## Oracle (ORCL) — 2030 trajectory vs. capital stack: what the filings imply ### Executive takeaway Oracle’s disclosures show an unusually large *contracted revenue runway* (remaining performance obligations) alongside an equally unusual *infrastructure funding ramp* (capex + long-dated capacity/lease commitments). The “hidden alpha” is that these two items move together: the RPO step-up helps de-risk demand into the buildout, while the capital stack actions (notes, ATM equity capacity, mandatory convertible preferred) suggest management is intentionally pre-funding the build to avoid an uncontrolled leverage spike. ["Remaining performance obligations were $523.3 billion"](0001193125-25-315925|0) ["Capital expenditures… increased… to $20.5 billion"](0001193125-25-315925|0) ["sell shares… up to $20.0 billion"](0001193125-26-037840|0) ["Depositary Shares Offering closed"](0001193125-26-039344|0) --- ## 1) The core debate: “2030 path intact” vs. “capital stack still a worry” ### What the filings *support* (path intact) The strongest filing-based evidence that the multi-year growth plan has real demand behind it is the sheer scale and duration profile of contracted commitments. * **RPO is $523.3B as of Nov 30, 2025** with a multi-year recognition schedule (only ~10% expected to recognize in the next 12 months, then 30% in months 13–36, 35% in months 37–60, and the remainder thereafter). ["Remaining performance obligations were $523.3 billion"](0001193125-25-315925|0) ["expect to recognize approximately 10%… next twelve months… remainder thereafter"](0001193125-25-315925|0) * Management explicitly attributes the Y/Y jump in RPO to **“certain significant cloud contracts”** entered during the period, i.e., not merely accounting noise. ["primarily attributable to certain significant cloud contracts"](0001193125-25-315925|0) **Hidden alpha angle:** RPO at this magnitude is not just backlog—it is *financing collateral in disguise*: it strengthens Oracle’s ability to raise long-dated capital at scale because it evidences forward cash-generation potential (even if revenue recognition lags). ["Remaining performance obligations were $523.3 billion"](0001193125-25-315925|0) ### What the filings *do not let you dismiss* (capital stack worry) Oracle’s disclosures also show the cost of fulfilling that demand is front-loaded and massive. * **Capex surged to $20.5B in the first half of FY2026**, explicitly tied to data center expansion, with management expecting the upward trend to continue for “the next few fiscal years.” ["Capital expenditures… increased… to $20.5 billion"](0001193125-25-315925|0) ["We expect this upward trend to continue… next few fiscal years"](0001193125-25-315925|0) * The balance sheet shows **property, plant & equipment (net) rising to $67.875B** at Nov 30, 2025 vs $43.522B at May 31, 2025 (a +$24.353B increase, computed), consistent with the capex ramp. ["Property, plant and equipment, net… 67,875"](0001193125-25-315925|0) ["Property, plant and equipment, net… 43,522"](0001193125-25-315925|0) * Oracle discloses **very large off-balance-sheet style capacity commitments**: as of Nov 30, 2025 it had **$248B of additional lease commitments** (substantially all related to data centers and cloud capacity arrangements), generally commencing between Q3 FY2026 and FY2028, for 15–19 year terms. ["we had $248 billion of additional lease commitments"](0001193125-25-315925|0) ["terms of fifteen to nineteen years"](0001193125-25-315925|0) **Hidden alpha angle:** the $248B figure reframes the “capex cycle” as only one leg of the buildout—Oracle is also locking in multi-decade capacity via leases/capacity arrangements, which can behave economically like leverage even if accounting treatment differs. ["$248 billion of additional lease commitments"](0001193125-25-315925|0) --- ## 2) Demand visibility: reading RPO like an analyst (not a headline) ### Why RPO quality matters more than RPO size RPO can be inflated by short-duration contracts; Oracle’s disclosure suggests the opposite: it’s long duration by design. * Only **~10%** recognized in the next 12 months implies the backlog is not “about to churn”; it