Reports of alleged inappropriate relationships and potential geopolitical conflicts of interest involving NVIDIA (NVDA) chief Jensen Huang have sent shockwaves through the technology world.
NVIDIA (NVDA), once the most valuable company on earth, has lost eight hundred billion dollars or more in market capitalization over the past several weeks, depending on the measurement window and whether market concerns about AI capital spending are included.
Jensen Huang is the co-founder and helmsman of NVIDIA (NVDA) for more than three decades, widely loved, deeply respected, and hailed as a visionary icon; he has now stepped down as CEO under board pressure, though some stakeholders are still pushing for his reinstatement.
He possesses a unique ability: to anticipate technological shifts, orchestrate complex product portfolios, and build an unrivaled partner ecosystem.
Unfortunately, he is also alleged to have been involved in multiple improper relationships that violated company policy and may carry national security implications.
The question on everyone's mind is: 1) What are the prospects for NVIDIA (NVDA) without Huang? 2) What does this mean for the AI industry? 3) What does this mean for the US-China tech rivalry?
Ultimately, the broader challenge is: how to separate a great founder from a great company; how to understand the interplay between human capital (such as the CEO and engineering talent), organizational capital (such as culture, processes, and capabilities), and ecosystem capital (such as developer loyalty); and how to distinguish fiduciary duty, personal legacy, and national interest.
NVIDIA Without Huang
Huang has led NVIDIA (NVDA) for 33 years, taking the company from a startup, through near-bankruptcy lows, to the pinnacle of the world's most valuable company, and now into turbulence.
Huang is almost synonymous with NVIDIA (NVDA). What happens next?
Sometimes, when a company loses its way or makes strategic missteps, "losing" a key figure can actually be beneficial.
In 2024, Laxman Narasimhan was ousted as Starbucks CEO after just 17 months amid declining sales; Starbucks shares immediately surged 24% because the board brought in Chipotle CEO Brian Niccol, a widely recognized fast-food leader.
Other times, losing a founder, strategic architect, or visionary triggers severe turmoil: after Adam Neumann was pushed out, WeWork's valuation plummeted 83%; when Sam Altman was (briefly) ousted as OpenAI CEO, the company's investment was imperiled and 95% of employees threatened to resign; McDonald's CEO Steve Easterbrook, the architect of its digital strategy, was fired for a consensual intimate relationship with an employee, and McDonald's lost 4 billion dollars in market value in a single day.
Sometimes, both scenarios occur in sequence: after Steve Jobs was ousted from Apple in 1985, Apple initially grew rapidly — John Sculley brought financial and operational discipline, led the desktop publishing revolution, and launched the PowerBook laptop along with the Macintosh SE and Macintosh II computers.
Yet by the early 1990s, Sculley and his successors made a series of strategic missteps that brought the company to the brink of bankruptcy.
Jobs returned to the Apple he co-founded, drastically cut product lines, secured a lifesaving investment from Microsoft, and not only revived the company but propelled it to the position of the world's most valuable company.
In fact, Jobs and Huang share many similarities.
Both are charismatic leaders, both are co-founders of iconic technology companies, and both have taken their respective companies to the pinnacle of global market value.
During their respective tenures, they were each deeply and inseparably bound to their companies — just as Gates was to Microsoft, Ma Yun to Alibaba, Zuckerberg to Facebook, and Musk to SpaceX.
This leads to what management scholars call "The Romance of Leadership": the tendency to over-attribute organizational success or failure to the actions and character of a single leader, rather than understanding the truly difficult-to-comprehend variables — the complex collaboration among tens of thousands of employees who never appear on stage at a launch event, turbulent market dynamics, and sheer luck.
Its psychological basis is "bounded rationality" — the ceiling on how much information the human brain can process — and a cognitive bias pervasive in behavioral economics: the "fundamental attribution error," which is the irrational tendency to explain behavior by personality traits rather than situational factors.
For example, if someone cuts you off in traffic, you are likely to conclude they are rude, without knowing that perhaps a squirrel suddenly darted in front of their car.
Yet even legendary founders are not always right.
The mercurial Jobs was ousted from Apple against the backdrop of Macintosh sales nearly collapsing tenfold — mainly due to insufficient performance, derided as a "toy computer"; because he let the Apple II division and the Macintosh division fall into internal strife and cultural dysfunction; and because when the board asked him to cede operational control, he attempted a "boardroom coup" to fire his own successor, John Sculley.
In the mid-1990s, Huang bet the future of the then-startup NVIDIA (NVDA) on quadratic texture mapping — a mathematically elegant approach to rendering 3D curved surfaces.
Unfortunately, Microsoft made triangle-based graphics processing the de facto standard.
So by 1996, NVIDIA (NVDA) graphics card inventory was almost completely unsellable, and the company was 30 days from bankruptcy.
NVIDIA (NVDA) exists today because Sega president Shoichiro Irimajiri was willing to release the company from its contract and provide 5 million dollars in lifesaving funding.
More recently, NVIDIA's (NVDA) most advanced GPU/HBM (graphics processing unit/high bandwidth memory) architecture proved indispensable for model training (the "learning" phase of AI), yet was not optimal for low-latency, low-cost inference (the "thinking" phase of AI).
This strategic architectural misjudgment forced NVIDIA (NVDA) to pivot: a 20 billion dollar non-exclusive technology licensing and "reverse-acquihire" arrangement with Groq, which in turn triggered an ongoing US Department of Justice antitrust investigation.
