Founders Who Retained Large Stakes in Their Companies

Oracle's September 2025 proxy recorded Ellison's beneficial holding at 40.6% of its ordinary shares. Retention was only part of the story: Oracle reported reducing its outstanding share count by 35% through repurchases between the start of its 2016 financial year and the end of 2025. A founder's percentage can therefore rise because the denominator shrinks, even without buying an equivalent proportion of new shares.

Meta's April 2025 disclosure showed Zuckerberg retaining roughly 14% of its shares through his holdings and associated vehicles, with about 61% of the votes. The gap comes from ten-vote Class B stock. Outside shareholders supplied capital and shared in the company's economic results, while the founder retained decisive voting authority. The holding includes entities used for family and philanthropic purposes, rather than one personal brokerage account.

Michael Dell's disclosed holding was 41.4% of all Dell Technologies common stock on 30 April 2025. That position followed a substantial restructuring of ownership: he partnered with Silver Lake to take Dell private in 2013, before the business returned to public trading. His case concerns buying and reorganising a company as well as keeping founder shares; a simple narrative about never accepting outside investors would miss the point.

At the end of 2024, Ek's wholly owned investment vehicle held about 12.5 million Spotify shares, roughly 6.1% of the company. The much larger 14.3% beneficial-ownership figure also included shares over which he held voting proxies for Tencent-related holders. Spotify's 2018 direct listing had allowed trading without a new capital raise. Ek illustrates why a founder's economic stake, regulatory ownership figure and voting influence need separate treatment.

Lorentzon retained about 20 million Spotify ordinary shares through his investment structure at the end of 2024, roughly 9.8% of the company. His disclosed voting power was 41.6%, boosted by beneficiary certificates that carried votes but no economic rights. Alongside Ek, he had taken a company financed by several investors onto the public market. The surviving founder block was substantial, but its voting strength greatly exceeded its claim on profits.

Airbnb attributed about 67.2 million Class A and B shares to Chesky on 7 April 2025, a little over a tenth of its outstanding common shares. Those holdings, including estate-planning trusts, carried 30.5% of the votes because Class B shares had twenty votes each. Public trading and a large founder holding coexisted; the share structure preserved influence without requiring Chesky to own most of the business.

Coinbase's March 2025 disclosure identified 24.1 million Class B shares in Armstrong's living trust. Its larger beneficial-ownership total also included exercisable options and shares of another trust for which he was trustee. Twenty votes per Class B share gave his reported holdings 50.2% of voting power. The distinction matters when reading founder fortunes: voting authority and regulatory attribution do not make every reported share the founder's personal economic property.

Moskovitz reinforced his Asana stake by putting more money into the business, including a $350 million purchase of newly issued shares in 2022. His August 2025 filing listed 56.9 million Class A shares held personally, alongside Class B stock and trust holdings. This was active concentration of capital as well as retention. The filing's broader beneficial total included other arrangements, so it should not be treated as an uncomplicated personal holding.

Snap reported roughly 169 million shares across its three classes for Spiegel at the end of 2024, including trust holdings. The economic block remained sizeable after the flotation, while high-vote Class C shares gave him 53.1% of total voting power. Snap's publicly traded Class A shares had no votes. Its structure made outside capital compatible with founder control, but left public investors with unusually little influence over corporate decisions.

Murphy's disclosed Snap holdings totalled roughly 181 million shares across three classes on 31 December 2024. Although that was more shares than Spiegel's reported total, Murphy held 46.4% of the votes because the founders' mix of share classes differed. The co-founder's continuing financial exposure was substantial. Comparing the two also shows why counting shares alone cannot explain who controls a company with several classes of stock.

Roku's April 2025 proxy attributed about 19.1 million Class A and B shares to Wood, including exercisable options and trust holdings. Most were Class B shares held in a revocable trust shared with his spouse, and the reported position carried 56.7% of voting power. Ten-vote stock preserved his authority through public ownership. The options and joint trust are material details when interpreting the figure as personal wealth.

Atlassian grew for eight years before accepting Accel's minority investment in 2010. On 30 September 2025, Cannon-Brookes still held about 48 million Class B shares through a trust, roughly 18.2% of all outstanding shares, with 42.59% of the votes. Early self-financing helped limit initial dilution; later outside investment and public trading did not erase the founders' substantial economic positions.

