Oracle Booked $638 Billion of AI Demand. The Stock Fell 8.5% the Next Day.

By Kresmion Research · June 12, 2026

Oracle closed its fiscal year with $638 billion of contracted future revenue and a record quarter, and the stock fell 8.5 percent the first day it traded. The report landed after Wednesday's close: fiscal Q4 revenue of $19.2 billion, up 21 percent, cloud revenue up 47 percent to $9.9 billion, GAAP earnings of $1.45 per share, and remaining performance obligations, the contracted revenue Oracle has booked but not yet recognized, at $638 billion, up 363 percent in a year (Oracle press release). On Thursday, while the S&P 500 rose 1.75 percent to 7,394.30 on a reported US-Iran de-escalation (Yahoo Finance), Oracle closed at $184.10, down 8.5 percent from Wednesday's $201.26 (Yahoo Finance historical data).

A stock falling that hard against a tape rising that fast is a specific kind of message. The market's objection was not demand. It was the bill for serving it.

Key takeaways

A record quarter, sold

The headline numbers were as strong as any in Oracle's history, and the selling concentrated on what funds them. Capital expenditure for the fiscal year came in at $55.7 billion, up 162 percent from the prior year, and the company signaled roughly $40 billion of additional debt and equity financing for fiscal 2027 to keep building data-center capacity against its AI contracts. Earnings coverage, citing Bank of America estimates, put more than half of the $638 billion obligation pool with a single customer, OpenAI (Blockonomi earnings coverage).

That concentration is the trade in one line: Oracle has contracted nearly two-thirds of a trillion dollars of future revenue, and delivering it requires capital spending large enough that the company plans to raise roughly $40 billion of new financing on top of it. Thursday's 8.5 percent decline, on a day the Dow rose more than 900 points, priced the financing, not the franchise.

What the Q1 filings showed: trimming, not fleeing

Kresmion's institutional database holds the Q1 2026 13F filings, the fund positions disclosed as of March 31. The clean Oracle common-stock rows show four large multi-strategy funds reducing in the same quarter:

Every row carries the same implied mark of $147.11 per share, which is exactly Oracle's real March 31 close (Yahoo Finance historical data), a consistency check that gives us confidence in the rows.

Three honest caveats before reading anything into this. First, these are modest cuts, between 5 and 15 percent, on positions that remain in the hundreds of millions of dollars: trimming, not fleeing. Second, a 13F is a quarter-end snapshot filed up to 45 days later. It says nothing about what these funds did in April, May, or this week. Third, the stock went the other way first: Oracle ran from $147.11 at the March quarter-end to a closing high of $248.15 on June 1 before round-tripping to $184.10. Whoever trimmed in Q1 watched the stock rise 69 percent past their mark before this week's repricing. This is positioning context, not foresight, and it is not a recommendation.

A signal we discarded on the way here

Transparency note: our screens initially surfaced what looked like an eight-fund Oracle accumulation cluster in the same Q1 filings. On verification it fell apart. The ticker-level grouping had merged three different security identifiers, one of which belongs to a different issuer entirely, OR Royalties, a gold royalty company whose identifier sorts adjacent to Oracle's. The clean Oracle common-stock rows, grouped by CUSIP and price-checked against the real March 31 close, show the reduction reported above. We publish what survives the check.

Elsewhere in the Q1 filings

The same filing set produced one convergence that did survive verification: seven funds added or initiated Blackstone in Q1, worth a combined $731.3 million at the quarter-end mark. D.E. Shaw went from 31,481 shares to 3,781,729 ($434.9 million), Renaissance Technologies initiated 1,431,400 shares ($164.6 million), Two Sigma and Soros Fund Management opened new positions, and Millennium, Point72, and Bridgewater all added, with every row marked at $114.99, Blackstone's real March 31 close. A smaller five-fund accumulation showed up in Broadcom, led by D.E. Shaw's increase to $2.02 billion. The same caveats apply: quarter-end marks, not cost bases, disclosed with a lag.

The wider market context

The tape Oracle fell against was strong and the inflation data behind it was not. May producer prices rose 1.1 percent on the month and 6.5 percent year over year, the largest 12-month increase since November 2022, with gasoline up 23.4 percent (BLS). The first FOMC meeting under chair Kevin Warsh is next week, June 16-17, and Kresmion's prediction-market read has a June hold priced at 98.45 percent and a July quarter-point cut at 47.57 percent, roughly half what this market priced earlier this month (single-venue markets, so one book's print rather than a cross-venue consensus).

Kresmion's macro regime score, which compresses growth, liquidity, risk appetite, and volatility readings into one number, recovered to -0.149 on June 12 from -0.303 on June 10, still Neutral with HIGH conviction. Growth is the lone positive factor at +0.44; the rebound this week came from the liquidity and volatility drags roughly halving after the CPI print. In crypto, US spot Bitcoin ETFs saw a net $22.5 million outflow on June 11, the twelfth negative session in thirteen (Farside Investors), Bitcoin traded near $63,800 (CoinGecko), and cross-venue odds of Bitcoin above $110,000 by year-end sat at 13.8 percent across Kalshi and Polymarket. Yesterday's note covered the on-chain side: The Whale Tape's Biggest Genuine Move This Week Was Solana, Not Bitcoin.

For how Kresmion separates verified data from narrative, see our methodology. For the fund-by-fund 13F view, the Oracle research page and Blackstone research page carry the underlying positions, and the related June 4 note on the Visa 13F split covers how we read divided positioning.

Frequently asked questions

Why did Oracle's stock fall after record results?

The results were records: $19.2 billion of quarterly revenue and $638 billion of contracted future revenue. The selling pressure followed the cost of delivery, $55.7 billion of fiscal-2026 capital expenditure, up 162 percent, plus roughly $40 billion of new financing signaled for fiscal 2027. The stock closed down 8.5 percent on June 11 while the S&P 500 rose 1.75 percent.

Did big funds sell Oracle before the June 2026 drop?

Q1 2026 13F filings show four multi-strategy funds, Millennium, D.E. Shaw, Bridgewater, and Point72, reduced Oracle positions by 5 to 15 percent as of March 31. That is a quarter-end snapshot disclosed with a lag of up to 45 days, and the stock first rose 69 percent past the March mark before falling, so it shows positioning, not prediction.

What are remaining performance obligations?

Remaining performance obligations, or RPO, are revenue a company has under signed contracts but has not yet recognized. Oracle reported $638 billion of RPO for fiscal Q4 2026, up 363 percent year over year, driven by AI infrastructure contracts. RPO converts to revenue only as capacity is built and delivered, which is why the capital spending required to serve it became the market's focus.

Sources

• Oracle Q4 FY2026 results, press release via StockTitan
• Yahoo Finance, ORCL historical prices
• Yahoo Finance live blog, June 11, 2026
• Blockonomi, Oracle Q4 2026 earnings coverage
• BLS, Producer Price Index, May 2026
• Farside Investors, Bitcoin ETF flow table
• CoinGecko, Bitcoin
• Kresmion institutional 13F database, regime engine, and prediction-market consensus (queried June 12, 2026)

Kresmion publishes event framing and verified data, not investment advice. Nothing here is a recommendation to buy or sell any security.