is structurally long-dated. ["expect to recognize approximately 10%… over the next twelve months"](0001193125-25-315925|0) ### The “invisible” implication for 2030 thinking Even without explicit “2030 targets” in the filings, the time-bucketed schedule plus “remainder thereafter” strongly implies Oracle is stacking obligations well beyond a typical 3-year planning window—useful for underwriting a multi-year thesis if execution holds. ["30%… month 13 to month 36… 35%… month 37 to month 60… remainder thereafter"](0001193125-25-315925|0) --- ## 3) Infrastructure buildout: the tell is not capex—it’s commitments ### Capex: the visible build * FY2026 first-half capex: **$20.535B**. ["Capital expenditures… (20,535)"](0001193125-25-315925|0) ### Commitments: the less visible build * **$248B additional lease commitments** (data centers/cloud capacity), starting mainly FY2026–FY2028, 15–19 years. ["$248 billion of additional lease commitments"](0001193125-25-315925|0) * Plus **$10B unconditional purchase and other obligations** as of Nov 30, 2025, primarily related to cloud capacity arrangements. ["unconditional purchase and certain other obligations… were $10 billion"](0001193125-25-315925|0) **Hidden alpha angle:** the combination of (1) capex, (2) long-dated lease commitments, and (3) capacity-related purchase obligations is effectively a “capacity flywheel.” The market often prices only capex; the filings show the true scale includes long-term fixed-ish commitments that will pressure free cash flow timing even if margins ultimately expand. ["Capital expenditures… increased… to $20.5 billion"](0001193125-25-315925|0) ["$248 billion of additional lease commitments"](0001193125-25-315925|0) ["obligations… primarily related to cloud capacity arrangements"](0001193125-25-315925|0) --- ## 4) Capital stack: what “financing set” actually means in disclosure terms ### A) Debt funding: repeated access at scale Oracle shows significant debt issuance activity (and a pattern of refinancing/maturity management). * In **Sept 2025**, Oracle issued **$18.0B** of fixed-rate senior notes (net $17.882B), for general corporate purposes including capex and debt repayment. ["issued $18.0 billion par value"](0001193125-25-315925|0) ["general corporate purposes, which may include capital expenditures"](0001193125-25-315925|0) * In the **first half of FY2026**, cash flow statement shows **$17.880B proceeds from issuances of senior notes and term loan credit agreements (net of issuance costs)**. ["Proceeds from issuances of senior notes and term loan credit agreements… 17,880"](0001193125-25-315925|0) * On **Feb 4, 2026**, Oracle consummated a large multi-tranche notes offering (including maturities out to 2066), with proceeds for general corporate purposes including capex. ["consummated the issuance and sale"](0001193125-26-037840|0) ["may include capital expenditures"](0001193125-26-037840|0) **Hidden alpha angle:** issuing very long-dated notes while simultaneously disclosing long-dated capacity commitments suggests Oracle is matching asset duration to liability duration—reducing refinancing risk for a buildout that will not pay back in a single cycle. ["terms of fifteen to nineteen years"](0001193125-25-315925|0) ["Notes due 2066"](0001193125-26-037840|0) ### B) Equity optionality: ATM program as a “pressure valve” Oracle put in place an **ATM program up to $20.0B**, with sales agent commissions up to 0.50%. * ATM size and structure: ["aggregate sales proceeds of up to $20.0 billion"](0001193125-26-037840|0) * Cost (agent commission): ["commission of up to 0.50%"](0001193125-26-037840|0) **Hidden alpha angle:** An ATM isn’t just “dilution risk.” It is a *liquidity and credit-rating defense mechanism*: it can be turned on opportunistically (or not used at all) to keep leverage/ratings stable during peak capex years, especially if debt markets get punitive. The disclosure explicitly notes Oracle can suspend/terminate and is not obligated to sell shares. ["not obligated to sell any Shares"](0001193125-26-037840|0) ["may at any time suspend"](0001193125-26-037840|0) ### C) Mandatory convertible preferred: equity-like cash