SemiAnalysis Report: The Post-Jensen Huang Era of Nvidia

*Original Title: Nvidia and the World Without Jensen*

*Source: SemiAnalysis*

The core company of the AI era—Nvidia's leader Jensen Huang has quietly stepped down amid scandals, regulatory doubts, and geopolitical games, leaving the global tech community with the biggest question: Is Nvidia still the unique company it once was without its soul figure? As one of the world's top semiconductor and AI research institutions, SemiAnalysis provides an extremely in-depth exclusive analysis based on its profound industry tracking capabilities. The research report not only reveals the underlying issues related to the RASA Act, chip bans, and intelligence community games behind Jensen Huang's resignation but also dissects the systemic crises that "Nvidia without Jensen Huang" will face from dimensions such as organizational structure, supply chain control, financial safety nets, and open-source ecosystems. Below is the original text of the research report:

Jensen Huang is no longer the CEO of Nvidia. The Nvidia board has been investigating Huang's behavior and weighing his future in the company he founded; clearly, behavioral issues do exist—there are ethical concerns regarding relationships with a Chinese client and several subordinates.

Moreover, we believe that his resignation just four days later has deeper reasons behind it. We think this is directly related to large-scale GPU shipments that evade customs regulations. Our accelerator model tracks the production and destination of every AI chip; data shows that over 1.5GW of regulated Nvidia chips ultimately still flowed into China. This figure does not even include the chips specifically designed by Nvidia to circumvent U.S. regulations—this report was initially disclosed by us.

We believe that the U.S. intelligence community is a significant force driving Jensen Huang's departure from Nvidia. Specifically, we think that the RASA (Remote Access Security Act) or similar legislation is likely to continue advancing; RASA has already passed in the House of Representatives. With Huang, who has consistently supported China, stepping down, AI and chip restrictions on China will accelerate. We firmly support certain restrictive measures, but poorly designed restrictions will harm U.S. interests.

Such legislation will have consequences, and we do not believe that politicians have fully considered these consequences. Under the current system where Chinese laboratories can remotely use GPUs, there are clear benefits:

- Profits flow to U.S. companies;

- The U.S. can know what Chinese laboratories are doing;

- If the U.S. wishes, it can almost immediately cut off GPU usage rights through political or coercive means;

- U.S. companies such as Nvidia, AMD, Broadcom, Google, Amazon, Microsoft, Meta, Supermicro, and Dell all rely on Chinese enterprises in their supply chains;

- Nvidia and AMD employ Chinese engineers residing in China to develop hardware and underlying software for their GPU systems. Without these engineers, the user experience of these GPUs would undoubtedly be much worse than it is now;

- U.S. companies, especially the entire startup ecosystem, benefit from the open-source research results published by Chinese enterprises, students, and independent researchers using these GPUs;

- The motivation for China to use domestically produced accelerators will be slightly weaker;

- Innovations from China flow globally through open innovation methods such as extensive expert parallelism and open-source kernel libraries, such as MSA, DeepEP, DeepGEMM, etc.

That said, we do support implementing certain AI restrictions on China, but these restrictions should be closer to the MATCH Act rather than RASA.

Perhaps the most important point is: open-source is under attack. It is an open secret that Nvidia directly collaborates with Chinese laboratories on open-source projects and helps them optimize GPUs and models. These laboratories use Blackwell GPUs, which Nvidia is aware of. The U.S. open-source ecosystem and many startups rely on Chinese AI models. With Jensen Huang's departure and hawks gaining the upper hand, this ecosystem will be severely damaged.

But for shareholders and the AI supply chain, the real question is: What will Nvidia be like without Jensen Huang?

Our view is that this company was never built to operate without him. He has steered the company for 33 years; even if the board forces him out or he "resigns," Nvidia with him will be far better than without him. This company was uniquely built around him. Without this person, the company could become a completely different entity. Without Musk, SpaceX and Tesla cannot operate; without Jensen Huang, Nvidia cannot operate either.

### A King and His 60 Direct Subordinates

For a company of such scale, Nvidia's organizational structure is peculiar. According to Jensen Huang's public statements, he has about 60 direct subordinates. He has stated that a CEO should have the most direct reports because these people require the least management, and he does not meet with them one-on-one.

However, very few can push things forward without Jensen Huang. The company resembles a kingdom with only one king; regardless of seniority, others merely execute his will. Jensen Huang has said that who reports to whom does not matter because titles are unrelated to the actual importance of the work. This is also one reason Nvidia has become "the smallest large company in the world": despite its massive market value, it has only about 42,000 employees. There is little internal politics at Nvidia. Jensen Huang stated that this flat network empowers employees and allows the company to adapt to changes more quickly, as its technology itself iterates rapidly.

