AP, GL, procurement) are derivable from regulatory documentation in months. Cross-domain interactions and vertical edge cases (banking AML, healthcare clinical pathways) require 1-3 years. The common data model's thousands of customer-learned interaction patterns are the hardest to replicate — but Cloud Apps at 10-11% CC for 7+ consecutive quarters suggests the market does not view this as compelling enough to consolidate onto.

Score confidence: MEDIUM. Revenue-weighted crystallized cognition is closer to 3.5 than 4. The common data model is the margin between 3 and 4. Trigger for downgrade: Cloud Apps CC decelerates below 10% for two consecutive quarters.

E = 3 — Under Challenge

The switching cost is quantifiable and declining. PL/SQL migration is 12-24 months today for mid-market. Mission-critical OLTP (banks, telcos, government) is 3-5 years. But AI-assisted migration tools are compressing these timelines. SnowConvert is free. Datometry translates PL/SQL in real-time. MongoDB/Postgres are capturing new workloads.

Revenue-weighted c_ℓ:

Layer% of Revenuec_ℓTrajectory
Mission-critical OLTP≈15%3-5 yearsStable
PL/SQL mid-market≈20%12-24 monthsDeclining
Cloud Apps≈25%12-18 monthsFlat
OCI compute≈30%Days-weeksCommodity
NetSuite (SMB)≈10%3-6 monthsLow

The revenue-weighted average switching cost is much lower than the headline "PL/SQL lock-in" suggests. OCI — the fastest-growing segment at 81% CC — has near-zero switching cost.

NRR silence is Bayesian evidence. CRWD discloses 112% NRR. NOW discloses 120%+. WDAY discloses 97% GRR. Oracle's refusal to report is informative — if NRR were strong, they'd say so. "Substantially all renew" covers the legacy support base where customers are effectively hostage to migration pain. That's inertia, not product moat.

Software support at 0% CC on 3-5% price escalation = 3-5%/yr net volume attrition. The installed base is a wasting asset.

Score confidence: MEDIUM. E = 3 holds at the low end. Trajectory is toward 2, not 4. Trigger for downgrade: software support CC turns negative, or a major bank/government migrates off Oracle DB.

Scenario: Corrected V if E and C Erode

ScenarioCEVκδTier
Base case433.280.28−0.11EMBEDDED
E erodes423.060.06−0.33EMBEDDED (barely)
C erodes333.030.03−0.36EMBEDDED (barely)
Both erode322.810.00−0.58AT_RISK

Joint probability of both eroding within 2 years: 12-18% (correlated — same driver: AI capability improvement reduces both re-derivation cost and migration friction simultaneously).

Thermodynamic Summary

Oracle's survival rests on two resistances: crystallized business process knowledge encoded across 49 years and 22+ verticals (C = 4), and PL/SQL infrastructure switching costs that keep ≈$20B/yr of support revenue structurally locked (E = 3). Both are real but declining — frontier AI models compress the re-derivation timeline while migration tools compress the switching cost. The fastest-growing revenue (OCI, +81% CC) offers no thermodynamic resistance. The stickiest revenue (database support, 0% CC) is a wasting asset. Intelligence flows around Oracle where it can (new workloads to Postgres/Snowflake, commodity compute to hyperscalers) and stalls where it can't (mission-critical OLTP, cross-domain ERP configurations). The equilibrium is slow erosion, not collapse.

Conviction Weight

κ_ORCL = (3.28 − 3.0)⁺ = 0.28

In a basket where w_i ∝ κ_i:

  w_SAP  = 1.09 / Σκ = 1.09 / 3.52 = 31.0%
  w_NOW  = 0.93 / 3.52 = 26.4%
  w_PANW = 0.56 / 3.52 = 15.9%
  w_PLTR = 0.33 / 3.52 =  9.4%
  w_ORCL = 0.28 / 3.52 =  8.0%
  w_CRM  = 0.10 / 3.52 =  2.8%
  w_WDAY = 0.05 / 3.52 =  1.4%
  w_ADBE = 0.00 / 3.52 =  0.0%  (FILTER — AT_RISK)

ORCL gets 8.0% of basket GMV. SAP gets 3.9x more weight. The math is doing the right thing — thin κ, negative δ, eroding E and C → minimal allocation.

Basket Verdict

KEEP. V = 3.28 > 3.0. κ = 0.28 > 0. ORCL stays in the basket at structural minimum weight.

But ORCL is the wrong name to overweight in this selloff. δ = −0.11 means the indiscriminate selloff is slightly generous to ORCL. RSI = 39.3 — it hasn't even capitulated yet while SAP (11.4), NOW (15.2), and ADBE (15.5) are at extremes. If ρ_intra normalizes and the market starts discriminating on structural V again, ORCL underperforms SAP and NOW.

Watch for upgrade: NRR disclosure >105% (E → 4, V → 3.50, κ → 0.50). Cloud Apps growth acceleration to 15%+ CC (C confirmed, common data model pull validated).

Watch for downgrade: Software support CC negative (E → 2). Cloud Apps CC below 10% for two quarters (C → 3). Major customer displacement from Oracle DB (E floor cracks).

What Would Change the Score

TriggerDimensionNew ScoreVκδ
NRR >110% disclosedE: 3→443.500.50+0.11
Cloud Apps >15% CC sustainedC: 4→553.530.53+0.14
MCP adoption + agent ARR disclosedA: 3→443.400.40+0.01
Software support CC negativeE: 3→223.060.06−0.33
Major DB displacement (bank/gov)E: 3→223.060.06−0.33
Cloud Apps CC < 10% two quartersC: 4→333.030.03−0.36

Sources

SEC filings: 10-K FY2025 (filed 2025-06-18), 10-Q Q3 FY2026 (filed 2026-03-11) Earnings transcripts: Q3 FY2026 (2026-03-10), Q2 FY2026 (2025-12-10) Market data: yfinance (2026-03-28) Peer V-Scores: SAP 4.09 (adjudicated), NOW 3.93, PANW 3.56, PLTR 3.33, CRM 3.10, WDAY 3.05, ADBE 2.78 Regression: OLS 2-factor (SPY + IGV), 74 trading days, statsmodels