Nvidia shares fell more than 10% on Thursday after CEO Jensen Huang’s departure, putting the company’s leadership transition in focus just five days before its earnings report. Investors are weighing not only demand for its chips, but also whether the broader AI investment network can hold together without a leader who helped drive Nvidia’s expansion and took part in several industry deal structures. AI-related stocks had already pulled back: Nvidia was down 20% from its November 9 peak, putting it in bear-market territory.
Huang resigned after acknowledging an extramarital relationship with a woman suspected by a US official of being a Chinese spy. Nvidia’s finance chief is serving as interim CEO. Intel also turned to its finance chief as interim CEO in 2018; that year, Intel shares fell 11.7% while the S&P 500 gained 6.8%. The comparison does not mean Nvidia will follow the same path, but the timing of the leadership change has sharpened questions about strategy as AI stocks come under pressure.
AI stocks retreat from recent highs
Several companies across the AI investment chain have fallen further than Nvidia. Data-center operator CoreWeave is down 55.96% from its 2026 high, while Oracle has dropped 47.89%. On November 5, after US midterm election results were announced, the S&P 500 briefly lost about $3 trillion in market value and the Nasdaq 100 fell 6.36%.
The 10 largest US companies account for about 36% to 40% of the S&P 500. That concentration means a pullback in big technology stocks can also affect retirement savings held in widely owned index funds.
$660 billion in spending draws attention to financing
Amazon, Microsoft, Alphabet, Meta and Oracle are expected to spend a combined $660 billion on AI infrastructure this year, 60% more than in 2025. At the same time, financing and procurement arrangements across the sector are becoming increasingly intertwined.
Anthropic’s IPO filing shows data-center commitments of about $895.4 billion through 2045. Broadcom could provide Anthropic with up to $42 billion in financing to pay for computing capacity, much of it involving Broadcom hardware. That puts Broadcom in the roles of both supplier and lender, and could also make it a future shareholder.
Nvidia is similarly central to multiple commercial relationships: it invests in, supplies or buys from companies including OpenAI, Anthropic and CoreWeave. Investor and analyst Nic Puckrin noted that Huang negotiated many of these arrangements and questioned whether his successor could build similar deal structures. Uncertainty about Nvidia’s future direction could also raise questions about the financing arrangements and valuations they support.
Columbia University professor Stijn Van Nieuwerburgh cautioned that vendor financing can help fund construction, but the costs ultimately have to be covered by revenue from end customers. Yields on CoreWeave bonds maturing in 2032 have topped 16%, reflecting investor concern about financing risks at some AI infrastructure companies.
Valuations, rates and leverage send mixed signals
Six of the eight warning signals in the AI bubble gauge maintained by market research account Bull Theory are rated as being in a “late-stage” phase. The account nevertheless puts a potential bubble peak in 2027 or 2028. The S&P 500’s cyclically adjusted price-to-earnings ratio, or CAPE, is near 40, not far below its 2000 peak of 44.2. FINRA data show margin debt used to buy stocks stood at $1.48 trillion in September, below the June record of $1.502 trillion.
Interest rates are also less supportive than they were. Higher-than-expected October inflation pushed the 10-year US Treasury yield back to around 5%. CME FedWatch data show the market-implied probability of a Federal Reserve rate hike in December above 90%.
Still, money has not fully left the AI sector. Anthropic is expected to seek up to $64.4 billion in funding next week at a valuation of as much as $1.4 trillion. AMD shares are up 40% from their November 5 low, while the S&P 500 remains just 2.5% below its record high.
Nvidia earnings and upcoming events test confidence
Further weakness in AI stocks could prompt investors to reconsider allocations across equities, bonds, gold and crypto assets. During the dot-com bust, the Nasdaq fell about 78% from March 2000 to October 2002. Bonds rose as the Federal Reserve cut rates, and gold later began a decade-long advance. Gold is now 25.67% below its January record, while bitcoin is down 16.76% for the year.
Bull Theory argues that if AI stocks retreat and the Fed is forced to cut rates, some capital could move into assets that have already fallen but retain a growth narrative. Gold advocate Peter Schiff, by contrast, sees bitcoin as the more pronounced bubble risk. These views illustrate the market’s disagreement; they do not establish where capital will flow.
Several dates should provide fresh signals. Nvidia is scheduled to report earnings on November 17, Anthropic is expected to go public on November 19, and the Federal Reserve will announce its rate decision on December 9. Nvidia’s demand outlook, the sustainability of AI infrastructure financing and the path of interest rates will all shape investors’ assessment of how resilient the AI boom will be after Huang’s departure.