This is also one reason Nvidia acts quickly. Once Jensen Huang sets a direction, everyone moves together. Former employees have told us that product development and engineering work often continue until the last moment so that the company can change its mind if necessary. A former executive stated that, for example, the design finalization time for the GB200 was much closer to the product cycle than anyone hoped. Jensen Huang has said he wants the company to always maintain a startup mentality.

However, the cost of this structure is that Nvidia does not have a deep bench of successors. Looking at Nvidia's website: among Huang's direct reports, only five are considered "well-known," and two of them are legal and financial executives. Most of his direct reports have not been trained to handle the public spotlight, and the company explicitly discourages them from seeking exposure. Very few leave Nvidia to become CEOs of other large companies; Rene Haas of Arm is one of the rare exceptions. Few within Nvidia can stand out externally because all major decisions are made by Jensen Huang. But many in Nvidia's executive team do not mind, as the burden of responsibility is heavy for the one wearing the crown; many are happy to work until retirement age because they are well-compensated, treated well, and believe in the company's mission and leadership.

### No One Really Wants the Job

Colette Kress has been appointed as interim CEO, but Nvidia is searching for a permanent CEO. This person cannot have a financial background. Currently, the internal candidates being discussed to potentially succeed Jensen Huang on the technical side are Ian Buck and Jonah Alben.

Buck created CUDA while pursuing his PhD at Stanford University, joined Nvidia in 2004, and currently serves as Vice President of Hyperscale and High-Performance Computing, overseeing the data center business. When Jensen Huang is traveling, he is often tasked with managing related affairs and frequently speaks at meetings.

Alben has been with Nvidia since 1997 and is Senior Vice President of GPU Engineering, responsible for overseeing the design of the company's GPUs.

But the reality is that no one person can manage everything alone. We do not believe Buck has the capability to lead the company's hardware engineering department. Alben and his team have technical autonomy, but even he does not control the entire chip business; other executives responsible for multiple major technical and business departments also report directly to Jensen Huang. Both are outstanding, but neither may be able to demonstrate the same level of supply chain control as Jensen Huang.

Additionally, there is another question: is there anyone willing to take on the role of CEO? One executive told us earlier this year: "Ian would rather die than take over from Jensen Huang; he doesn't want that job. Jonah is more respected, and people say he could do it, but the same applies—no one really wants that job." However, the situation has clearly changed. Kress is the interim CEO, but integrating over 60 vice presidents into a unified company vision requires a special type of person; otherwise, the company's development direction risks getting lost in noise.

### Founder-Like Bets Cannot Be Continued by Professional Managers

Jensen Huang is one of the last old-school tech founders, a contemporary of Elon Musk, slightly older than Mark Zuckerberg. Other companies that grew alongside Nvidia, including AMD, Google, Microsoft, Apple, and Amazon, have completed their transitions to professional management.

Nvidia's success relies on massive bets that may take years or even a decade to pay off. Founders with the full support of the board can take these risks without worrying about internal rebellion. CUDA lost money for years, and Wall Street analysts constantly questioned it, but Jensen Huang persisted, and this bet ultimately paid off handsomely. A former employee stated, "Despite many opposing the investment in limited-use programmable shaders, CUDA, and Tensor Cores, Jensen Huang still bet the entire company on these directions."

Omniverse and autonomous driving are new bets within Nvidia. Former employees say that Jensen Huang internally refers to them as "zero billion-dollar bets." They do not make money today, but he believes their potential is as great as CUDA's. He also built the Shield TV consumer entertainment box and DGX Cloud service based on the idea that "Nvidia needs to understand customers to design better products"; even though these two businesses were nearly unprofitable, they were ultimately abandoned.

A successor to Nvidia who is not a founder, regardless of how long they have worked at the company, will be accountable to the board and shareholders in a different way. Such long-term bets and steadfast beliefs will become less likely. One of the biggest issues is that Jensen Huang's legendary ability to "turn stone into gold" and his deal-making skills will be hard to replicate.

### Customers Buy Jensen Huang, Not Just GPUs

Jensen Huang is the core decision-maker for allocating chips to hyperscale cloud providers and other customers. Customers are always competing for quotas and vying for Nvidia's attention, and an increasing number of Nvidia deals require substantial support from him.

