Insights Article
In September 2025, Oracle Corporation's stock reached an intraday record of $345.72 following the disclosure of a five-year cloud infrastructure agreement with OpenAI valued at more than $300 billion.¹ Three months later, after the company's newly appointed leadership warned that fiscal 2026 capital expenditures would run forty-three percent above the figure provided just months earlier, and as analyst consensus projected gross margins to compress from seventy-seven percent in fiscal 2021 to approximately forty-nine percent by fiscal 2030, Oracle's stock had fallen approximately forty-three percent from that peak, with the fourth-quarter decline on pace to be the company's sharpest quarterly drop since the dot-com collapse of 2001.²
What very few discussions of that drawdown have addressed is what was disclosed in Oracle's definitive proxy statement filed September 26, 2025. As of September 19, 2025, Lawrence J. Ellison, Oracle's Chairman, Chief Technology Officer, and largest stockholder, had pledged 346,000,000 shares of Oracle common stock as collateral to secure certain personal indebtedness—equivalent to roughly thirty percent of his beneficial stake.³ At the September 19 closing price, those pledged shares had a market value of approximately $107 billion.⁴
This is not a trivia footnote. It is a structural risk that institutional investors are systematically failing to price.
In November 2025, Michael Burry resurfaced publicly to argue that the hyperscalers—Alphabet, Amazon, Meta, Microsoft, and Oracle—are understating depreciation expense by extending the useful lives of artificial intelligence hardware that, in his framing, has an economic life closer to two to three years than the five to six years now commonly assumed.⁵ Buxton Helmsley has previously written on the depreciation question and finds the analytical framework persuasive.⁶ But the more proximate exposure for the artificial intelligence mega-cap cohort—the one most likely to convert a moderate share-price decline into a broader cascade—is hiding in a different proxy disclosure altogether.
Item 403(b) of Regulation S-K requires every public company, in the security ownership table of its annual proxy statement, to indicate by footnote or otherwise the amount of shares beneficially owned by directors, director nominees, and named executive officers that have been pledged as security.⁷ The disclosure obligation is mechanical, but the regulatory environment around it is not. Neither the New York Stock Exchange nor Nasdaq listing standards prohibit insider pledging, and the Securities and Exchange Commission, while it has required pledging disclosure since the 2006 amendments to Item 403, has never moved to limit the practice itself.⁸
The proxy advisors have largely been left to police the issue alone. Institutional Shareholder Services has classified significant pledging as a risk-oversight failure since 2013, and recommends voting against members of the board committee responsible for risk oversight where it identifies excessive pledged amounts.⁹ Glass Lewis evaluates the practice on a case-by-case basis, considering, among other factors, the volatility of the company's shares, the participation of executives in pledging, and the adequacy of pledging policies.¹⁰ BlackRock's stewardship guidelines look to boards to establish policies "prohibiting the use of equity awards in a manner that could disrupt the intended alignment with shareholders' financial interests, such as the excessive pledging or hedging of stock."¹¹
The aggregate effect has been a quiet retreat from the practice. As of May 2022, more than ninety percent of S&P 500 companies had adopted policies restricting insider pledging, with sixty-eight percent imposing a complete ban and twenty-two percent permitting it only with prior approval or in narrowly defined circumstances.¹² The remaining holdouts have tended to be founder-led companies whose largest individual holders predate the modern stewardship regime.
Get the latest insights delivered to your inbox.
The historical record on what happens when pledged stakes are forced to liquidate is unambiguous, and it is worse than most investors remember.
