What I'm seeing

ORCL peaked at $345.72 on Sep 10, 2025, then bled 57% to a low of $114.50 in July 2026 a textbook Weinstein Stage 4 markdown, and it happened while the AI-infrastructure story ORCL once led turned against it. The current bounce to ~$147 is real, but unproven:

57% drawdown from the ATH one of the sharpest de-ratings in the name's history
~28% rip off the $114.50 low in roughly two weeks violent, but that's exactly what bounces inside downtrends look like
Price has reclaimed the lower end of the wholesale zone (~$135–$193), but still sits below every meaningful reclaim level
The average buyer from the $200–$345 range is deeply underwater — overhead supply stacked at every level above
No confirmed Stage 2. A bounce is not a base. A base is not a breakout.

Why it matters

The fundamentals did NOT cause this. FY2026 revenue grew ~17% to roughly $67B. Earnings rose ~36%. Cloud infrastructure was compounding north of 50% YoY. And in the middle of the drawdown, Oracle booked a reported ~$300B OpenAI cloud deal and a $7B Pentagon contract the largest backlog in company history.

So what actually got repriced? Not the top line the cost of it. The market de-rated ORCL on AI-capex intensity: the fear that the spend required to build this cloud empire compresses margins and free cash flow before the revenue proves the ROI. Stack cost-overrun reports and a fresh Michael Burry short on top of that, and you get a narrative repricing sitting on top of legitimate margin questions.

The levels:

$135 — line in the sand (wholesale floor). Held the July structure. Lose it convincingly → $119, then a retest of the $114.50 low.

$156–165 — first real test. Prior breakdown congestion, plus the reclaim level that flips short-term structure. Nothing changes until this goes.

$193 — top of the wholesale zone. Reclaim and hold here and the discount thesis graduates toward equilibrium ($229+) where Wall Street's average target ($258.50) actually lives.

What I expect

Default assumption after a 57% markdown: overhead supply exists at every level above, and rallies get sold until proven otherwise. The $114 → $147 bounce has proven a floor, not a trend.

Two conditions, no predictions:

Weekly acceptance above ~$156 = first evidence the repricing is exhausting and a base is building.
Convincing loss of $135 = the wholesale thesis is early; the $119–$114 retest is back on the table.

Here's the part I keep coming back to: even Wall Street's single most bearish analyst targets $145 basically today's price, and the floor of the wholesale zone. The consensus average ($258.50) sits inside the retail zone. So the map lets you accumulate where the bears' worst case lives, and distribute where the crowd's target lives. That's what favorable risk/reward looks like on a chart. (Strong Buy — 27 Buy / 4 Hold / 0 Sell, 31 analysts.)

Note the calendar: earnings land mid-September. Known volatility event size accordingly.

Strong business + reversing narrative + reclaimed structure earns real attention. It does not earn full size. Confirmation earns size.

Patience is a position. Not financial advice educational breakdown of my process.