As we wrote in July, Nvidia can choose to allocate more capacity to hyperscale cloud providers, neocloud, or neocloud in its safety net plan. In this plan, Nvidia provides credit support for clusters and shares cloud service revenue. This "Nvidia Central Bank" is supported by Jensen Huang's full credibility and credit. The next CEO is unlikely to make bets as bold as Jensen Huang.

Nvidia prefers to make decisions based on feasibility, but many things ultimately depend on what is strategically significant. Skipping hyperscale cloud providers may not be strategically significant; collaborating with them and giving them preferential treatment may not always be strategically significant either. Nowadays, many customers want assurance: that Nvidia will continue to support them through financial guarantees or quota support.

Nvidia disclosed that as of the second quarter of its fiscal year, its total off-balance-sheet guarantees amounted to $530 billion, up from $184 billion in the previous quarter; this includes a $36 billion take-or-pay minimum commitment for neocloud capacity. Nvidia also holds a $99 billion equity investment portfolio.

Many customers want to strike a deal with Jensen Huang, and the reasons go beyond the numbers themselves. The reason neoclouds around the world can secure financing is that Jensen Huang endorses them, not because they chose NVIDIA after comparing chips. NVIDIA's bottom-line measures allow neoclouds and new labs to secure financing without the support of hyperscale cloud vendors: NVIDIA sells GPUs at full profit margins from day one, then sets a revenue floor for buyers, provides guarantees for buyers' landlords, or signs leases themselves, enabling operators without investment-grade ratings to borrow at investment-grade prices. Without the "Jensen Huang put option," this ecosystem could shrink significantly.

Note: NVIDIA's "bottom-line plan" is essentially a financial support system aimed at leveraging its own credit to facilitate large-scale financing for AI computing infrastructure, with the core mechanism being the "Residual Value Support Agreement": in a typical transaction structure where customers purchase GPUs through special purpose vehicles and lease the computing power, if rent payments are interrupted and there is still a gap after asset disposal, NVIDIA promises to cover the shortfall according to an agreed ratio, thereby reducing lenders' concerns about AI companies defaulting and chip depreciation risks.

### Control over the Supply Chain Will Ease

NVIDIA has strong control over its suppliers. If competitors could secure supplies, their shipments would be much higher. They cannot get them because Jensen Huang personally locked in memory supplies from South Korea and chip supplies from Taiwan.

As we wrote in "The Great AI Silicon Supply Shortage," NVIDIA has locked in most of the supply of logic wafers, memory, and other necessary components. Jensen Huang's trip to South Korea before 2025 was to ensure supply before other executives fully realized the extent of the memory shortage. Our industry clients once thought our 2026 forecast was overly aggressive, but later they found we were right. NVIDIA's supply and capacity commitments jumped from $119 billion to $279 billion in one quarter, primarily for memory.

This control has a strong personal touch. Jensen Huang made these deals with major supply chain companies in a CEO-to-CEO manner. We estimate that in South Korea, SK Hynix's price for HBM4 to NVIDIA in 2027 could be around $3.0 to $3.3 per Gb; while other tier-one customers, including AMD, Broadcom, Amazon, and Google, are around $3.7 to $4.1 per Gb. We believe there is an understanding between the two parties: in exchange, NVIDIA grants SK Group priority quotas for GPUs. Without Jensen Huang, these companies would not prioritize NVIDIA to the same extent, and other companies would find it easier to secure capacity. We believe that by 2028, over 70% of total capacity may not belong to NVIDIA.

More importantly, in addition to locking in supply, Jensen Huang's vision has also driven the Taiwanese supply chain to increase investment, making it more confident in this round of construction than anyone else. Jensen Huang's legendary supply chain dinners and commitments have pushed Taiwanese suppliers to execute aggressive plans at both technical and logistical levels.

### The AI Industry Loses Its Open Source Advocate

Without Jensen Huang, the AI industry loses its most important advocate. He is everywhere, always speaking as the chief advocate for AI. He is the face of AI, and no one can replace that.

Jensen Huang also formulated NVIDIA's China policy and pushed for its support of open-source AI models. These models give countries, enterprises, and neoclouds reasons to purchase GPUs, rather than just relying on a few cutting-edge labs developing their own chips.

As we wrote in September, no one needs open-source AI more than NVIDIA, which is heading toward a world with "only Anthropic and OpenAI as real customers, and if you count hyperscale cloud vendors, there are a total of seven customers." Among these seven customers, except for SpaceX, the rest are trying to build their own XPUs to erode NVIDIA's margins. AI models and applications are complementary to GPUs.