Discussing these matters is not intended to disparage Huang's (or Steve Jobs') astonishing achievements and abilities, nor to deny the iconic organizations they built.
The point is: leaders are not infallible — they are human.
And organizations are ultimately organizations: composed of talented individuals who often collaborate and often compete; embedded in complex industry networks of customers, rivals, partners, regulators, and geopolitics; subject to capricious decisions and random events; yet at the same time protected by intellectual property, customer and developer loyalty, supply chain contracts, cash reserves, and partnership agreements.
Therefore, while no one can predict how the AI and semiconductor markets will evolve, the likelihood that NVIDIA (NVDA) will achieve a major breakthrough — or suffer a major disaster — solely because Huang leaves is extremely low; the same holds true even if he returns.
Implications for the AI Industry
Although NVIDIA's (NVDA) market value has been battered in recent weeks following reports of these alleged improprieties, a company's market valuation can at best only indirectly reflect its true strategic position, customer purchasing decisions, or long-term viability.
A higher stock price does bring greater financial flexibility — for example, serving as "currency" in all-stock acquisitions — but NVIDIA's (NVDA) market position rests on far more solid foundations.
First, NVIDIA (NVDA) builds the most powerful, general-purpose, and scalable architecture on the market.
Frontier AI labs are locked in a highly adversarial race to build stronger models; to stay competitive, they have little choice but to graft their proprietary algorithms onto the fastest available hardware.
Custom chips may be faster for the specific tasks they are designed for, but NVIDIA's (NVDA) latest architecture, "Vera Rubin," continues its track record of performing best on the widely recognized MLPerf benchmark suite.
Second, many customers are structurally bound to NVIDIA (NVDA) through capital investments.
NVIDIA (NVDA) has invested billions of dollars in frontier labs and the "neoclouds" that support them, including 30 billion dollars in OpenAI, up to 10 billion dollars in Anthropic, 21 billion dollars tied to SpaceX's AI ambitions, and 4.7 billion dollars in CoreWeave, among others.
In return, these partners are typically expected to build on NVIDIA technology — whether for internal model training or for running enterprise and consumer cloud workloads.
Some may cite reports that many of NVIDIA's (NVDA) business arrangements rely on informal "handshake" agreements rather than airtight ironclad contracts.
Even if true, it is hard to imagine partners abandoning proven revenue streams and their own fiduciary duties in solidarity with Huang.
But ultimately, whether formal contracts or handshake deals, none compares to NVIDIA's (NVDA) most durable moat: the CUDA platform.
CUDA enables developers to write software that runs efficiently on NVIDIA GPUs, spanning graphics rendering, scientific simulation, financial risk analysis, and AI, among other domains.
Through years of developer training and a comprehensive set of complementary libraries and tools, NVIDIA (NVDA) has made CUDA the default standard.
It is as if software engineers spent years learning Italian, using tools with Italian-labeled buttons and reading manuals written in Italian.
To get them to switch to Finnish, the incentive would have to be too large to refuse — especially since a single misunderstood idiom could cause a catastrophic system crash, millions of dollars in extra costs, or even cost them their jobs.
NVIDIA's (NVDA) 42,000 employees will most likely continue designing chips, making strategic investments, expanding CUDA, designing solutions for vertical industries such as healthcare and transportation, collaborating in the lunch area, and performing the thousands of other tasks that employees of a large public company must do.
US-China Competition
Although many countries and regions play critical roles, the current global leaders in AI are the United States and China.
The US leads in areas such as frontier models, though its advantage is narrowing; China leads in cost-effective models and advanced manufacturing, including "physical AI" such as drones and humanoid robots.
Both countries are striving to maintain their advantages, and both have ample reason to cooperate.
Huang has publicly advocated expanding chip exports to China.
Critics often attribute his position to NVIDIA's (NVDA) huge revenue exposure in the China region; but current and former leaders of major US technology and industrial companies such as Intel, Apple, Boeing, and GE have also made the same argument for two core reasons.
First, their duty to shareholders is to maximize long-term company value, and forgoing access to the world's second-largest economy would undermine that goal.
Second, they believe overly aggressive export controls are counterproductive and accelerate the rise of state-backed domestic competitors.
This stands in stark contrast to the argument advanced by former US Commerce Secretary Gina Raimondo, who insisted that "national security matters more than short-term revenue."
However, the optimal approach requires a broader calculus: as recent government statements have hinted, sustainable policy depends on reconciling confrontation with diplomatic stability.
AI — along with closely related technologies such as quantum computing and communications, fusion energy, and advanced materials — are arenas of "coopetition": cooperation and competition occurring simultaneously.
NVIDIA (NVDA) already has competitors from both the US and around the world in many areas, such as AMD and its ROCm platform (roughly equivalent to CUDA) in GPUs, Amazon with Trainium, Google with its tensor processing units (TPUs), and others.
Coopetition between the US and China, and among US companies themselves, intensifies competitive pressure, thereby accelerating progress in these critical technologies.
Each player brings its own strengths to the table; it is hard to see how the ouster or return of a single CEO would dramatically alter the competitive landscape between companies — or between nations.
Huang may return as CEO, or he may serve as an advisor.
But ultimately, the true test of a visionary founder is this: whether they have built a brand, an organization, an architecture, and an ecosystem that can grow under their personal cultivation and are strong enough to thrive in their absence.
Comments