Pushed Out of Your Own Company: Why Votes, Not Vision, Decide Who Stays

As of November 13, 2026. The Nvidia story is still developing, and details may change.

Jensen Huang co-founded Nvidia in 1993 and ran it for 33 years. He became the face of the entire AI compute era. And on 12 November 2026, Nvidia announced he was out as CEO and off the board, after reports of two affairs, one with a woman suspected of spying for China.

If your first reaction was but that's his company, you are not alone.

It isn't. Not in the way that matters. That gap between "the company he built" and "the company he controls" sent me down a rabbit hole, and what I found is something every founder should understand before they incorporate anything.

Founding Is Not Owning

Here is the chain of command at every company, private or public:

Shareholders own the company. Each share usually carries one vote. They elect the board.
The board of directors oversees the company on shareholders' behalf. It hires, pays and fires the CEO.
The CEO runs things day to day and reports to the board.

Notice where the founder appears in that list: nowhere. A founder is just a shareholder who happens to have a job title. Whoever controls a majority of votes picks the board, and whoever controls the board picks the CEO.

Nvidia's 2026 proxy statement listed Jensen as owning about 3.58% of the company as of March 2026. That is an absurd amount of money and almost no voting power. Nvidia has a single class of shares, one vote each, so his votes counted exactly like anyone else's.

And the board had real reasons to act. Nvidia sits at the center of US export controls on chips to China. A credible allegation that the CEO could be a security risk is not a PR problem, it is a regulatory and national security problem. Directors have a legal duty to the company, not to the founder. A board that ignored this could get sued by its own shareholders.

Jensen is not the first:

Year	Company	Founder	Outcome
2026	Nvidia	Jensen Huang	Out as CEO and off the board
2023	OpenAI	Sam Altman	Fired by the board, back within days
2017	Uber	Travis Kalanick	Forced out by investors
1985	Apple	Steve Jobs	Pushed out, returned in 1997

The Founders You Cannot Fire

Now compare that to SpaceX. When it filed to go public this year, its structure made Elon Musk effectively unfireable. According to the S-1, public investors get Class A shares with one vote each, while Musk's Class B shares carry ten. He ends up with about 85% of the voting power while owning a much smaller slice of the equity. The charter says he can only be removed by a vote of the Class B shares, which he controls. The only person who can fire Musk is Musk.

This is a dual-class structure, and the math is simple.

Say you own 15% of the company and the public owns the other 85%. If your shares carry 10 votes each and theirs carry 1, your 15% turns into 150 votes against their 85. That is 150 out of 235 total votes, or about 64%. You own less than a sixth of the company and control nearly two thirds of the vote.

Now run the same numbers with one vote per share. Your 15% is 15 votes out of 100. You control exactly 15%, which is Jensen's situation. Same founder, same stake, completely different fate.

Musk is not the only one:

Meta. Zuckerberg's super-voting shares give him majority control. In 2022, Meta's stock fell more than 60%, the worst performance in the S&P 500 that year, while he poured billions into the metaverse through Reality Labs, a division that lost $9.4 billion in the first three quarters of 2022 alone. Plenty of investors wanted it cut. Nobody could outvote him. Meta stayed profitable throughout because the ads business paid for the bet.
Alphabet. Larry Page and Sergey Brin stepped down as executives in 2019 and handed the CEO job to Sundar Pichai. But they kept their 10-vote shares. Per Alphabet's 2026 proxy, Page holds 27.4% of the total vote and Brin holds 25.3%, a combined 52.7%, which means they control the board, which means Sundar ultimately works for them. They walked away from the job without walking away from the power.

That last point is underrated. Dual-class shares do not just protect founders from being fired. They let founders leave on their own terms and come back when they want to, which Brin did when the AI race heated up.

The Trade-off Nobody Escapes

It is tempting to read all this and conclude founder control is strictly good. It is not. It is a bet on one person.

Scenario	Founder control	Board accountability
Founder wants a long bet investors hate	✓	✗
Wall Street demands short-term cuts	✓	✗
Founder steps back but wants to steer	✓	✗
Founder becomes a security or legal risk	✗	✓
Founder is simply wrong and will not stop	✗	✓

The Nvidia situation is exactly the row boards exist for. At SpaceX, which holds sensitive defense and NASA contracts, there is no equivalent emergency brake. Super-voting shares remove the brake for good reasons and bad ones alike.

