Jensen Huang Exits, Tech Stocks Dodge the "Breakup Fee"

Author: Joanne

Editor: Su Yang

Source: Tencent Tech

After the announcement of Jensen Huang's departure, Nvidia's stock price fluctuated sharply, but the market has not yet seen panic selling.

On November 12, Nvidia announced that Jensen Huang would resign from his roles as CEO and director, with CFO Colette Kress taking over as interim CEO. Following the announcement, Nvidia's stock fell about 14% during intraday trading, before closing down about 4%.

By comparison, the S&P 500 and the Nasdaq Composite fell 0.35% and 0.39% respectively, while the Dow Jones Industrial Average dropped 0.36%. That day, the Nasdaq 100 closed down 0.30% at 28,289.19 points.

The shock of Huang's exit has investors worried about whether it will spread from Nvidia to the broader U.S. stock market's AI supply chain. Over the past few years, Nvidia has not only sold GPUs to customers but also helped some customers expand procurement and build data centers through equity investments, credit support, and computing partnerships. Customers' financing capacity, Nvidia's chip orders, and the expansion of AI infrastructure are thus tightly linked. Whether the new management will adjust these arrangements will determine whether the shock further spreads across the entire supply chain.

Huang's Exit and the AI Trade Priced Separately

On November 12, Nvidia closed down 4.3%, while AMD and Broadcom both rose nearly 6%. The gains by competitors suggest that at least some investors are still looking for opportunities from AI chip demand and reassessing the market share different suppliers might capture.

On the other side, AI infrastructure companies came under greater pressure. CoreWeave fell 6.9%, Nebius dropped 16.8%, and data center-related companies IREN and Applied Digital declined 6.5% and 5.4% respectively. However, Nebius also released quarterly results and guidance that day, and these companies' performance was affected by a combination of their own operating conditions, financing expectations, and the market environment.

These companies' ties to Nvidia go beyond just purchasing chips. Some have received investment or other business support from Nvidia, and their expansion depends on continued access to external funding. Therefore, the market needs to assess not only whether GPU demand remains strong, but also whether customer financing and capital support arrangements can continue after Nvidia's leadership change.

This pressure ultimately fed through to the broader market. That day, all three major U.S. stock indices fell about 0.4%, and the Philadelphia Semiconductor Index dropped 0.5%; the energy sector rose 1.5%, partially offsetting the drag from tech stocks.

Judging from that day's market action, the market has not entirely abandoned the AI trade but is instead distinguishing between different risks: which companies might gain competitive opportunities, which are more dependent on the financing environment, and which have growth tightly linked to Nvidia's support.

The New Nvidia

Over the past few years, building AI data centers has required purchasing large quantities of GPUs, but GPUs are expensive, and power, land, server rooms, and networking infrastructure also require massive investment. For companies specializing in renting out AI computing power, such as CoreWeave, Lambda, and Nscale, whether they can secure sufficient financing directly determines how many chips they can buy and how large a data center they can build.

Nvidia has gradually developed a special industry role: it is both an AI chip supplier and an investor and financing supporter for some of its customers.

Nvidia provides equity investments, credit guarantees, or other financing support to these companies to help them obtain capital; these companies then use the financing to purchase Nvidia GPUs and rent computing power to model companies such as OpenAI and Anthropic.

Ultimately, financial support is converted into chip demand.

Nvidia's investments span the full AI stack: grouped by company type, showing its 50 largest investments since 2023

This is also the source of what the market calls "circular financing" concerns: Nvidia sells chips to customers on the one hand, while helping customers obtain the funds needed to buy chips on the other. The faster customers expand, the more orders Nvidia receives; but if customers cannot generate enough revenue to repay debt, financing risks could in turn flow back to Nvidia.

Nvidia's latest disclosures list financial commitments totaling $164.5 billion, including data center lease guarantees and capacity purchase commitments; in addition, the company has made direct equity investments in AI cloud service providers.

Nvidia has committed up to $164.5 billion to backstop its customers' business difficulties

This model matters all the more because Nvidia's major customers are changing.

Microsoft, Amazon, and Google have their own cloud businesses and massive cash flows, are also developing custom chips, and can support computing expansion through their own balance sheets. By contrast, many emerging cloud service providers do not have cash flows of comparable scale and must rely on external financing to purchase GPUs.

Nvidia's financial reports show that the share of non-hyperscale cloud providers and large enterprise customers in its data center revenue has risen significantly. In the July quarter disclosed by the company, for example, such customers contributed about $40 billion in data center revenue, or about 45%, up notably from the same period a year earlier. They have become an important part of Nvidia's growth.

Therefore, the risk brought by Huang's departure is not that customers will stop buying GPUs tomorrow, but whether Nvidia is still willing to bear the financing and business risks it took on in the past.

Uncertainty Over Continued "Life Support"

Nvidia's financial support for the AI industry already runs deep across multiple projects. But these deals are at different stages, and not all investment intentions should be treated as commitments that have already been finalized.

The most closely watched is OpenAI.