In November 2015, Goldman Sachs liquidated approximately 1.3 million shares of Valeant Pharmaceuticals International stock that the company's Chairman and Chief Executive Officer, Michael Pearson, had pledged to secure a $100 million personal loan. The forced sale prompted Valeant's market value to decline by approximately twenty percent.¹³
In May 2012, Robert Stiller, the founder and then-Chairman of Green Mountain Coffee Roasters, faced a margin call following a sharp share price decline and was forced to sell approximately five million shares—equal to roughly 3.2 percent of common shares outstanding—for an estimated $125.5 million. The sale further depressed the stock, and Stiller was subsequently removed from his role as Chairman.¹⁴
In October 2008, Aubrey McClendon, the co-founder, Chairman, and Chief Executive Officer of Chesapeake Energy Corporation, was required to sell approximately 31 million of his 33,469,359 Chesapeake shares over the course of three trading days to meet margin calls on personal loans collateralized by those shares. Chesapeake's already distressed shares fell approximately forty percent during the period of the involuntary disposition.¹⁵ The same month, Sumner Redstone's National Amusements, Inc.—the holding entity through which he controlled CBS Corporation and Viacom Inc.—disclosed a $233 million sale of CBS and Viacom shares precipitated by covenant violations on a $1.6 billion loan whose terms were tied to the trading prices of those securities.¹⁶
Most consequential of all, Bernard Ebbers, the Chief Executive Officer of WorldCom, accumulated more than $400 million in personal debt collateralized by WorldCom common stock. By the time of his departure in April 2002, every share Ebbers owned had been pledged. The Special Investigative Committee of WorldCom's Board of Directors later concluded that the company's extension of loans and guaranties to Ebbers, ultimately consolidated into a $408 million promissory note at his departure and intended to forestall the involuntary disposition of his shares, "was a 19-month sequence of terrible decisions—badly conceived, and antithetical to shareholder interests—and a major failure of corporate governance."¹⁷
These episodes share a common architecture. An insider whose net worth is concentrated in a single security borrows against that security; the security declines; and the resulting liquidation drives the price lower still, harming all other shareholders. The pledging insider's incentives diverge from those of common stockholders in two distinct phases. Before the margin call, the insider has an incentive to take whatever steps are necessary to avoid the price decline that would trigger forced liquidation; the Ebbers case, in particular, stands as a stark illustration of where that incentive can lead. After the margin call, the lender's interests dominate, and shares are sold without regard to fundamentals, market conditions, or the timing implications for other holders.
Among the companies whose share prices have most directly absorbed the artificial intelligence capital expenditure thesis, the divergence in approach to insider pledging is striking. The disclosure record from the most recent annual proxy filings reveals two distinct cohorts.
The first cohort permits pledging. Oracle, Tesla, Inc., and Meta Platforms, Inc. have each affirmatively approved structures permitting their founder-level holder to pledge company stock. Ellison's 346 million Oracle shares, valued at approximately $107 billion as of September 19, 2025, sit at the top of the list.¹⁸ Elon Musk's 236 million pledged Tesla shares, valued at approximately $100 billion at the time of the company's 2025 proxy disclosure, follow.¹⁹ Mark Zuckerberg's pledged position is materially smaller—approximately 12 million Class B shares, representing roughly 3.5 percent of his beneficial ownership—and is bounded by a board-approved framework that caps the pledged percentage at twenty percent of his ownership and limits the aggregate loan amount to five percent of the fair market value of his beneficial holdings.²⁰
The second cohort prohibits pledging. Microsoft Corporation's corporate governance guidelines and proxy disclosure expressly prohibit directors and executive officers from pledging Microsoft stock as collateral for a loan.²¹ NVIDIA Corporation's insider trading policy, filed as an exhibit to its annual report on Form 10-K, prohibits the pledging of NVIDIA securities as collateral for any loan, without exception.²²
The line drawn between these two cohorts is not, on close reading, principally a matter of governance philosophy. It is a matter of founder concentration. The companies that permit pledging are those whose founder-level holders predate the modern stewardship regime and possess the personal voting power to defeat any policy change put to a stockholder vote. The companies that prohibit pledging are either those without a founder of comparable continuing centrality, or those whose pledging policy was set during a later and more institutionalized governance era. The structural risk to common stockholders is greatest at the firms least equipped to constrain it through ordinary governance channels.
