Nvidia Board Announces Jensen Huang Resignation After 33 Years

Nvidia begins its first leadership transition in company history without the co-founder who built it, after its board announced that Jensen Huang had stepped down as chief executive 33 years after he started the chipmaker.

The board must now replace a founder who set the company’s technical direction personally and, by his own account, never prepared anyone to follow him.

The resignation was announced by Nvidia’s board, whose members now face the choice of hiring another innovator from outside, promoting an operations-focused executive, or searching deeper in the organization for someone with Huang’s instincts.

Steve Blank, writing in an expert perspective for The Cipher Brief, argued the board’s decision will determine whether Nvidia keeps reinventing itself or coasts on momentum and brand.

Three reinventions built the company

Huang transformed Nvidia’s business model three times in under two decades, each shift changing what the company sold and to whom, Blank wrote. In 2006, Huang launched CUDA, turning a graphics chip made for PC gamers into a general-purpose computing platform, and paid for it for years before it earned anything. In 2012, after researchers used Nvidia chips to win the ImageNet contest, he steered the company toward deep learning. In 2020, the acquisition of Mellanox moved Nvidia from selling boards with GPUs to selling integrated racks that power AI data centers.

“These were not just product transitions. They were radical business model transitions involving new channels, new customers and new markets,” Blank wrote in The Cipher Brief. In his judgment, a chief executive focused on executing an existing business model would never have funded CUDA through those years without revenue.

Blank, who writes about innovation and entrepreneurship at steveblank.com, compared Huang to Steve Jobs, who reinvented Apple three times between 2001 and 2008. “Visionary CEOs are not just great at ensuring world-class execution of a successful business model. They are also world-class innovators,” he wrote.

The warning from Redmond

The precedent Blank put before Nvidia’s board is Microsoft. Bill Gates handed the company to Steve Ballmer in January 2000, after running it for 25 years. Ballmer organized the company around Windows and Office, tripled sales to $78 billion and more than doubled profits, from $9 billion to $22 billion.

Those numbers, Blank wrote, hid five missed markets that reshaped computing: search, where the company lost to Google; smartphones, lost to Apple; mobile operating systems, lost to both; media, lost to Apple and Netflix; and the cloud, lost to Amazon. Only in early 2014, when Satya Nadella reorganized Microsoft around Azure, freed Office and Azure from Windows and shut the phone business, did the company recover.

Today Microsoft is worth about $4 trillion, one of the few companies that regained its former standing after a board replaced a salesman with a product insider.

Apple offers a softer version of the same story.

Tim Cook ran the company for 15 years and grew revenue from $108 billion to $416 billion, lifting market value from $350 billion to about $4 trillion, Blank wrote. But Apple stumbled with Vision Pro, missed on AI despite Siri’s early lead, and this year turned to Google to help fix Siri. In September 2026, Cook handed the job to John Ternus.

First statement from Jensen Huang since resignation as Nvidia CEO — CNBC Television

One visionary, 60 direct reports

The scale of the replacement problem at Nvidia is unusually large. Huang had about 60 direct reports, discouraged one-on-one meetings and ran without long-term plans, Blank wrote. Like Jobs, he was the point where technical, product and business decisions converged. He said he did not believe in succession planning.

“He said he didn’t believe in succession planning,” Blank wrote, adding that Huang anointed no successor before leaving. When visionary founders depart, their operating executives believe it is their turn to run the company, he noted; Gates chose Ballmer, Jobs chose Cook.

Huang chose no one, and left the board the same day he left the job.

An execution-focused successor’s first moves carry their own risks, Blank argued. No one else can run the company with 60 direct reports, so the next chief executive will likely add layers of management.

That brings predictability, but it can start a creative death spiral in which talented innovators leave and process-oriented hires replace them. Amazon points the other way: before succeeding Jeff Bezos, Andy Jassy had founded and led Amazon Web Services — an innovator who had built a new business inside the company.

Competitors inside the customer list

The threat to the business is visible now, not hypothetical, according to the analysis. Nvidia’s biggest customers — Google, Amazon, Microsoft, Meta, OpenAI and Tesla — are building their own AI chips. Huang’s answer was to bring a new generation of chips to market every year, wall off the platform with CUDA software, and invest roughly $50 billion in the AI labs that run on Nvidia hardware.

“These may be the very bets an execution-focused CEO trims first,” Blank wrote. His conclusion about the stakes: “In the GPU market, the difference between market dominance and Chapter 11 may be a hands-on CEO who understands products, customers and business models.”

Timing offers the board no shelter. An inherited order book and the installed base around CUDA could make the next two or three years look strong regardless of who takes over, Blank wrote — unless a meaningful share of that revenue involves chips diverted to China, a question inherited along with a Congress and investigators asking how many restricted Nvidia chips reached that country.

Huang also set Nvidia’s foreign policy personally, as its face in Washington, Beijing and Taipei; no Nvidia operating vice president has done that job.

Four problems the board inherits

Blank laid out four difficulties for the board. First, distinguishing execution from innovation: Jobs and Gates both chose execution executives as successors, confusing operational excellence with product passion and market insight. Nvidia’s version is sharper — with the founder gone and no successor named, the decision falls to a board in crisis under its first-ever chair, and boards in crisis reach for stability, which produces an execution chief executive by a different route. Second, retention. If the board hires an innovator from outside, the strongest executor — the No. 2 or No. 3 executive — will likely leave, convinced the job was theirs. Third, the founder’s shadow.

Huang, like Jobs, Mark Zuckerberg, Elon Musk, Walt Disney and Henry Ford before him, is part of the company’s brand and a touchstone for internal decisions. Years from now, employees may still ask what Huang would have done rather than what the changing market requires. Fourth, geopolitics without retreat.

Adopting the language of the industry cards the sources put forward, the arithmetic runs against caution: “Executing today’s business is essential. So is reinventing it before someone else makes it obsolete,” Blank wrote. “Execution pays your salary. Innovation pays your pension.”

What the board has not yet said is which kind of chief executive it wants — and no name has emerged for the job.

Photo of author

World Desk

World Desk at Archyde.

MakeUseOf: 5 TV Settings to Fix Before Playing GTA VI