today, dilution tomorrow—with guardrails Oracle issued **depositary shares** representing interests in **6.50% Series D Mandatory Convertible Preferred**, with conversion into common stock by Jan 2029 based on VWAP mechanics. * Offering closed: ["Depositary Shares Offering closed on February 5, 2026"](0001193125-26-039344|0) * Dividend rate and period: ["accumulate dividends… at an annual… rate of 6.50%"](0001193125-26-039344|0) * Automatic conversion window and bounds: ["automatically convert… into between 499.8126 and 624.7657 shares"](0001193125-26-039344|0) **The capital stack “catch” (often missed): distribution/repurchase restrictions** The preferred terms include a key protection for preferred holders: **restrictions on common dividends and repurchases** if preferred dividends are in arrears (subject to exceptions). * Summary in 8‑K: ["no dividend or distribution will be declared or paid… and no Common Stock… purchased… unless all accumulated and unpaid dividends… have been declared and paid"](0001193125-26-039344|0) * Detailed term in Certificate of Designations: ["no dividend… on Common Stock… and no Common Stock… purchased, redeemed… unless all accumulated and unpaid dividends… have been declared and paid"](0001193125-26-039344|2) **Hidden alpha angle:** this feature can *implicitly re-prioritize capital returns* (buybacks/dividends) behind buildout funding during stress scenarios—potentially stabilizing credit at the expense of near-term equity yield optics. The market may underprice this governance-like constraint until it matters. ["no Common Stock… purchased, redeemed"](0001193125-26-039344|2) --- ## 5) Is the capital stack still a worry? A filing-based scorecard ### Pressure points (why it can still be a worry) 1. **Capex intensity is explicit and persistent** (management guidance in MD&A). ["We expect this upward trend to continue"](0001193125-25-315925|0) 2. **Off-balance-sheet-like commitments are enormous** ($248B additional lease commitments). ["$248 billion of additional lease commitments"](0001193125-25-315925|0) 3. **Debt load is rising**: non-current notes/borrowings increased to **$99.984B** at Nov 30, 2025 vs **$85.297B** at May 31, 2025 (+$14.687B, computed). ["Notes payable… non-current… 99,984"](0001193125-25-315925|0) ["Notes payable… non-current… 85,297"](0001193125-25-315925|0) ### Offsetting mitigants (why it may be manageable) 1. **Cash balance rose** to **$19.241B** at Nov 30, 2025 (from $10.786B at May 31, 2025). ["Cash and cash equivalents… 19,241"](0001193125-25-315925|0) ["Cash and cash equivalents… 10,786"](0001193125-25-315925|0) 2. **Proactive multi-channel financing**: debt issuance + ATM + mandatory convertible preferred, i.e., management is not relying on one market staying open. ["Notes Offering"](0001193125-26-037840|0) ["ATM Offering"](0001193125-26-037840|0) ["Depositary Shares Offering"](0001193125-26-039344|0) 3. **Contracted demand signal (RPO)** is unusually strong versus typical “build first, hope later” capex cycles. ["Remaining performance obligations were $523.3 billion"](0001193125-25-315925|0) --- ## 6) Additional “hidden alpha” signals from disclosures (often overlooked) ### (i) The data center capacity risk is explicitly recognized (and framed as a profitability driver) Oracle discloses a two-sided risk: underbuild (lost business) vs overbuild (locked into multi-year commitments and termination fees), meaning execution on capacity planning is a *first-order driver of margin outcomes*, not a footnote. ["If we underestimate customer demand… lose business"](0000950170-25-087926|0) ["locked into multi-year commitments… resulting in lower profitability"](0000950170-25-087926|0) ### (ii) RPO composition shift implies “bigger contracts” dynamics Oracle flags that seasonal patterns and the timing of large contracts can move RPO materially; the FY2026 jump is linked to “certain significant cloud contracts,” suggesting customer concentration / megadeal dynamics could be increasingly relevant for forecasting. ["timing of booking of large contracts"](0000950170-25-087926|0) ["certain significant cloud contracts"](0001193125-25-315925|0) ### (iii) Capital