These positions become consequential precisely when investor confidence in the artificial intelligence capital expenditure thesis weakens. Oracle's remaining performance obligations are substantially concentrated in a single counterparty, OpenAI, whose own revenue and user growth recently fell short of internal targets, prompting Oracle's stock to decline approximately four percent on the day the report became public.²³ Oracle's total debt has risen approximately sixty percent year-over-year to a record $153.1 billion, and the company is reportedly raising approximately $50 billion through a mix of debt and equity to fund the data center build-out the OpenAI commitment requires.²⁴ Oracle's shares now trade approximately forty-three percent below their September 2025 peak.²⁵
In supplemental proxy materials issued October 6, 2025, Tesla clarified that Musk currently has no loans collateralized by his pledged shares, but maintains the lending arrangement "to give him the flexibility to take loans that are secured by the pledge."²⁶ Tesla's policy caps the maximum loan amount at the lesser of $3.5 billion or twenty-five percent of the pledged stock value, and any draw is subject to consultation with the company. Even within those constraints, the underlying disclosure permits a multi-billion-dollar lending facility against company stock to be activated at the founder-CEO's discretion, without further board action or stockholder consent.
A continued repricing—driven by any combination of OpenAI's revenue trajectory, Oracle's debt service obligations, the durability of hyperscaler depreciation policies addressed in the prior section, or a broader rotation out of artificial intelligence infrastructure exposure—would test pledging arrangements that institutional investors have simply assumed would never be tested.
The boilerplate disclosure typically offered to address that risk is that the pledging insider "has the financial capacity to repay his personal term loans without resorting to the sale or transfer of pledged shares."²⁷ That assertion, however, is unverifiable from public filings. Common stockholders have no visibility into the insider's other assets, their liquidity, the loan-to-value ratio of the underlying facilities, the price levels at which maintenance covenants would trigger forced sale, or the identities and concentration of the lending counterparties. The board may have access to such information. Common stockholders do not.
The voluntary disclosure regime that has built up around insider pledging is inadequate to the magnitude of the positions now outstanding at the largest companies whose share prices have been most aggressively rerated. Buxton Helmsley believes institutional investors should require, at minimum, the following enhancements before voting in favor of incumbent boards or compensation plans at any company at which a director, executive officer, or significant beneficial owner has pledged a meaningful percentage of his or her holdings.
First, quantitative disclosure of the loan-to-value ratio of the pledging arrangement, the price level at which a margin call would be triggered, and the percentage decline in the company's stock price required to reach that price.
Second, identification of the lending counterparties and disclosure of whether the loans are recourse or non-recourse to assets other than the pledged shares. Non-recourse facilities create the most acute risk of forced sale during a stock price decline because the insider has no economic incentive to fund a margin call from other sources.
Third, a contractual lockup or right of first refusal in favor of the company prior to any involuntary disposition by the lender, so that the company has the opportunity to manage the timing and market impact of any sale rather than ceding that decision entirely to the lender.
Fourth, exclusion of pledged shares from any stock ownership guideline applicable to the insider. Shares that a margin lender can sell at any time are not, in any meaningful economic sense, owned by the insider for purposes of alignment with common stockholders.
Fifth, with respect to companies whose commitments to capital expenditure or counterparty exposure render them particularly sensitive to share price volatility—the artificial intelligence infrastructure cohort being the obvious example—a requirement that any new pledging activity be put to a vote of disinterested stockholders rather than approved unilaterally by the board.
The pledged share is not a problem unique to any one founder, any one company, or any one technology cycle. It is a structural feature of concentrated insider ownership combined with permissive borrowing. What has changed in the current cycle is the size of the underlying positions, the magnitude of the share price moves required to trigger them, and the degree to which the corporate strategies of some of the largest companies in the United States have come to depend on share prices that may not durably support those strategies. The historical record is clear about what happens when those conditions converge. The forensic accounting case for insisting on better disclosure—and, where appropriate, on the unwinding of the underlying positions—is equally clear.
