On November 12, Nvidia announced that co-founder Jensen Huang had resigned as president, CEO, and a member of its board of directors. The board had accepted his resignation the previous day, according to a regulatory filing. CFO Colette Kress was named interim CEO, while Harvey Jones became Nvidia’s first-ever board chair. The company also began searching for a successor.
The announcement ended Huang’s 33-year tenure at the helm of Nvidia. Since co-founding the company in 1993, he had overseen several major strategic shifts, transforming a chip startup into one of the world’s most influential technology companies.
Yet less than five months earlier, Huang had received approximately 99% support in the director elections at Nvidia’s annual shareholders’ meeting. The contrast between that vote of confidence and his sudden departure has renewed attention on a corporate governance question long overshadowed by Nvidia’s commercial success.
Why did Huang, an Nvidia co-founder and the company’s longtime leader, never secure control through a dual-class share structure, as Google’s Larry Page and Sergey Brin and Meta’s Mark Zuckerberg did?
Dual-class share structures were uncommon in 1999
Part of the answer may lie in the ownership structure Nvidia adopted when it went public in 1999.
At the time, Silicon Valley technology companies commonly followed the principle of one share, one vote when listing on public markets. Founders could influence corporate decisions through their shareholdings, executive positions, and reputations, but they did not automatically receive greater voting rights than ordinary shareholders. Nvidia followed this convention, issuing only common stock.
Dual-class share structures were not an invention of the internet era. Several prominent publicly traded US companies had adopted such arrangements in the previous century, using shares with different voting rights to preserve control. Ford and The New York Times were notable examples.
What emerged later, during the internet era, was the practice of using dual-class shares specifically to protect technology companies’ long-term strategies. Such structures became increasingly common among major technology and internet companies pursuing IPOs.
When Google went public in 2004, it adopted a dual-class share structure. Its Class A shares, held by public investors, carried one vote each, while Class B shares, held by the founders and other insiders, carried ten votes each.
Google’s IPO prospectus argued that a conventional public-company structure could undermine corporate independence. Its dual-class arrangement was intended to preserve management’s long-term control, allowing the company to pursue strategic objectives without undue interference.
Nvidia, by contrast, had gone public five years earlier. Its decision to retain a one-share, one-vote structure was consistent with the corporate governance conventions prevailing among technology companies at the time.
Moreover, Nasdaq rules restrict publicly traded companies from disproportionately reducing existing shareholders’ voting rights after listing, including through the subsequent issuance of shares carrying superior voting rights.
For an already listed company such as Nvidia, introducing supervoting shares to protect its founder would therefore face substantially greater regulatory and governance obstacles than establishing such a structure before its IPO.
Huang’s ownership stake kept shrinking
Nvidia’s relatively dispersed ownership was evident from its IPO.
According to its 1999 prospectus, Huang owned approximately 13.1% of Nvidia before the offering, a stake that fell to 10.8% afterward. By 2026, his ownership had declined to 3.58%.
Fundraising, additional share issuances, equity compensation, and changes in Huang’s personal holdings reduced his ownership percentage over time. His voting power declined accordingly, leaving him with limited ability to determine the board’s composition or decisions through his shareholding alone.
Silicon Valley had seen founders lose control as their ownership stakes diminished long before Huang’s departure.
In 1985, Steve Jobs became embroiled in a dispute with Apple CEO John Sculley over the company’s direction. Although Jobs remained Apple’s largest individual shareholder, his ownership had declined to approximately 11%, leaving him without sufficient voting power to dictate the board’s decisions.
The board sided with Sculley. Following a restructuring of the company’s operations, Jobs was stripped of his management responsibilities. He left Apple, the company he had co-founded, in September that year.
Microsoft offers a contrasting example. After Microsoft went public in 1986, Bill Gates retained approximately 45% of its shares. Although his ownership gradually declined, he still held around 14% in 2000, the year he stepped down as CEO.
Gates benefited from his substantial ownership at the time of Microsoft’s IPO, the relatively gradual dilution of his stake, and the leadership authority he had accumulated over many years. Together, these factors helped him retain considerable influence throughout his tenure as CEO.
Huang was in a different position. When Nvidia went public, he owned approximately 11% of the company, substantially less than Gates’ stake in Microsoft at its listing.