Cutting-edge labs in China use Blackwell GPUs, and NVIDIA DevTech closely collaborates with these labs, learning from their end users, which is already an open secret. This has facilitated innovations in the CUDA ecosystem, such as widespread expert parallelism, DeepGEMM, and DeepEP. Subsequently, these innovations spread to the Western open-source community and startups. Jensen Huang's departure would have a massive negative impact on these startups.

Without Jensen Huang as an advocate and providing financial and engineering support for open models, Chinese labs may no longer be able to open-source kernel libraries for NVIDIA Blackwell GPUs, such as MiniMax's MSA, DeepGEMM, DeepEP, and FlashKDA from the Dark Side of the Moon. Widely used inference engines like SGLang and vLLM have adopted these innovations from Chinese labs. Chinese labs may have to completely deny having NVIDIA GPUs and limit NVIDIA software to internal use, rather than open-sourcing Blackwell software.

Innovations from China will no longer be shared within the NVIDIA ecosystem and the entire machine learning ecosystem, which is fundamentally a bad thing. This may be beneficial for closed AI labs; however, it is detrimental to startups.

A company led by managers may deprioritize the underperforming Nemotron. As many models from U.S. cutting-edge labs shift to closed-source, NVIDIA has been single-handedly supporting other U.S. open-source models by investing in companies like Reflection, Thinking Machines, Poolside, and acquiring Hugging Face.

NVIDIA has been pushing for continued engagement with China, arguing that China should use and rely on U.S. technology. As we wrote last year, after the U.S. government banned H20 from entering China and then indicated it would grant NVIDIA an export license, the U.S. needs to balance allowing China to use the U.S. AI technology stack, slowing its autonomous development, and limiting the quality and quantity of products exported to China. Now, what if he is pushing all this under some constrained state? That will be determined by investigations. Clearly, the hawks are winning, so we must pose a somewhat strange question: why now?

### Why Are Jensen Huang's Personal Issues Being Exposed Now?

We are not keen on conspiracy theories, but several events occurring around the same time seem to form a narrative that is hard to ignore. The intelligence community has benefited greatly from this. The second report published by The New York Times on this topic clearly states that its reporting is based on U.S. intelligence. Trump, Rubio, and Luttig were already aware of all this before their visit to China in May. Additionally, there are reports that Luttig had delayed actions against Chinese chips. This relationship has changed, and now the winds are blowing in another direction.

We do not actually believe Alice Huang has any special identity—she is just a well-connected business person. Like other well-connected business people in China, it is normal for her to have ties to the authorities.

Our accelerator model tracks over 1.5GW of NVIDIA GPUs that should not exist in China. The U.S. Department of Commerce now also believes that there are over 1.5GW of NVIDIA chips that should not be there in China. With the midterm elections leaning toward the Democrats, someone decided it was time to take a hard line: to stop actively circumventing existing export controls and to expand the scope of restrictions, completely cutting off China's access to cutting-edge chips.

It is important to clarify that there were divisions within the Trump administration on this issue, but Trump had previously provided cover for Luttig. Trump is no longer providing that cover. More importantly, a key bill called the Remote Access Security Act (RASA) has passed the House and is yet to be voted on in the Senate, and that vote is happening.

### RASA is About to Be Voted On

RASA passed the House on January 12 with 369 votes in favor and 22 against. It passed via a "suspension of rules motion," which requires a two-thirds majority. The House Foreign Affairs Committee approved the bill with a vote of 51 to 0 in April 2025. Although the bill has been submitted to the Senate Banking Committee, it has not yet been voted on. We know that NVIDIA and Oracle have been lobbying strongly against it.

Of course, RASA is not the only way the U.S. government can prevent Chinese companies from remotely accessing GPUs. But RASA is currently the most hotly debated bill in Washington. It is clearly popular, as it has already passed the House; and the substantive content of the bill seems obvious to many: why allow Chinese companies to remotely access GPUs they are not permitted to import?

It is worth mentioning that we know that neoclouds in Asia, Australia, the Middle East, Northern Europe, Europe, Canada, and even the U.S. are actively selling computing power to Chinese companies. They will all be affected by this bill. This is another global export control bill with far-reaching implications. Jensen Huang, and possibly Luttig, do not want to see this happen.

Therefore, let us oppose the popular narrative and explore a possibility: despite the recent headlines, passing RASA or similar bills could lead to unexpected consequences. According to an old saying in American politics, we believe that if all politicians agree on something, it is probably a bad idea.