Why Some Founders Limit Their Own Power

This is where sunset clauses come in: provisions that convert super-voting shares into ordinary ones after a fixed time or a specific event. Most dual-class companies actually have one, which might surprise you. A study of 259 US dual-class companies found almost four in five had at least one sunset provision, though most are tied to events or dilution rather than a fixed date, and only 24 used a time-based sunset. The reasons make sense once you see them:

Investors accept it more easily. Large institutions dislike permanent founder control. A sunset makes the IPO easier to sell.
Control should not outlive the person. Without an event-based sunset (death, departure, selling shares), super-voting stock can pass to heirs or estates who never built anything.
Judgment does not scale forever. Control over a five-person startup should not automatically become permanent control over a fifty-thousand-person company.
Sometimes it is required. India's SEBI framework for superior-voting shares mandates sunset provisions.

These are rarely pure generosity, though. They usually come out of negotiation, and research suggests founders with more bargaining power are less likely to accept time-based sunsets. Founders with enough leverage, like Zuckerberg or Musk, have avoided any fixed expiry date on their control. Investors wanted in badly enough to accept the terms.

What This Means If You Are Building Something

There are two ways founders end up with control:

Set it up early. When you and your co-founders are the only shareholders, adding a founder share class is a simple charter change. Nobody to negotiate with. The hard part is keeping it when investors arrive.

Add it later with leverage. Facebook created its dual-class structure in 2009, years after taking outside money. Google adopted its structure around its 2004 IPO. Both were hot enough that investors said yes.

In the early years, the things that actually protect you are more boring:

Keep a meaningful ownership stake. Early on, a large stake of ordinary shares is control enough. Track dilution every round.
Control the board. Before an IPO, the split of founder versus investor seats decides who is really in charge.
Set up vesting properly. It protects co-founders from each other.
Earn your leverage. A dual-class structure written into your charter is not permanent protection on its own. If your company struggles and needs money, a new investor can make collapsing those super-voting shares a condition of their check, and you may have no choice but to agree. What actually keeps the structure intact is being the founder investors badly want to back, so they accept your terms instead of setting their own.

Jensen built one of the most valuable companies in history and still lost it, because building and controlling are different things. The structure decides who stays. Know which one you are setting up before you need it.

Disclaimer

This post is for informational and educational purposes only. It is not legal, financial, tax, or investment advice. When you incorporate, a startup lawyer will set up share classes and vesting properly, and it is far cheaper to do it right on day one than to fix it later.

The content is based on public reporting, company filings, and other sources available as of November 13, 2026, all linked in the references below. Details about ongoing events may change, and while I have tried to keep everything accurate, I cannot guarantee it is complete or error-free. Allegations described here are reported claims, not established facts.

The author is not responsible for any decisions made or actions taken based on this post. Please consult a qualified professional before acting on anything you read here. The views expressed are my own and do not represent any organization I am affiliated with.

If you spot an error, or if you would like any section corrected or removed, please reach out through the contact page and I will review it promptly.

References

Jensen Huang ousted as CEO of Nvidia, investingLive
Huang, Face of the AI Boom, Ousted by Nvidia Board as CEO, Bloomberg Law
Nvidia CEO Holds 3.58% of Shares, With No Super-Voting Class, FourWeekMBA (from Nvidia's 2026 proxy statement)
SpaceX files for Nasdaq IPO with Musk retaining 85.1% voting control, investingLive
A look at the upcoming historic SpaceX IPO, SpaceQ
Who Owns Google?, AI Funding Tracker (from Alphabet's 2026 proxy statement)
Alphabet 2026 Proxy Statement (Form DEF 14A), SEC
Removal of Sam Altman from OpenAI, Wikipedia
Sam Altman returns as OpenAI CEO with a new board, Fortune
Uber CEO Travis Kalanick resigns under investor pressure, Reuters via Ahram Online
Uber CEO Travis Kalanick resigns under investor pressure, AP via ABC7
Why did Apple's board fire Steve Jobs in 1985?, The Corporate Governance Institute
Showdown at Apple: John Sculley vs. Steve Jobs, Mac History
Meta just dropped a metaverse bomb on investors, The Motley Fool
Year in review: how badly did Meta do in 2022?, Analytics India Magazine
Facebook creates dual-class structure, but no IPO, AP via NBC News (also covers Google's 2004 structure)
What are the implications of Facebook's multiple class equity structure?, Columbia Business School CaseWorks
Dual Class: "Sunset" Practices, TheCorporateCounsel.net (summarizing Minerva Analytics)
The Rise of Dual-Class Stock IPOs, CLS Blue Sky Blog, Columbia Law School