As Huang’s ownership continued to decline, his already limited voting influence weakened further. He could not rely on his shareholding alone to determine Nvidia’s corporate decisions, even as his operational authority grew.
Huang’s 33-year leadership rested on his track record
Huang’s influence over Nvidia cannot be explained solely by the voting rights attached to his shares.
Corporate leadership also rests on credibility earned through performance, strategic judgment, and the board’s continued confidence in senior executives.
From graphics chips to programmable GPUs and sustained investment in the CUDA ecosystem, Nvidia repeatedly committed to long-term technological strategies before their commercial potential became widely apparent.
CUDA was particularly significant. Although it struggled to generate substantial immediate returns in its early years, the platform gradually established a competitive advantage through the integration of hardware and software.
Later, the rise of generative artificial intelligence drove a surge in demand for computing power, helping Nvidia evolve from a chip supplier into a central player in global AI infrastructure.
These strategic decisions, and the commercial success they delivered, reinforced Huang’s position as Nvidia’s leader.
In day-to-day operations, Huang maintained a central role in setting product development priorities, allocating R&D resources, coordinating supply chains, and managing customer relationships. Nvidia’s relatively flat management structure further strengthened his influence over operational decisions.
For the board and institutional investors, supporting Huang meant more than recognizing his past achievements. It also reflected confidence in his ability to guide the company’s future.
As a result, although Huang never possessed outright legal control of Nvidia, he exercised considerable authority over its strategic direction for decades.
His track record helped him maintain that position for 33 years without the protection of a controlling ownership stake. The approximately 99% support he received in the June 2026 director elections demonstrated the confidence shareholders continued to place in him.
But authority based on performance and confidence is conditional. It can strengthen as a company succeeds, yet weaken if business performance deteriorates or concerns arise over a leader’s judgment or conduct.
For someone like Huang, who lacked outright voting control, a loss of confidence could jeopardize the board’s support and, ultimately, his position at the company. His departure highlights the vulnerability of founder-led businesses whose leaders lack special voting protections.
Who do dual-class shares really protect?
Nvidia’s governance structure gives its board considerable authority. Under the company’s bylaws, a majority of directors can remove senior executives at any time, with or without cause.
Even as a co-founder, Huang could not use his ownership stake of less than 4% to veto the board’s personnel decisions.
Without voting control, his ability to remain CEO depended on maintaining the board’s confidence, supported by business performance and shareholder backing.
Because Huang could not control the board through his own voting rights, directors retained the ability to exercise meaningful oversight and, when necessary, initiate changes in senior management.
Seen in this light, Huang’s departure illustrates the implications of Nvidia’s longstanding ownership structure. His influence was substantial, but it did not override the board’s formal authority.
In an interview with TechCrunch, Yale University professor Jeffrey Sonnenfeld observed that Huang’s ability to lead Nvidia for more than three decades without a dual-class share structure offered a compelling argument against such arrangements.
Yet, a counterpoint exists: had Nvidia granted Huang supervoting rights when it went public, removing him from leadership would have been considerably more difficult.
Google and Meta are notable examples of companies that use dual-class share structures.
Supervoting rights give founders greater control, allowing them to withstand short-term market volatility and shareholder pressure while pursuing long-term strategies. But when concerns arise over a founder’s judgment or conduct, ordinary shareholders may have limited ability to change the company’s governance through voting.
Dual-class shares therefore strengthen founders’ strategic autonomy while making it harder for outside shareholders to intervene.
Nvidia’s decision not to adopt such a structure was consistent with the circumstances surrounding its 1999 IPO. It also left the company with a governance model fundamentally different from those of technology giants such as Google and Meta.
Rather than permanently entrenching its founder’s control through special voting rights, Nvidia operated under a structure in which Huang’s leadership depended on continued confidence in his performance and judgment. The arrangement preserved the board’s authority to oversee management and intervene when necessary.
Huang’s resignation underscores a distinction in corporate governance: however much influence founders accumulate through business success, those without controlling voting rights remain subject to the authority of their boards.
KrASIA features translated and adapted content that was originally published by 36Kr. This article was written by Huang Yida for 36Kr.





