Oracle's $988M CEO Pay Went Underwater in Eight Months — Now the Board Is Trapped Inside Its Own Strike Prices

The proxy lays out the timing cleanly. Clay Magouyrk and Mike Sicilia were promoted to co-CEO on September 22, 2025 — nine days after the $300 billion OpenAI compute deal sent Oracle Corp shares to their fiscal-year high. Their one-time promotional grants were struck on September 24, 2025, at an exercise price of $308.46, just under the prevailing market. Ellison received his annual grant on October 23, 2025, at $280.07. By May 31, 2026, the proxy explicitly states, \"the stock options, including the PSOs, had no intrinsic value at fiscal year end.\"

What the committee thought it was pricing was a cloud-AI multiple: $300B of contracted OpenAI demand, a $455B remaining performance obligations figure from Q1 FY26 that was already up 359% year-over-year, and 92% growth in cloud infrastructure. The market agreed, briefly. Oracle Corp's stock traded above $322 in the 52 weeks ended this week. By Friday, it closed at $137.10 — a 57% drop from peak.

The package is backloaded. Both Magouyrk and Sicilia's awards are 80% time-based stock options and 20% performance stock options (PSOs) tied to revenue metrics over a three-year performance period ending May 31, 2028. Even the time-vesting portion lapses only if shares clear the strike — and at $137, every share the executives could exercise is currently worthless. The compensation committee's intended target value is calculated by dividing the approved equity value by the closing price on the grant date; Black-Scholes drove the grant-date fair value higher, which is why Magouyrk's $621.7M grant-date value exceeded the $250M headline target announced on his promotion.

The options weren't struck on hope — they were struck on a specific thesis about how Oracle Corp would convert its $638B RPO into operating cash flow. That conversion is going the wrong way. The FY26 10-K shows capital expenditures of $55.66 billion against operating cash flow of $31.98 billion, leaving free cash flow at negative $23.69 billion. Capex consumed 82.6% of total revenue, the highest ratio in any Oracle fiscal year on record, and 174% of operating cash flow.

The cash burn forced Oracle to raise $43 billion in debt and $5 billion in equity during fiscal 2026, and management has guided to a further $45–$50 billion of gross cash proceeds in calendar 2026 — including a $20 billion at-the-market equity program. This is what makes the underwater options a board-level issue rather than a comp-table footnote: the comp committee priced governance on a multiple that required capex to convert into revenue on schedule. Capex is on schedule; revenue is too — the question is whether the multiple still belongs to the old model.

The proxy reveals three governance facts that compound the comp story. First, Ellison personally has 413 million Oracle Corp shares pledged as collateral for personal indebtedness as of September 21, 2026 — a 19% increase over2025, or roughly $57 billion of notional exposure at Friday's close. Second, on September 12, 2026, Ellison cancelled a 10b5-1 plan that would have allowed him to sell up to 50 million shares (~$7.5 billion at the time of adoption) — the day after the plan was disclosed. Third, the same proxy holds the comp committee's stated rationale that the package combines time-based and performance-based vesting tied to \"Oracle's long-term financial performance,\" with the PSOs tied to revenue metrics over a period ending May 31, 2028.

Read together, the underwater options are not a moral-hazard problem — the executives cannot monetize them. They are an information problem: the board paid for outcomes that required a specific capex-to-revenue conversion path, the capex is happening, the revenue is arriving, and the multiple nevertheless compressed. The November 18, 2026 annual meeting — already on the calendar — will vote on director elections and the compensation committee's report. With the proxy disclosing413 million pledged shares alongside $988 million of underwater options, that vote becomes the closest thing to a shareholder referendum on whether the FY26 capex decision was itself well-priced.

The PSOs vest on a three-year performance period ending May 31, 2028 — the same horizon as Oracle Corp's stated FY28 outlook window. For those PSOs to clear, two things must be true simultaneously. First, the FY27 capex of $90–$95B must convert into RPO-to-revenue at a rate that returns Oracle to positive free cash flow by FY28. Second, the equity multiple must re-rate above the strike levels, requiring either an acceleration in IaaS revenue (Q1 FY27's 121% growth is the leading indicator) or a market-wide re-rating of cloud-infrastructure peers. Neither is impossible; both are hostage to the same execution.

The peer table is the structural risk. Microsoft spent roughly twice as much on AI infrastructure in FY26, generated 2.4x as much capex-to-revenue headroom, and still produced $67 billion of free cash flow — enough to fund its dividend, buybacks, and the AI buildout without dilution. Oracle Corp spent a smaller absolute capex figure but at a multiple of its revenue that produced a $90 billion FCF gap to the hyperscaler benchmark. To clear the $308.46 strike by2028, Oracle has to either catch up to Microsoft's revenue scale, narrow the capex-to-revenue ratio by 50 points, or convince the market that a negative-FCF cloud is still a hyperscaler.

The $988 million underwater options and the $55.7 billion capex bill are the same trade. The board authorized one and paid for the other at a multiple that compressed. The fact that the compensation package struck at the peak is, by itself, an industry-wide phenomenon — peers have re-priced comp similarly — but Oracle's compression is sharper because the capex intensity is higher. The market is pricing Oracle Corp for a slower RPO-to-cash conversion than the compensation committee priced in.

For investors, the trade breaks into two questions. First, does Oracle convert? Q1 FY27's 121% IaaS growth and $664B RPO say yes on revenue, but the same release confirmed $28.5B of capex in a single quarter and a maintained $90–$95B FY27 guide — the FCF trough is getting deeper before it gets better. Second, does the board re-price governance? The proxy vote on November 18 is the first formal opportunity; if shareholders ratify the comp committee's actions, the FY28 PSO cliff will arrive with the same strike prices and the same capex thesis. If they don't, the FY27 capex commitment becomes the next governance fight, and the comp package becomes the exhibit.