return flexibility may be structurally reduced during the preferred’s life If preferred dividends ever become constrained, the restrictions can limit common dividends/buybacks (with exceptions), which effectively acts like a contingent capital conservation tool. ["no Common Stock… purchased, redeemed"](0001193125-26-039344|2) --- ## 7) What to monitor next (filing-tethered) 1. **RPO conversion vs. capacity build:** Does deferred revenue/RPO translate into revenue/margin without disproportionate expense growth? ["infrastructure expenses"](0001193125-25-315925|0) 2. **Capex + “unpaid capital expenditures” trend:** Unpaid capex is disclosed as **$6.849B** as of Nov 30, 2025 (non-cash investing activity indicator). ["Unpaid capital expenditures… 6,849"](0001193125-25-315925|0) 3. **Lease/capacity commitments trajectory:** $248B is the new reference point; watch if it rises further. ["$248 billion of additional lease commitments"](0001193125-25-315925|0) 4. **Equity issuance usage:** whether Oracle actually taps the ATM (the agreement is capacity, not a requirement). ["not obligated to sell any Shares"](0001193125-26-037840|0) --- ## Bottom line Based strictly on filings, the long-range operating path looks *demand-supported* (RPO scale and duration), but the capital stack question is legitimate because the buildout is not just capex—it’s also massive long-dated capacity commitments. Oracle’s recent financing actions (multi-tranche notes, ATM capacity, mandatory convertible preferred) read as intentional “over-preparation” for that funding burden, which is constructive for plan continuity—but it comes with real equity trade-offs (future dilution and potential capital return constraints). ["Remaining performance obligations were $523.3 billion"](0001193125-25-315925|0) ["Capital expenditures… $20.5 billion"](0001193125-25-315925|0) ["$248 billion of additional lease commitments"](0001193125-25-315925|0) ["aggregate sales proceeds of up to $20.0 billion"](0001193125-26-037840|0) ["no… Common Stock… purchased, redeemed"](0001193125-26-039344|2)

## Oracle (ORCL) — 2030 trajectory vs. capital stack: what the filings imply ### Executive takeaway Oracle’s disclosures show an unusually large *contracted revenue runway* (remaining performance obligations) alongside an equally unusual *infrastructure funding ramp* (capex + long-dated capacity/lease commitments). The “hidden alpha” is that these two items move together: the RPO step-up helps de-risk demand into the buildout, while the capital stack actions (notes, ATM equity capacity, mandatory convertible preferred) suggest management is intentionally pre-funding the build to avoid an uncontrolled leverage spike. ["Remaining performance obligations were $523.3 billion"](0001193125-25-315925|0) ["Capital expenditures… increased… to $20.5 billion"](0001193125-25-315925|0) ["sell shares… up to $20.0 billion"](0001193125-26-037840|0) ["Depositary Shares Offering closed"](0001193125-26-039344|0) --- ## 1) The core debate: “2030 path intact” vs. “capital stack still a worry” ### What the filings *support* (path intact) The strongest filing-based evidence that the multi-year growth plan has real demand behind it is the sheer scale and duration profile of contracted commitments. * **RPO is $523.3B as of Nov 30, 2025** with a multi-year recognition schedule (only ~10% expected to recognize in the next 12 months, then 30% in months 13–36, 35% in months 37–60, and the remainder thereafter). ["Remaining performance obligations were $523.3 billion"](0001193125-25-315925|0) ["expect to recognize approximately 10%… next twelve months… remainder thereafter"](0001193125-25-315925|0) * Management explicitly attributes the Y/Y jump in RPO to **“certain significant cloud contracts”** entered during the period, i.e., not merely accounting noise. ["primarily attributable to certain significant cloud contracts"](0001193125-25-315925|0) **Hidden alpha angle:** RPO at this magnitude is not just backlog—it is *financing collateral in disguise*: it strengthens Oracle’s ability to raise long-dated capital at scale because it evidences forward cash-generation potential (even