Get the latest insights delivered to your inbox.
[1] Jordan Novet, Oracle Stock on Pace for Worst Quarter Since 2001 Amid AI Concerns, CNBC (December 26, 2025); Wall Street Journal (September 10, 2025) (first reporting the OpenAI–Oracle agreement of approximately $300 billion in cloud computing services over approximately five years).
[2] Id. As of the December 24, 2025 close referenced in Novet, Oracle had fallen approximately forty-three percent from its September 10, 2025 intraday peak of $345.72 (CNBC: "the stock has since lost 43% of its value, closing on Wednesday at $197.49"); the calendar fourth-quarter 2025 decline of approximately thirty percent put the stock on pace for its sharpest quarterly drop since 2001 (CNBC: "Oracle shares have plummeted 30% so far this quarter…on pace for its sharpest decline since 2001 and the dot-com bust"). Margin compression figures reflect analyst consensus polled by FactSet, as reported by CNBC.
[3] Oracle Corporation, Definitive Proxy Statement on Schedule 14A (filed September 26, 2025), Review of Pledging Arrangements; see also Christopher Whittall, Larry Ellison Pledges 30% of His Oracle Shares in Loans, International Financing Review (October 3, 2025).
[4] International Financing Review, supra note 3 (calculating market value at September 19, 2025 closing price).
[5] Michael J. Burry (@michaeljburry), Post on X (November 10, 2025) ("Understating depreciation by extending useful life of assets artificially boosts earnings—one of the more common frauds of the modern era. Massively ramping capex through purchase of Nvidia chips/servers on a 2-3 yr product cycle should not result in the extension of useful lives of compute equipment. Yet this is exactly what all the hyperscalers have done."); see also Hakyung Kim, 'Big Short' Investor Michael Burry Accuses AI Hyperscalers of Artificially Boosting Earnings, CNBC (November 11, 2025).
[6] See Buxton Helmsley, The Useful Life Question, Insights (April 28, 2026).
[7] 17 C.F.R. § 229.403(b); see also Securities and Exchange Commission, Compliance and Disclosure Interpretations, Regulation S-K, Section 129 (Item 403).
[8] See generally New York Stock Exchange Listed Company Manual, Section 303A; Nasdaq Stock Market Rules, Section 5600 series (corporate governance requirements; neither set of listing standards prohibits insider pledging); Securities and Exchange Commission, Final Rule, Executive Compensation and Related Person Disclosure, Release No. 33-8732A (Aug. 29, 2006), 71 Fed. Reg. 53,158 (Sept. 8, 2006) (effective Nov. 7, 2006) (amending Item 403 to require footnote disclosure of pledged shares).
[9] Institutional Shareholder Services, U.S. Proxy Voting Guidelines (2013 update) (treating significant pledging as a risk-oversight failure under the Board Governance Failure policy framework, applied to director elections); see also Davis Polk & Wardwell LLP, The Impact of the ISS Policy Against Pledging (July 1, 2013); Joanna Wu et al., Do Anti-Pledging Policies Have Unintended Consequences for Corporate Governance?, CLS Blue Sky Blog (September 18, 2019).
[10] Glass Lewis & Co., 2022 Policy Guidelines: United States; see also Orrick, Herrington & Sutcliffe LLP, Considerations for Company Insiders When Contemplating Pledging Shares (January 19, 2023).
[11] BlackRock Investment Stewardship, Proxy Voting Guidelines for U.S. Securities (effective January 2025), available at blackrock.com/corporate.
[12] Institutional Shareholder Services, Share Pledges Lose Popularity as Companies Clamp Down, ISS-Corporate Insights (citing ISS Governance QualityScore data as of 2022).
