On 12 November, Nvidia announced the departure of founder Jensen Huang, with chief financial officer Colette Kress taking over as interim chief executive officer with immediate effect. Nvidia shares plunged as much as 14.3% during trading before paring losses to close 4.34% lower, wiping about US$182 billion off its market value based on the closing price.
Shares in fellow chipmakers AMD and Broadcom rose on the same day, gaining more than 4% and 5% respectively.
With one giant hit by an unforeseen governance shock and its rivals enjoying a recovery in valuations, the divergent share movements signal a market reassessment: beyond the benefits of a booming sector, an entrepreneur’s personal influence is becoming a hidden risk that artificial intelligence (AI) giants must take seriously.
Retail investors trade on the sector’s growth prospects, and institutional investors scrutinise corporate governance. Nvidia’s latest share price plunge was the result of several overlapping factors, not only the leadership change. These factors included profit-taking in AI stocks, rising oil prices and an approaching earnings release.
Nvidia’s business model relies heavily on Huang’s personal standing in the industry, creating a form of “personal collateral” that is difficult to capture in financial statements.
The chip giant has built an off-balance-sheet guarantee network worth hundreds of billions of US dollars, providing backstops for several computing companies in forms ranging from residual-value guarantees to revenue-sharing arrangements. Industry insiders often jokingly call Nvidia the AI sector’s “central bank of computing power”.
Although these off-balance-sheet guarantees are signed in the company’s name, the market largely ties their fulfilment to the founder’s personal business credibility.
This key-person risk is particularly pronounced in the financing of the computing infrastructure supply chain. Take CoreWeave, a prominent example of a highly leveraged cloud services provider. Its five-year credit default swap (CDS) spread rose sharply during trading on 15 October, at one point reaching 1,194.5 basis points.
This implied a five-year default probability of about 63%, underscoring deep market concerns about the fragility of the underlying financing chain.
In sharp contrast, CDS default protection costs for established chipmakers such as AMD and Broadcom fell markedly on 12 November, reinforcing the signal from their rising share prices.
Higher CDS spreads reflect an increase in the cost to institutions of insuring against debt defaults, and directly indicate a shift in the market’s appetite for exposure to the companies concerned.
The unexpected leadership upheaval closely coincided with an investigation into Megaspeed by Singapore’s police and customs authorities, drawing attention to compliance risks involving Nvidia’s overseas customers.
According to a statement by Singapore’s Ministry of Home Affairs, the authorities began investigating as early as February 2025, focusing on suspected false declarations in server procurement. The investigation is unrelated to espionage.
Over the past two years, Huang has continued lobbying in Washington, urging the US government to ease restrictions on AI chip exports to China and preserve Nvidia’s access to the Chinese market.
Compared with Nvidia’s latent governance risks, the valuation advantages of AMD and Broadcom lie precisely in governance that is less dependent on individuals. Broadcom’s long-term orders are all secured through commercial contracts, leaving its performance unaffected by changes involving a single executive.
AMD’s product development follows an established, publicly available roadmap, rather than changing direction with the departure or arrival of one executive. Its operations do not depend heavily on a key leader’s personal, on-the-spot judgement.
It was Huang’s highly concentrated authority and personal influence that drove Nvidia’s rapid expansion over the past decade. But a governance model so dependent on a central leader may also sow the seeds of future governance problems.
This bout of market turmoil offers a lesson: do not stake the future of a country’s computing industry on any supposedly irreplaceable entrepreneur. The AI industry’s long-term competitiveness must ultimately rest on institutional governance, stable orders and standardised business processes, rather than one entrepreneur’s personal connections and credibility.
When a company’s expected risks are closely tied to the person at its helm, an unexpected upheaval in leadership alone can unsettle capital market expectations.
The author is a financial columnist.