ORCL — 66% OFF THE HIGH AND BACK AT THE 2024 SHELFOracle CorporationBATS:ORCLJahirS Oracle ran from roughly 60 in mid-2022 to about 345 in late 2025 — an AI-infrastructure re-rating that briefly took the company near a trillion dollars in market value. It has since given back around 66%, and at 118.20 roughly 630B of that market cap is gone. Price is now back inside the 100–125 zone it consolidated in through 2023 and 2024. That is the first genuinely structural level on the way down, and it is the reason this chart is worth a post rather than a shrug. Everything above here — the 175 breakdown, the 230 bounce high — failed. This is where the move started. The question is not whether 100–125 is support. It clearly is, historically. The question is whether a shelf built during a lower-capex, higher-margin era still applies to a balance sheet that now carries 3.94x debt-to-equity. STRUCTURE I'M WATCHING Support shelf ....... 100 – 125 ... The 2023–24 consolidation. Price is inside it now. Hard invalidation ... 100 ........ Round number and the floor of the shelf. Next support ........ 82 – 85 .... The 2021–22 pivot. Where a shelf failure would likely head. Cycle low ........... 60 – 63 .... The 2022 base. Full round trip scenario. First resistance .... 175 – 180 .. Early-2026 breakdown. First real test of any recovery. Second resistance ... 230 – 240 .. The spring-2026 bounce high. Reclaiming it changes the structure. Old high ............ 345 ........ Not a level to plan around from here. On the projection to 400: that path requires a new all-time high and a full re-validation of the AI earnings thesis. Forward estimates are currently moving the other way (see below). I would not publish it as a target and I would not size against it. WHAT THE CHART IS TELLING ME - The shelf is real and it is wide. Two years of two-way volume between 100 and 125 means genuine price memory here, not a single wick. - A 66% mega-cap drawdown in under a year is a regime change, not a dip. The market is repricing the thesis, not just the multiple. - Every bounce has failed lower. 345 > 175 > 240 > 118 is a sequence of lower highs. Support tests inside that sequence fail more often than they hold. - Volume expanded on the decline through 2026. That is distribution, and it is heavier than anything seen on the way up. - Risk is at least definable. From 118 with invalidation at 100 and a first objective at 175, the geometry is roughly 3:1. That is the strongest argument the chart makes. FUNDAMENTAL BACKDROP — WHERE THE TENSION SITS Market cap ......... 340.4B ... Down from roughly 970B at the peak P/E trailing ....... 20.29 .... Reasonable on delivered earnings P/E forward ........ 30.15 .... HIGHER than trailing — consensus expects earnings to FALL Net margin ......... 25.37% ... Still a genuinely high-quality business ROIC ............... 11% ...... Respectable, above cost of capital Debt / equity ...... 3.94 ..... The problem. Debt-funded datacentre capex Price / book ....... 9.06 ..... No asset-value floor Dividend yield ..... 0.89% .... Token support, not a reason to own it Piotroski F-score .. 4 / 9 .... Mediocre; the leverage and margin trend weigh on it The bull and bear case sit in the same table. Net margin above 25% and ROIC of 11% describe a business most companies would envy. But a forward multiple ABOVE the trailing multiple means the market is pricing earnings decline, and 3.94x debt-to-equity means that decline arrives on a leveraged balance sheet with rising depreciation from the capex build. That is the whole story: Oracle borrowed heavily to buy growth, the growth is being questioned, and the depreciation is not optional. Verify the latest quarter, RPO conversion and capex guidance yourself — that is what decides which side of the table wins. INVALIDATION A decisive weekly close below 100 ends the idea. That breaks a two-year shelf and opens 82–85, then the 60–63 base. On this leverage profile I would not average into that. Softer warning: repeated weekly closes inside 100–125 with no volume expansion means the shelf is being worn down rather than defended. RISKS WORTH NAMING - Catching a falling mega-cap. Nothing on this chart says the decline is finished. A held level is not a reversal. - Leverage. At 3.94x debt-to-equity, an earnings miss hits equity far harder than it would have in 2023. - The capex cycle is the thesis. If AI infrastructure demand or pricing disappoints, depreciation stays and revenue does not. - Concentration risk. Large backlog concentrated in a small number of counterparties turns customer-specific news into ORCL-specific news. - Forward estimates can keep falling. A 30x forward multiple on declining estimates gets more expensive, not cheaper, as price drops. TRADE PLAN Entry ......... Weekly reversal with volume expansion inside 100 – 125 Invalidation .. Weekly close below 100 Target 1 ...... 175 – 180 Target 2 ...... 230 – 240 Geometry ...... Roughly 3:1 from 118 to 175 against a stop at 100 Sizing ........ Defined-risk level trade. Not a position to add to on weakness No position on a drift lower without a volume signal. Not financial advice — my own chart interpretation, shared for discussion. Fundamental figures are read from a data panel and should be verified against current filings before acting. Manage your own risk. Tags: ORCL, Oracle, AI, Datacenter, CloudComputing, WeeklyChart, SupportResistance, Drawdown, Valuation, Leverage, RiskManagement