Larry Ellison has never been one of the risk-averse leaders, but Oracle ‘s latest financial plan pushes the boundaries even by Silicon Valley standards. The company has announced its intention to raise between $45bn and $50bn in 2026, an unprecedented injection of capital to fund a rapid expansion of its cloud infrastructure. While the goal is clear – to satisfy the hunger for computing power of giants like Nvidia, OpenAI and xAI – the path to achieve it is causing growing concern among bondholders.
Oracle’s strategy is based on a breakneck balance between equity and debt markets. Half of the amount is to come from equity issues and hybrid instruments, including a new $20 billion ‘at-the-market’ equity sale programme. The other half will be financed by new bonds that will hit the market early next year. It is a bold move at a time when the market price of insuring Oracle’s debt against insolvency has reached levels not seen in half a decade.
For executives and investors, the key question is no longer whether Oracle can build data centres, but whether the foundations on which these investments grow are stable. The fate of the Austin-based giant is becoming inextricably linked to the financial health of OpenAI. Sam Altman’s start-up, which is one of the key tenants of Oracle Cloud Infrastructure, remains unprofitable and has not provided a clear funding path for its ambitious plans.
In the background of these events, a legal battle is unfolding that casts a shadow over the company’s transparency. A lawsuit filed by bondholders suggests that Oracle deliberately concealed the scale of debt needed so as not to hound the markets prematurely. For the business, there is a lesson here about the high price of dominance in the AI era: technological superiority today requires not only engineering genius but, above all, nerves of steel in managing the balance sheet. If Ellison’s bets on AI don’t pay off, Oracle could wake up with a future-ready infrastructure that no one will be able to pay for.