Nvidia previously committed to providing at least $30 billion in financial support to OpenAI, and is also involved in up to $105 billion in lease credit support for an Ohio data center project. Nvidia also plans to participate in Anthropic's IPO with an intended investment of about $10 billion, and has discussed providing credit guarantees for data centers leased by Anthropic.

For Nvidia, these deals are not just financial investments but also an important way to ensure that large model companies continue to have access to computing power and continue to purchase GPUs.

However, model companies like OpenAI and Anthropic do not share exactly the same financing logic as emerging cloud service providers. They have greater computing power needs and more potential sources of funding. Even if Nvidia tightens support for smaller cloud service providers, that does not mean it will immediately withdraw from cooperation with large model companies.

CoreWeave is a different case.

The company is one of the AI cloud service providers Nvidia has backed heavily. The two not only have a chip procurement relationship, but Nvidia has also invested in CoreWeave multiple times and provided credit support for its leasing of Nvidia chips.

CoreWeave's business model is to first raise financing to build data centers and purchase GPUs, then rent out computing power to AI companies. As long as customer demand continues to grow, this model can expand on a rolling basis; but if the financing environment deteriorates or computing rental prices fall, debt pressure will quickly emerge.

After Huang's departure, the market is precisely worried about whether Nvidia will still step in when CoreWeave encounters financing difficulties. Even if Nvidia does not immediately withdraw existing support, approval of new guarantees in the future may become more cautious.

The case of Firmus illustrates the point even more clearly. The Australian AI cloud service provider had planned an IPO but withdrew it due to investor doubts about valuation and business prospects, then turned to seeking at least $2 billion in private financing. Nvidia had previously participated in multiple funding rounds for Firmus and provided credit support for its server purchases. The related arrangements also involved Nvidia providing support in exchange for a share of certain cloud business revenue.

The significance of such deals is that Nvidia provides not just capital but also helps customers cross financing thresholds, allowing GPU procurement and data center construction that would otherwise be difficult to complete independently to continue. But Firmus's IPO setback also exposes a reality: even with Nvidia's backing, capital markets may not be willing to unconditionally pay for every AI infrastructure company.

These cases require distinguishing between two things: investments and guarantees already signed will not automatically become invalid because the CEO leaves; investments, credit support, and cooperation projects still under negotiation are more likely to be affected by the new management's decision-making style.

Kress has long been responsible for Nvidia's financial work and is very familiar with the company's cash flow, risk exposure, and capital allocation. After she takes over as interim CEO, what is most worth watching is not whether she will immediately stop investing, but whether Nvidia will reset its investment thresholds: which customers are worth continuing to support, which projects need more collateral and return guarantees, and which risks Nvidia will no longer bear.

The "Breakup Fee" That Has Not Yet Appeared

The impact of Nvidia's leadership change on U.S. stocks cannot be measured solely by how much market value Nvidia itself has lost.

Nvidia once accounted for more than 6% of the S&P 500's weight and was one of the key companies driving U.S. stocks higher over the past few years. Its stock price fluctuations directly affect index performance, but the S&P and Nasdaq ultimately posted only slight declines on November 12, which also shows that the broader market does not rely entirely on Nvidia alone. Gains by competitors such as AMD and Broadcom and strength in other sectors temporarily offset Nvidia's drag.

If Nvidia tightens credit support, the first to be affected may be emerging cloud service providers and data center operators—rising financing costs would lead them to delay GPU purchases, slow data center construction, and even renegotiate computing leasing contracts. Only then might the impact spread to chip suppliers, server manufacturers, storage companies, and data center power equipment companies.

This also explains why the market showed clear divergence on November 12: companies with close ties to Nvidia's financial support, such as CoreWeave and Nebius, came under pressure, while AMD and Broadcom rose on potential share shifts.

Investors are not turning bearish on AI; they are reassessing who can generate revenue from the next phase of computing investment and who must rely on external funding to sustain expansion.

Another layer of impact comes from valuations.

In the past, the valuations of AI infrastructure companies were largely built on expectations that computing demand would continue to grow in the future. As long as financing channels remained open, companies could build data centers in advance, purchase GPUs, and then repay debt with future computing rental revenue. But when capital markets begin to demand higher risk compensation, project return cycles, customer contract quality, and debt repayment capacity will be re-examined.

This is not specific to Nvidia. On November 12, the U.S. 10-year Treasury yield rose to about 5.07%, and oil prices also climbed amid geopolitical risk. Higher interest rates raise the financing costs of data center projects and also compress the valuations of tech companies with a high proportion of long-dated cash flows. The combination of Nvidia's management change and macroeconomic rate pressure could put AI infrastructure companies' valuations under greater strain.

However, Nvidia's own performance remains a key variable. The company is scheduled to report third-quarter results on November 17. What the market needs to confirm is not only whether GPU orders continue to grow, but also which customers the orders come from, whether those customers have the ability to keep paying, and whether Nvidia needs to continue promoting these orders through investments and credit guarantees.

If the earnings report shows that demand from large cloud providers and enterprise customers is strong enough, Nvidia will have a chance to prove that its growth does not depend entirely on the investment network Huang built. Conversely, if new orders still rely heavily on cloud service providers with weaker financing capacity, investors will pay closer attention to these customers' debt and Nvidia's guarantee obligations.