Farquhar's September 2025 Atlassian disclosure also showed about 48 million Class B shares, held through an investment partnership. That was roughly 18.2% of the whole company, rather than the 50% figure applying only to its Class B stock. He had already left the co-chief-executive role while retaining the stake. Ownership can outlast day-to-day management, and planned share sales need not mean a founder has sold out.

Zoom's March 2025 disclosure recorded 21.6 million Class B shares in a revocable trust for which Yuan and his spouse were co-trustees, plus about 84,000 shares due from awards. The 51.1% figure applied to Class B stock alone; across both classes the stake remained a minority interest. Its 31.6% voting power demonstrates how public ownership and substantial founder influence can coexist without majority economic ownership.

Pinterest's disclosure for 26 March 2025 attributed about 38.8 million shares to Silbermann, roughly 5.7% of both classes combined. The same holding represented 46.97% of Class B stock and 34.54% of total voting power, three very different measures. He had stepped away from the chief-executive role in 2022 while retaining an ownership position. Twenty-vote shares extended his influence beyond the financial size of that stake.

Shopify's April 2025 circular recorded 78.9 million Class B shares and about 632,000 Class A shares owned, controlled or directed by Lütke, roughly 6.1% of the ordinary share base. A separate Founder Share helped bring the associated voting power to 40.03%. That arrangement preserved influence as the company grew with outside shareholders, but it did not turn a six-percent economic interest into a forty-percent entitlement to its value.

Inditex's accounts recorded Ortega's 59.294% stake on 31 January 2024, held through Pontegadea and Partler. The group's 2001 flotation had offered about 26% of its shares to the market, leaving a large ownership block intact. Unlike examples based mainly on high-vote stock, this was a majority economic interest. The dividends attached to those shares were separate from Inditex's sales, cash balances and total stock-market value.

Yanai, who developed Uniqlo from the family clothing business, personally held 17.4% of Fast Retailing's shares excluding treasury stock on 31 August 2025. Its shareholder table listed his sons and investment companies separately. That distinction prevents a broad family holding from being described as his alone. Public ownership supported a global retailer while leaving the founder of its Uniqlo business with substantial direct exposure to its results.

SoftBank Group reported Son's holding at 29.68% on 31 March 2025, excluding treasury stock from the denominator. The company separately listed other shareholders and entities. His large stake sat at the parent-company level: it did not make every business in SoftBank's funds his personal property. The structure combined a concentrated founder interest with capital from public shareholders, lenders and outside fund investors.

PDD's March 2025 ownership table attributed 24.8% of its ordinary shares to entities affiliated with Huang, held through a trust benefiting him and his family. He had left the chairmanship while retaining that substantial financial connection to the business. The table showed no outstanding Class B shares at that date. Earlier accounts of overwhelming founder voting control therefore cannot simply be carried forward to describe this later ownership snapshot.

Xiaomi's end-2024 disclosure traced about 4.07 billion Class A shares to a holding company ultimately owned by a trust established for Lei and his family. Those shares alone represented roughly 61.9% of voting rights on ordinary shareholder matters. They were only part of the economic holding, which also included Class B shares. The weighted-vote structure preserved control alongside outside financing, while the trust makes family attribution important.

Baidu's January 2025 report attributed about 452.5 million shares and near-term share entitlements to Handsome Reward, a company wholly owned by Li, representing 16.4% under its reporting rules. His broader beneficial-ownership figure reached 19.1% partly because employees had granted him voting proxies. The retained investment was substantial without counting those employees' economic interests as his. High-vote stock and proxies explain why his voting power was greater still.

JD's February 2025 table attributed 11.7% of ordinary shares and 71.7% of aggregate voting power to Liu. The founder remained heavily invested after the business expanded beyond its original retail operation and attracted outside shareholders. Class B shares carried twenty votes each, so the two percentages described different rights. The disclosed position also involved holding entities and share-plan arrangements, rather than a simple percentage of company assets held personally.

Geely Automobile's end-2024 report recorded a 41.15% interest for Li through controlled corporations and a further 0.23% held personally. Its notes traced the corporate holdings through Geely Holding, Geely Group and related interests. This is a substantial founder-linked block in the listed carmaker, but the attributed holding should not be confused with sole personal ownership of every Geely-branded business or associate.