if revenue recognition lags). ["Remaining performance obligations were $523.3 billion"](0001193125-25-315925|0) ### What the filings *do not let you dismiss* (capital stack worry) Oracle’s disclosures also show the cost of fulfilling that demand is front-loaded and massive. * **Capex surged to $20.5B in the first half of FY2026**, explicitly tied to data center expansion, with management expecting the upward trend to continue for “the next few fiscal years.” ["Capital expenditures… increased… to $20.5 billion"](0001193125-25-315925|0) ["We expect this upward trend to continue… next few fiscal years"](0001193125-25-315925|0) * The balance sheet shows **property, plant & equipment (net) rising to $67.875B** at Nov 30, 2025 vs $43.522B at May 31, 2025 (a +$24.353B increase, computed), consistent with the capex ramp. ["Property, plant and equipment, net… 67,875"](0001193125-25-315925|0) ["Property, plant and equipment, net… 43,522"](0001193125-25-315925|0) * Oracle discloses **very large off-balance-sheet style capacity commitments**: as of Nov 30, 2025 it had **$248B of additional lease commitments** (substantially all related to data centers and cloud capacity arrangements), generally commencing between Q3 FY2026 and FY2028, for 15–19 year terms. ["we had $248 billion of additional lease commitments"](0001193125-25-315925|0) ["terms of fifteen to nineteen years"](0001193125-25-315925|0) **Hidden alpha angle:** the $248B figure reframes the “capex cycle” as only one leg of the buildout—Oracle is also locking in multi-decade capacity via leases/capacity arrangements, which can behave economically like leverage even if accounting treatment differs. ["$248 billion of additional lease commitments"](0001193125-25-315925|0) --- ## 2) Demand visibility: reading RPO like an analyst (not a headline) ### Why RPO quality matters more than RPO size RPO can be inflated by short-duration contracts; Oracle’s disclosure suggests the opposite: it’s long duration by design. * Only **~10%** recognized in the next 12 months implies the backlog is not “about to churn”; it is structurally long-dated. ["expect to recognize approximately 10%… over the next twelve months"](0001193125-25-315925|0) ### The “invisible” implication for 2030 thinking Even without explicit “2030 targets” in the filings, the time-bucketed schedule plus “remainder thereafter” strongly implies Oracle is stacking obligations well beyond a typical 3-year planning window—useful for underwriting a multi-year thesis if execution holds. ["30%… month 13 to month 36… 35%… month 37 to month 60… remainder thereafter"](0001193125-25-315925|0) --- ## 3) Infrastructure buildout: the tell is not capex—it’s commitments ### Capex: the visible build * FY2026 first-half capex: **$20.535B**. ["Capital expenditures… (20,535)"](0001193125-25-315925|0) ### Commitments: the less visible build * **$248B additional lease commitments** (data centers/cloud capacity), starting mainly FY2026–FY2028, 15–19 years. ["$248 billion of additional lease commitments"](0001193125-25-315925|0) * Plus **$10B unconditional purchase and other obligations** as of Nov 30, 2025, primarily related to cloud capacity arrangements. ["unconditional purchase and certain other obligations… were $10 billion"](0001193125-25-315925|0) **Hidden alpha angle:** the combination of (1) capex, (2) long-dated lease commitments, and (3) capacity-related purchase obligations is effectively a “capacity flywheel.” The market often prices only capex; the filings show the true scale includes long-term fixed-ish commitments that will pressure free cash flow timing even if margins ultimately expand. ["Capital expenditures… increased… to $20.5 billion"](0001193125-25-315925|0) ["$248 billion of additional lease commitments"](0001193125-25-315925|0) ["obligations… primarily related to cloud capacity arrangements"](0001193125-25-315925|0) --- ## 4) Capital stack: what “financing set” actually means in disclosure terms ### A) Debt funding: repeated access at scale Oracle shows significant debt issuance activity (and a pattern of refinancing/maturity management). * In **Sept 2025**, Oracle issued **$18.0B** of