SemiAnalysis Report: Nvidia's Post-Jensen Huang Era

Original title: "Nvidia and the World Without Jensen"

Original source: SemiAnalysis

The helmsman of Nvidia, the core company of the AI era, Jensen Huang, has departed in disgrace amid scandal, regulatory uncertainty, and geopolitical maneuvering, leaving the global tech world with one overriding question: without its soul, is Nvidia still that one-of-a-kind company? As one of the world's leading semiconductor and AI research firms, SemiAnalysis has leveraged its deep industry-tracking capabilities to deliver an exceptionally in-depth exclusive analysis. The report not only uncovers the hidden threads behind Jensen Huang's resignation involving the RASA Act, chip bans, and intelligence-community maneuvering, but also dissects, across dimensions including organizational structure, supply chain control, financial backstop systems, and the open-source ecosystem, what systemic-level crisis "Nvidia without Jensen Huang" will face. Below is the original report:

Jensen Huang is no longer Nvidia's CEO. Nvidia's board has been investigating Jensen Huang's conduct and weighing his future at the company he founded; clearly, the conduct issues are real—relationships with a Chinese customer and multiple subordinates raise ethical problems.

Moreover, we believe there are deeper reasons behind his resignation just four days later. We believe this is directly tied to large-scale GPU shipments that evaded customs oversight. Our accelerator model tracks the production and destination of every AI chip; the data shows that over 1.5GW of controlled Nvidia chips ultimately flowed into China. This figure does not even include the chips Nvidia specifically designed to circumvent U.S. regulations—a story we first disclosed.

We believe the U.S. intelligence community was a major force behind Jensen Huang's departure from Nvidia. Specifically, we believe RASA (the Remote Access Security Act) or similar legislation will likely continue to advance; RASA has already passed the House. With the departure of Jensen Huang, who had consistently supported China, restrictions on AI and chips toward China will accelerate and escalate. We firmly support certain restrictions, but poorly designed restrictions harm U.S. interests.

Legislation of this kind has consequences, and we do not believe politicians have fully considered them. Under the current system in which Chinese labs can remotely use GPUs, there are clear benefits:

· Profits flow to U.S. companies;

· The U.S. can know what Chinese labs are doing;

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

· U.S. companies including Nvidia, AMD, Broadcom, Google, Amazon, Microsoft, Meta, Supermicro, and Dell all depend on Chinese companies in their supply chains;

· Nvidia and AMD employ Chinese engineers living in China to develop hardware and low-level software for their GPU systems. Without these engineers, the user experience of these GPUs would undoubtedly be far worse than it is today;

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

· China has slightly less incentive to use domestic accelerators;

· Innovation from China flows to the world through open innovation approaches such as widespread expert parallelism and open-source kernel libraries, for example MSA, DeepEP, DeepGEMM, and others.

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

Perhaps most importantly: open source is under attack. It is an open secret that Nvidia works directly with Chinese labs on open-source projects and helps them optimize GPUs and models. These labs use Blackwell GPUs, and Nvidia knows it. The U.S. open-source ecosystem and many startups depend on Chinese AI models. With Jensen Huang gone and hawks gaining the upper hand, this ecosystem will be severely damaged.

But the question that really matters for shareholders and the AI supply chain is: what will Nvidia look like without Jensen Huang?

Our view is that this company was never built to operate without him. He has been at the helm for 33 years; even if the board forced him out, or he "resigned," Nvidia is far better off with him than without him. The company is uniquely constructed around him as an individual. Without this person, the company could become something entirely different. SpaceX and Tesla cannot function without Musk; Nvidia cannot function without Jensen Huang.

A King and His 60 Direct Reports

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

But few of them can move things forward without Jensen Huang. The company is like a kingdom with only one king; regardless of seniority, everyone else merely executes his will. Jensen Huang has said that it doesn't matter who reports to whom, because titles are unrelated to the actual importance of the work. This is one reason Nvidia has become "the smallest big company in the world": despite its enormous market capitalization, it has only about 42,000 employees. There is relatively little internal politics at Nvidia. Jensen Huang says this flat network empowers employees and allows the company to adapt quickly to change, because its technology itself iterates extremely fast.

This is also one reason Nvidia moves quickly. Once Jensen Huang sets a direction, everyone pushes forward together. Former employees told us that product development and engineering work often continue until the last moment so the company can change its mind if necessary. A former executive said, for example, that the GB200's design was finalized far closer to the product cycle than anyone would have liked. Jensen Huang has said he wants the company to remain a startup forever.