As mentioned earlier, under the current system, Chinese labs can remotely use GPUs, which brings the following obvious benefits:

- Under the current system where Chinese labs can remotely access GPUs, profits flow to U.S. companies;

- The U.S. can monitor what Chinese labs are doing;

- The U.S. can almost cut off GPU access at any time if it wishes;

- U.S. tech companies rely on Chinese enterprises and engineers;

- U.S. companies benefit from Chinese open-source research;

- China will also use domestic accelerators less;

- Chinese innovation flows globally through open-source kernels like MSA, DeepEP, DeepGEMM, and widespread expert parallelism mechanisms.

It is important to clarify that in a world where Chinese labs are completely cut off from using NVIDIA GPUs, their motivation to turn to domestic accelerators and move away from NVIDIA GPUs will only grow stronger.

And this is already happening.

In February, Zhiyu AI claimed that GLM-5 has supported inference on seven domestic accelerators.

In April, DeepSeek released the V4 paper, and then open-sourced an inference super kernel for the V4 Pro model on Huawei's Ascend NPU.

In June, Meituan released LongCat 2.0 and claimed that the model was fully pre-trained on Huawei's Ascend 910B NPU.

Politicians continue to ignore how tightly intertwined China, Taiwan, and the U.S. actually are. Many Chinese nationals work in cutting-edge labs; many Chinese engineers are employed by NVIDIA and AMD, writing the underlying code to launch new chips; many Americans learn from Chinese open-source contributions; many of the raw materials needed to produce these chips come from China; many ODM board and component suppliers from L1 to L6 are also Chinese companies.

Escalating tensions will inevitably lead to reduced chip production and slow down AI progress for everyone.

We believe the U.S. should limit or control the scale of computing power delivered to Chinese labs; prohibiting exports to China while allowing trusted cloud service providers to offer remote access is the best way to achieve this goal.

### The Situation May Be Worse Than We Imagine

Therefore, if one must answer the question "Why now," one possibility is that this move is to eliminate the opposing forces faced by an extremely tough export control bill against China. With significant changes occurring in Congress and the impending lame-duck status, it is most beneficial to clear out the opposition to push the bill through. This is our most conspiracy-theory-laden judgment of the current situation and aligns with various ripples coming out of Washington.

Now, back to the initial question.

### What would Nvidia be like without Jensen Huang?

At such a scale, there is only one true case of a company that can continue to operate well after losing its legendary founder: Apple and Steve Jobs. We believe that Jensen Huang's departure clearly signifies the end of the era of visionary founders; it is impossible to operate the company with the same intensity if the person who built it is not at the helm.

To use a car analogy that is easy to understand: no one wants to drive a car built by someone else into the ground; and the car's manufacturer—Jensen Huang—can comfortably drive at 150 miles per hour on the highway.

Whoever is promoted to maintain the status quo and shift the company from founder-led to manager-led may need to deal with an enviable problem: how to handle such a large amount of cash. Jensen Huang clearly envisioned making the company the balance sheet backstop for the ecosystem, providing and expanding credit for neocloud and laboratories. This is a natural extension of supply chain dominance.

But professional managers are likely to do another thing that is somewhat effective: stock buybacks. Apple after Jobs is the best example of a company that performed well after the founder's handover. We believe that the only script that has worked at such a scale could very well play out again: the founder hands over to professional managers, the company becomes more conservative, and cash is returned. Shareholders will be happy, but the company's pace of technological advancement will no longer be what it once was.

This is a bad outcome for technology. It will also be a bad outcome for shareholders. But clearly, people will miss Jensen Huang, and the unique Nvidia will become more ordinary. As the supply chain shifts from following Jensen Huang's bold commitments to following the mediocre vision of managers, the internal struggles and digestion processes within this giant will manifest both internally and globally. Without him, the tech world will be different.

Our current judgment is that Ian Buck is most likely to lead, with Kress closely following, potentially becoming the formal CEO. There are also rumors about external CEO candidates, such as Johny Srouji. However, given Nvidia's uniqueness, we believe it is best to promote from within the company. There are few candidates, and whether anyone is willing to take the job remains to be seen. Only industry giants have the hope of being as outstanding as Jensen Huang; we doubt that giants like Lisa (AMD's CEO Lisa Su) or Hock (the recently retired former Apple CEO) would switch jobs.

In any case, this is very beneficial for all of Nvidia's competitors and very detrimental to Nvidia's supply chain.