fixed-rate senior notes (net $17.882B), for general corporate purposes including capex and debt repayment. ["issued $18.0 billion par value"](0001193125-25-315925|0) ["general corporate purposes, which may include capital expenditures"](0001193125-25-315925|0) * In the **first half of FY2026**, cash flow statement shows **$17.880B proceeds from issuances of senior notes and term loan credit agreements (net of issuance costs)**. ["Proceeds from issuances of senior notes and term loan credit agreements… 17,880"](0001193125-25-315925|0) * On **Feb 4, 2026**, Oracle consummated a large multi-tranche notes offering (including maturities out to 2066), with proceeds for general corporate purposes including capex. ["consummated the issuance and sale"](0001193125-26-037840|0) ["may include capital expenditures"](0001193125-26-037840|0) **Hidden alpha angle:** issuing very long-dated notes while simultaneously disclosing long-dated capacity commitments suggests Oracle is matching asset duration to liability duration—reducing refinancing risk for a buildout that will not pay back in a single cycle. ["terms of fifteen to nineteen years"](0001193125-25-315925|0) ["Notes due 2066"](0001193125-26-037840|0) ### B) Equity optionality: ATM program as a “pressure valve” Oracle put in place an **ATM program up to $20.0B**, with sales agent commissions up to 0.50%. * ATM size and structure: ["aggregate sales proceeds of up to $20.0 billion"](0001193125-26-037840|0) * Cost (agent commission): ["commission of up to 0.50%"](0001193125-26-037840|0) **Hidden alpha angle:** An ATM isn’t just “dilution risk.” It is a *liquidity and credit-rating defense mechanism*: it can be turned on opportunistically (or not used at all) to keep leverage/ratings stable during peak capex years, especially if debt markets get punitive. The disclosure explicitly notes Oracle can suspend/terminate and is not obligated to sell shares. ["not obligated to sell any Shares"](0001193125-26-037840|0) ["may at any time suspend"](0001193125-26-037840|0) ### C) Mandatory convertible preferred: equity-like cash today, dilution tomorrow—with guardrails Oracle issued **depositary shares** representing interests in **6.50% Series D Mandatory Convertible Preferred**, with conversion into common stock by Jan 2029 based on VWAP mechanics. * Offering closed: ["Depositary Shares Offering closed on February 5, 2026"](0001193125-26-039344|0) * Dividend rate and period: ["accumulate dividends… at an annual… rate of 6.50%"](0001193125-26-039344|0) * Automatic conversion window and bounds: ["automatically convert… into between 499.8126 and 624.7657 shares"](0001193125-26-039344|0) **The capital stack “catch” (often missed): distribution/repurchase restrictions** The preferred terms include a key protection for preferred holders: **restrictions on common dividends and repurchases** if preferred dividends are in arrears (subject to exceptions). * Summary in 8‑K: ["no dividend or distribution will be declared or paid… and no Common Stock… purchased… unless all accumulated and unpaid dividends… have been declared and paid"](0001193125-26-039344|0) * Detailed term in Certificate of Designations: ["no dividend… on Common Stock… and no Common Stock… purchased, redeemed… unless all accumulated and unpaid dividends… have been declared and paid"](0001193125-26-039344|2) **Hidden alpha angle:** this feature can *implicitly re-prioritize capital returns* (buybacks/dividends) behind buildout funding during stress scenarios—potentially stabilizing credit at the expense of near-term equity yield optics. The market may underprice this governance-like constraint until it matters. ["no Common Stock… purchased, redeemed"](0001193125-26-039344|2) --- ## 5) Is the capital stack still a worry? A filing-based scorecard ### Pressure points (why it can still be a worry) 1. **Capex intensity is explicit and persistent** (management guidance in MD&A). ["We expect this upward trend to continue"](0001193125-25-315925|0) 2. **Off-balance-sheet-like commitments are enormous** ($248B additional lease commitments). ["$248 billion of additional lease commitments"](0001193125-25-315925|0) 3. **Debt