However, the cost of this structure is that Nvidia has no deep bench of successors. Look at Nvidia's website: among Jensen Huang's direct reports, only five are even somewhat "known," and two of them are legal and finance executives. Most of his direct reports have not been trained to handle the public spotlight, and the company explicitly discourages them from seeking exposure. Few leave Nvidia to become CEOs of other large companies; Arm's Rene Haas is a rare exception. Few people inside Nvidia can shine externally, because all major decisions are made by Jensen Huang. But many on Nvidia's executive team don't mind, because heavy is the head that wears the crown; many are happy to work until retirement age because they are well compensated, well treated, and believe in the company's mission and its leader.

No One Really Wants the Job

Colette Kress has been appointed interim CEO, but Nvidia is searching for a permanent CEO. This person cannot have a finance background. The candidates currently discussed internally as possible technical successors to Jensen Huang are Ian Buck and Jonah Alben.

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

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

But the reality is that no single person can manage everything alone. We do not believe Buck has the ability to lead the company's technical hardware engineering organization. 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 units also report directly to Jensen Huang. Both are outstanding, but neither may be able to demonstrate Jensen Huang's command of the supply chain.

There is also the question of whether anyone wants to be CEO. An executive told us earlier this year: "Ian would rather die than succeed Jensen Huang; he doesn't want that job. Jonah is more respected, and people say he could do it, but it's the same thing—no one really wants that job." Clearly, however, the situation has changed. Kress is interim CEO, but integrating more than 60 vice presidents into a unified company vision requires a certain type of person; otherwise, the company's direction risks getting lost in the noise.

Founder-Style Bets Cannot Be Continued by Professional Managers

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

Nvidia's success depends on huge bets that may take years or even a decade to pay off. A founder with full board support can take these risks without fear of internal rebellion. CUDA lost money for years, and Wall Street analysts constantly questioned it, but Jensen Huang persevered, and the bet ultimately paid off enormously. A former employee said: "Despite many people opposing spending transistors on programmable shaders with limited uses, CUDA, and Tensor Cores, Jensen Huang bet the entire company on these directions."

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

A non-founder successor at Nvidia, no matter how long they have been at the company, will be accountable to the board and shareholders in a different way. Such long-term bets and firm convictions will become less likely. One of the biggest problems is that Jensen Huang's legendary "Midas touch" and dealmaking skills will be very difficult to replicate.

Customers Buy Jensen Huang, Not Just GPUs

Jensen Huang is the core decision-maker on allocating chips to hyperscalers and other customers. Customers are always competing for allocation and for Nvidia's attention, and a growing number of Nvidia deals require substantial support from the company.

As we wrote in July, Nvidia can choose to allocate more capacity to hyperscalers, neoclouds, or the neoclouds in its backstop program. In that program, Nvidia provides credit support for clusters and shares cloud service revenue. This "Nvidia central bank" is backed by Jensen Huang's full faith and credit. The next CEO is unlikely to make bets as bold as Jensen Huang's.

Nvidia likes to make decisions based on feasibility, but many things ultimately come down to what is strategically meaningful. Skipping hyperscalers is not necessarily strategically meaningful; working with them and giving them preferential treatment is also not always strategically meaningful. Today, many customers want assurances that Nvidia will continue to support them through financial guarantees or allocation.

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

Many customers want to do deals with Jensen Huang for reasons that go beyond the numbers themselves. Neoclouds around the world are able to raise financing because Jensen Huang backs them, not because they compared chips and chose Nvidia. Nvidia's backstops allow neoclouds and new labs to raise financing without hyperscaler support: Nvidia sells GPUs at full margin on day one, then sets a floor on the buyer's revenue, guarantees the buyer's landlord, or signs leases itself, enabling operators without investment-grade ratings to borrow at investment-grade prices. Without the "Jensen Huang put option," this ecosystem could shrink dramatically.

BlockBeats note: Nvidia's "backstop program" is essentially a financial support system that uses its own credit as a fulcrum to lever up large-scale financing for AI compute infrastructure. At its core is a mechanism called the Residual Value Support Agreement: in a typical transaction structure where customers borrow through special purpose vehicles to buy GPUs and rent out compute, once rent payments are interrupted and a shortfall remains after asset disposal, Nvidia commits to covering the difference at an agreed ratio, thereby reducing lenders' concerns about AI company defaults and chip depreciation risk.

The Chokehold on the Supply Chain Will Loosen

Nvidia has strong control over its suppliers. If competitors could obtain supply, their shipments would be much higher. They can't get it because Jensen Huang personally locked up memory supply in Korea and chip supply in Taiwan.