[13] Joanna Wu et al., supra note 9 (citing Financial Times, November 6, 2015, for approximately twenty percent decline in Valeant's market value following the Goldman Sachs liquidation); see also Ronald C. Anderson & Michael Puleo, Insider Share-Pledging and Equity Risk, Journal of Financial Services Research (2020).
[14] Anderson & Puleo, supra note 13; see also Maxwell Murphy, Margin Call: The Most Exposed, The Wall Street Journal CFO Journal (May 15, 2012).
[15] Chesapeake Energy Corporation, Press Release (October 10, 2008), republished in Chesapeake's McClendon Forced to Sell "Substantially All" Shares, Natural Gas Intelligence (October 13, 2008); see also Anna Driver & Brian Grow, Special Report: The Lavish and Leveraged Life of Aubrey McClendon, Reuters (June 7, 2012); William D. Cohan, Aubrey McClendon Attracted Controversy, Outrage, and Lots of Natural Gas, Fortune (March 3, 2016).
[16] Roger Parloff, Viacom Shareholders Feel Redstone's Pain, CNN Money / Fortune (October 24, 2008); see also National Amusements, Inc., Schedule 13D/A filings (October 2008).
[17] Report of Investigation by the Special Investigative Committee of the Board of Directors of WorldCom, Inc. (March 31, 2003) (Dennis R. Beresford, Nicholas deB. Katzenbach & C.B. Rogers, Jr.), §VIII (Corporate Governance Issues); see also United States v. Ebbers, 458 F.3d 110 (2d Cir. 2006).
[18] See sources cited supra notes 3-4.
[19] Tesla, Inc., Definitive Proxy Statement on Schedule 14A (filed September 17, 2025); see also International Financing Review, Third of Musk's Tesla Shares Tied Up in Margin Loans (September 12, 2025); Tesla, Inc., Form DEFA14A (filed October 6, 2025) (disclosing "approximately 236 million shares pledged as collateral (valued at nearly $100 billion)").
[20] Meta Platforms, Inc., Definitive Proxy Statement on Schedule 14A (filed April 16, 2026), Pledging Arrangement.
[21] Microsoft Corporation, Notice of Annual Shareholders Meeting and Proxy Statement (2025) (filed October 21, 2025); Microsoft Corporation, Corporate Governance Guidelines (Derivatives Trading, Hedging, and Pledging Policy).
[22] NVIDIA Corporation, Annual Report on Form 10-K (fiscal year 2026), Exhibit 19.1 (Insider Trading Policy).
[23] Yuvraj Malik, ORCL Stock Drops Premarket: OpenAI Reportedly Misses Targets, Stocktwits (April 28, 2026); see also CNBC, OpenAI Reportedly Missed Revenue Targets; Shares of Oracle and These Chip Stocks Are Falling (April 28, 2026).
[24] Stocktwits, supra note 23 ("In the past year, Oracle's rising debt, data center investment, and customer concentrations have worried investors. The company's total debt increased 60% to a record $153.1 billion last quarter").
[25] Oracle stock price as of early May 2026 (approximately $194–$198 per share) compared with the September 10, 2025 intraday peak of $345.72 per share (a decline of approximately forty-three to forty-four percent). Sources: Yahoo Finance, Investing.com.
[26] Tesla, Inc., Form DEFA14A (filed October 6, 2025) (response to inquiry from proxy advisory firm regarding pledged share disclosure).
[27] Oracle Corporation, Definitive Proxy Statement on Schedule 14A (filed September 26, 2025), Review of Pledging Arrangements.
This article is published by Buxton Helmsley USA, Inc. for informational and educational purposes only. It does not constitute investment advice, a solicitation, or a recommendation to buy or sell any security. The views expressed are those of Buxton Helmsley and are based on publicly available information as of the date of publication. Investors should conduct their own due diligence and consult with qualified professionals before making investment decisions.
Insights and updates delivered directly to your inbox.