load is rising**: non-current notes/borrowings increased to **$99.984B** at Nov 30, 2025 vs **$85.297B** at May 31, 2025 (+$14.687B, computed). ["Notes payable… non-current… 99,984"](0001193125-25-315925|0) ["Notes payable… non-current… 85,297"](0001193125-25-315925|0) ### Offsetting mitigants (why it may be manageable) 1. **Cash balance rose** to **$19.241B** at Nov 30, 2025 (from $10.786B at May 31, 2025). ["Cash and cash equivalents… 19,241"](0001193125-25-315925|0) ["Cash and cash equivalents… 10,786"](0001193125-25-315925|0) 2. **Proactive multi-channel financing**: debt issuance + ATM + mandatory convertible preferred, i.e., management is not relying on one market staying open. ["Notes Offering"](0001193125-26-037840|0) ["ATM Offering"](0001193125-26-037840|0) ["Depositary Shares Offering"](0001193125-26-039344|0) 3. **Contracted demand signal (RPO)** is unusually strong versus typical “build first, hope later” capex cycles. ["Remaining performance obligations were $523.3 billion"](0001193125-25-315925|0) --- ## 6) Additional “hidden alpha” signals from disclosures (often overlooked) ### (i) The data center capacity risk is explicitly recognized (and framed as a profitability driver) Oracle discloses a two-sided risk: underbuild (lost business) vs overbuild (locked into multi-year commitments and termination fees), meaning execution on capacity planning is a *first-order driver of margin outcomes*, not a footnote. ["If we underestimate customer demand… lose business"](0000950170-25-087926|0) ["locked into multi-year commitments… resulting in lower profitability"](0000950170-25-087926|0) ### (ii) RPO composition shift implies “bigger contracts” dynamics Oracle flags that seasonal patterns and the timing of large contracts can move RPO materially; the FY2026 jump is linked to “certain significant cloud contracts,” suggesting customer concentration / megadeal dynamics could be increasingly relevant for forecasting. ["timing of booking of large contracts"](0000950170-25-087926|0) ["certain significant cloud contracts"](0001193125-25-315925|0) ### (iii) Capital return flexibility may be structurally reduced during the preferred’s life If preferred dividends ever become constrained, the restrictions can limit common dividends/buybacks (with exceptions), which effectively acts like a contingent capital conservation tool. ["no Common Stock… purchased, redeemed"](0001193125-26-039344|2) --- ## 7) What to monitor next (filing-tethered) 1. **RPO conversion vs. capacity build:** Does deferred revenue/RPO translate into revenue/margin without disproportionate expense growth? ["infrastructure expenses"](0001193125-25-315925|0) 2. **Capex + “unpaid capital expenditures” trend:** Unpaid capex is disclosed as **$6.849B** as of Nov 30, 2025 (non-cash investing activity indicator). ["Unpaid capital expenditures… 6,849"](0001193125-25-315925|0) 3. **Lease/capacity commitments trajectory:** $248B is the new reference point; watch if it rises further. ["$248 billion of additional lease commitments"](0001193125-25-315925|0) 4. **Equity issuance usage:** whether Oracle actually taps the ATM (the agreement is capacity, not a requirement). ["not obligated to sell any Shares"](0001193125-26-037840|0) --- ## Bottom line Based strictly on filings, the long-range operating path looks *demand-supported* (RPO scale and duration), but the capital stack question is legitimate because the buildout is not just capex—it’s also massive long-dated capacity commitments. Oracle’s recent financing actions (multi-tranche notes, ATM capacity, mandatory convertible preferred) read as intentional “over-preparation” for that funding burden, which is constructive for plan continuity—but it comes with real equity trade-offs (future dilution and potential capital return constraints). ["Remaining performance obligations were $523.3 billion"](0001193125-25-315925|0) ["Capital expenditures… $20.5 billion"](0001193125-25-315925|0) ["$248 billion of additional lease commitments"](0001193125-25-315925|0) ["aggregate sales proceeds of up to $20.0 billion"](0001193125-26-037840|0) ["no… Common Stock… purchased, redeemed"](0001193125-26-039344|2)