As we wrote in "The Great AI Silicon Shortage," Nvidia has locked up most of the supply of logic wafers, memory, and other necessary components. Jensen Huang traveled to Korea in 2025 to secure supply before other executives fully realized the extent of the memory shortage. Our industry clients once thought our 2026 forecasts were too aggressive, but later found we were right. Nvidia's supply and capacity commitments jumped from $119 billion to $279 billion in a single quarter, mainly for memory.

This control is highly personal. Jensen Huang struck these deals with major global supply chain companies CEO-to-CEO. We estimate that in Korea, SK Hynix's HBM4 price to Nvidia in 2027 may be around $3.0 to $3.3 per Gb; while other tier-one customers including AMD, Broadcom, Amazon, and Google pay around $3.7 to $4.1 per Gb. We believe there is an implicit understanding: in exchange, Nvidia gives SK Group priority allocation of GPUs. Without Jensen Huang, these companies will no longer prioritize Nvidia to the same degree, and other companies will find it easier to obtain capacity. We believe that by 2028, more than 70% of total capacity may no longer belong to Nvidia.

More importantly, beyond locking up supply, Jensen Huang's vision also drove Taiwan's supply chain to invest more heavily, making it believe in this buildout more than anyone else. Jensen Huang's legendary supply chain dinners and commitments drove Taiwanese suppliers to execute aggressive plans at the technical and logistical levels.

The AI Industry Loses Its Open-Source Champion

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

Jensen Huang also shaped Nvidia's China policy and pushed it to support open-source AI models. These models give countries, companies, and neoclouds a reason to buy GPUs, rather than relying only on a handful of frontier 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 two real customers, Anthropic and OpenAI, or seven if you count the hyperscalers." And of those seven customers, all except SpaceX are trying to build their own XPUs to erode Nvidia's margins. AI models and applications are complements to GPUs.

It is an open secret that Chinese frontier labs use Blackwell GPUs, and that Nvidia's DevTech works closely with Chinese frontier labs, learning from their end users. This has driven innovation in the CUDA ecosystem, such as widespread expert parallelism, DeepGEMM, and DeepEP. These innovations then spread to the Western open-source community and startups. Jensen Huang's departure will have a massively negative impact on these startups.

Without Jensen Huang as an advocate and providing financial and engineering help 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 Moonshot's FlashKDA. Widely used inference engines globally, such as SGLang and vLLM, have adopted these innovations from Chinese labs. Chinese labs may have to completely deny that they have Nvidia GPUs and restrict Nvidia software to internal use, rather than open-sourcing Blackwell software.

Innovation from China no longer being shared within the Nvidia ecosystem and the broader machine learning ecosystem is fundamentally a bad thing. For closed AI labs, it may be beneficial; but it is bad for startups.

A manager-led company may deprioritize the underperforming Nemotron. As many models from U.S. frontier 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, and Poolside, and acquiring Hugging Face.

Nvidia has consistently pushed for continued engagement with China on the grounds that China should use and depend on American technology. As we wrote last year, after the U.S. government banned H20 from entering China and then said it would grant export licenses to Nvidia, the U.S. needed to strike a balance between letting China use the American AI stack, slowing its independent development, and limiting the quality and quantity of products shipped to China. Now, what if he was pushing all of this while being held in check? That is for the investigation to determine. Clearly, the hawks are winning, so we must ask a somewhat strange question: why now?

Why Is Jensen Huang's Private Life Being Exposed Now?

We are not fond of conspiracy theories, but several things happening at the same time seem to form a narrative that is hard to ignore. The intelligence community benefits greatly from it. The New York Times' second report on this topic explicitly stated that its reporting was based on the U.S. intelligence community. Trump, Rubio, and Lutnick already knew all of this before their visit to China in May. In addition, there are reports that Lutnick delayed a crackdown on Chinese chips. The relationship has changed, and now the wind is blowing the other way.

We actually do not believe Alice Huang really has any special identity—she is simply a well-connected businessperson. Like other well-connected businesspeople in China, it is normal for her to have official contacts.

Our accelerator model tracked more than 1.5GW of Nvidia GPUs that should not exist in China. The U.S. Commerce Department now also finally believes that more than 1.5GW of Nvidia chips exist in China that should not be there. As the midterm elections tilt toward the Democrats, someone decided it was time for tough action: both to stop active evasion of existing export controls and to expand restrictions to completely cut off China's access to frontier chips.

To be clear, the Trump administration is divided on this issue