Nvidia said Jensen Huang resigned as president, chief executive officer and director effective immediately, handing interim leadership to chief financial officer Colette Kress as the company weighs a pipeline of AI investments, credit guarantees and infrastructure commitments exceeding $140 billion over the past three years.
The abrupt change put pressure on Nvidia shares, which fell nearly 7% intraday after the disclosure and had declined more than 25% over the preceding two weeks. CoreWeave and Iren, two cloud-computing companies linked to Nvidia's AI expansion, each fell about 7% on the day.
Kress now takes control during a period when Nvidia's role in AI has extended far beyond selling graphics processors. The company has used equity investments, lease guarantees and financing support to help AI labs and cloud providers secure the capital needed to build data centers packed with Nvidia hardware. Those arrangements can support chip demand, but they also tie Nvidia more closely to customers whose spending depends on debt markets, equity financing and long-term computing contracts.
Kress takes charge of financing decisions
Kress, 59, joined Nvidia in 2013 after holding senior finance positions at Microsoft and a Cisco business unit. She was one of the few Nvidia executives, alongside Huang, regularly visible on the company's quarterly earnings calls.
As interim CEO, she will oversee a financing strategy that has become central to Nvidia's AI ecosystem. The company's commitments have ranged from direct investments in model developers to large guarantees that help cloud operators obtain leases for data-center equipment.
Several commitments appear positioned to continue. Nvidia is expected to proceed with a planned $5 billion investment in Safe Superintelligence, according to the information provided. It is also tied to a proposed investment of roughly $2.5 billion in Thinking Machines Lab, the AI startup founded by former OpenAI chief technology officer Mira Murati. That investment represents about half of Thinking Machines Lab's planned $4 billion to $5 billion financing round.
Other deals are less settled. Nvidia has been discussing a potential $10 billion investment in Anthropic's planned initial public offering after previously providing $105 billion in data-center lease credit guarantees to the AI developer. The scale of those guarantees illustrates how Nvidia has moved from a semiconductor supplier into a financial partner for companies building large AI-computing operations.
Large commitments remain in negotiation
Figure AI, the robotics company in which Nvidia has already invested twice, has been linked to talks over an additional $1 billion investment. No outcome has been disclosed.
Nvidia has also been negotiating more than $5 billion in credit guarantees for Nscale, a cloud provider planning a 460-megawatt data center. The agreement had not been signed by the end of October, leaving the project exposed to a change in Nvidia's deal priorities.
In Australia, cloud provider Firmus is seeking at least $2 billion in private financing after pulling an IPO plan. Nvidia has taken part in discussions, though no final arrangement has been reached. Lambda, another cloud-computing operator, has commitments for nearly $20 billion in lease guarantees from Nvidia and a separate $1.5 billion chip sale-leaseback arrangement. Lambda is pursuing another funding round, but Nvidia's participation remains unconfirmed.
Reflection AI has already received more than $1 billion from Nvidia and has been associated with discussions about further funding. People close to the company said Reflection AI is not currently raising capital.
These financing structures give Nvidia a way to support customers that may otherwise struggle to pay for vast clusters of advanced chips. They also place the company closer to the credit risk and execution risk associated with data-center construction. Competitors AMD and Broadcom have not matched Nvidia's reported use of large credit backstops at the same scale.
Infrastructure platform has yet to complete a transaction
Nvidia's $500 billion AI computing infrastructure financing platform, announced in August with Apollo Global Management, BlackRock, Blackstone, Brookfield Asset Management, Goldman Sachs and KKR, has yet to complete a transaction.
The platform relies in part on Nvidia's residual value support agreement, which is capped at 25% of a transaction's size. That feature is designed to provide lenders with protection against the possibility that AI equipment loses value faster than expected. People familiar with the process said Nvidia may not be able to announce a transaction until January.
Broadcom has used a more aggressive structure in comparable AI infrastructure financing. Its AI XPV platform, developed with Apollo and Blackstone, supported a $35 billion debt transaction in May, including $30 billion in backstop support for Anthropic to lease Google-produced TPU chips. Broadcom later disclosed in an 8-K filing that it would provide roughly $43 billion, or 72.5%, of backstop support for a separate $63 billion transaction.
The contrast puts pressure on Nvidia's new management to decide whether it wants to expand its balance-sheet support or keep tighter limits on financial exposure. A conservative approach could slow projects that depend on Nvidia-backed financing. A more expansive one could deepen the company's exposure to customers building ahead of proven revenue.
Cloud spending could affect decentralized computing tokens
The leadership change has also raised concerns among operators and traders in decentralized computing networks, whose tokens are often treated as a proxy for demand for AI infrastructure. Any slowdown in debt-funded data-center construction could weaken the narrative that rapidly expanding AI workloads will create immediate demand for distributed GPU capacity.
That outcome would not automatically translate into lower activity across every decentralized computing protocol. Networks such as Render and Bittensor have different token models, customer bases and incentives. Yet their markets remain sensitive to the same spending cycle that drives demand for chips, cloud capacity and high-performance computing.
Render had fallen to $1.79 in the figures provided, down 86.78% from its March 2024 record of $13.53. Bittensor's market capitalization was cited at $3.2 billion, while its daily trading activity was described as materially below last year's levels. Those declines show how quickly AI-linked crypto assets can react when expectations for infrastructure spending are reassessed.
Nvidia's next decisions under Kress will therefore be watched well beyond the company's own stock. The pace of its commitments to AI labs, cloud providers and data-center developers could shape how quickly new computing capacity comes online and how much financial support the AI buildout requires to sustain itself.
Navigate Nvidia’s AI shake-up by mastering crypto volatility—start with our guide on funding rates in crypto today.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
