Oracle's Bull Case Is Airtight. Its Chart Is a Crime Scene.

Here's a fun exercise: read our own deep-research report on Oracle and you'd think this thing was headed to $268. Then look at the tape and Oracle closed within pennies of its 52-week low, down roughly 65% from the $345 peak it hit less than a year ago. Both things are true at the same time, which tells you everything about where this stock actually is: caught between a genuinely good story and a market that has stopped listening to stories.

Let's start with the story, because it's not fake. Q4 FY2026 revenue came in at $19.2 billion, up 21% year-over-year. OCI — the cloud infrastructure business that's supposed to be Oracle's ticket into the AI arms race — grew 93%, hitting $9.3 billion and now accounting for more than half of total revenue. Remaining performance obligations, the backlog that's supposed to convert into years of future revenue, sit at $638 billion, up from $553 billion just one quarter earlier. Add a fresh 10-year, up-to-$7 billion Department of War contract and a Stargate joint venture with OpenAI and SoftBank anchoring AI capacity out of Abilene, Texas, and you've got the makings of a genuine platform shift — "database vendor" morphing into "AI infrastructure landlord." Forward P/E of 11 and a PEG of 0.42 look, on paper, almost embarrassingly cheap for that kind of growth.

So why is the stock sitting at $119.44 instead of $249?

Because growth like that isn't free, and Oracle is paying for it in a currency the market cares about more than backlog: cash. Capex has exploded from $8.5 billion to $16.5 billion a quarter. Trailing twelve-month free cash flow is negative $24.5 billion. Total debt has ballooned from $105 billion to $156 billion in a single year, against just $42.5 billion of equity — a debt-to-equity ratio pushing 4x. S&P already clipped the credit rating to BBB-, one notch above junk, explicitly citing the AI buildout and customer concentration. Oracle just raised $20 billion in a share sale plus a $40 billion debt-and-equity package to keep funding this thing. That's not a software company printing cash anymore. That's a leveraged infrastructure bet wearing a database company's ticker.

And the chart knows it. Price is more than 30% below the 50-day average and over 36% below the 200-day. RSI sits at 29.6 — technically oversold — except it's been technically oversold for most of the way down from $345, which is the market's polite way of saying "oversold doesn't mean bottom." Every "this is too cheap to ignore" call on Oracle over the last two months has been wrong, directionally, on the tape. That's not a knock on the thesis; it's a warning about timing.

Here's my actual take: the bull case and the bear case are the same case. Oracle's entire valuation now hinges on whether $638 billion of RPO converts into recognized revenue faster than the balance sheet deteriorates. GPU lead times of 3–7 months, a power grid that needs to somehow deliver 2.8 gigawatts to new data centers, and a credit rating one downgrade from junk are all the same risk wearing different hats: can Oracle build the thing before the debt bites? Until that question gets answered with an FCF inflection or credit stabilization — not another contract press release — this stock trades on fear, not backlog math.

So: not a table-pounder, but not a name to ignore either. The Pentagon deal landed at a genuinely extreme technical setup — 2.45% short float, RSI in the high 20s — which is exactly the kind of coiled spring that produces violent squeezes if $119.44 holds. A weekly close below $118 voids that thesis entirely and opens the door to $100, and there's no heroism in catching that on the way down. But a small starter position here, with a hard stop, and real conviction adds only on a reclaim of $128 and then $140, respects both realities at once: Oracle might be building the future of enterprise AI, and it might also be one bad quarter away from the market deciding the leverage isn't worth the story. Own the optionality. Don't marry the narrative.