Doug Leone Says the Biggest Mistake in Venture Capital is Selling Winners Too Early; True Super Returns Come from Long-Term Holding Rather Than Constantly Seeking New Projects
Doug Leone, a partner at Sequoia Capital, stated in an interview that the biggest mistake in the venture capital industry is not investing in the wrong companies, but selling true winners too early.
He pointed out that Sequoia sold its approximately 20% stake in NVIDIA, sold after being the first institutional investor in Apple, sold after holding about a quarter of Cisco, and sold after holding about 10% of Google. If they had not sold Google, that holding would be worth about $500 billion today. He noted that great companies can compound growth for over 20 years after their IPO.
Leone stated that if a company is confirmed to be facing an infinite market, no shares should be sold. Traditional funds emphasize paper IRR and a 10-year exit, locking in returns to distribute to LPs, which leads to missing out on subsequent compounding. He gave an example that leaving $5 billion in the market could grow to $50 billion.
He pointed out that all venture capitalists have made this mistake because great companies continue to compound after their IPO. Sequoia has historically sold high percentages of early holdings due to fund cycles. Leone emphasized that true super returns come from long-term holding rather than constantly seeking new projects.
Market Mechanism: Early exits transfer long-term compounding assets to secondary market buyers, benefiting short-term IRR but missing out on subsequent multi-fold growth; investors who can identify infinite markets and extend holding periods benefit from continuous compounding, while institutions constrained by closed fund structures face reinvestment pressure.
Source: Public Information
ABAB AI Insight
Doug Leone has been with Sequoia since 1988 and has co-led the firm with Michael Moritz since 1996, overseeing investments including ServiceNow and driving Sequoia's entry into China and India. After stepping down as global managing partner in 2022, he will become chairman in 2026 and reflected on past exit decisions in an interview with David Senra.
In terms of capital path, Sequoia operates with a 10-year closed fund structure, distributing returns to LPs after holding high percentages of Apple, Cisco, Google, and NVIDIA early on, thus ceding subsequent compounding growth to the public market. Leone advocates never selling shares in companies confirmed to be in infinite markets, keeping capital on the compounding path rather than locking in paper returns early.
This contrasts with partners like Mike Moritz who choose to hold some investments long-term, similar to Warren Buffett's approach of holding quality companies rather than frequent trading. Current venture capital is transitioning from a short-term IRR focus to identifying long-term compounding opportunities, with Sequoia as a representative of early tech investment, its exit model influencing industry standards.
Essentially, this relates to capital concentration and the transfer of pricing power: when fund structures force exits within a limited cycle, the pricing power of long-term compounding assets shifts from early investors to secondary market holders. The mechanism is that closed funds need to return cash to LPs to lock in performance, leading to early reductions in holdings of companies in infinite markets, allowing subsequent decades of growth to be captured by public market investors, while early capital cannot continue to share the full compounding path.
ABAB News · Cognitive Law
Winners who cash out too early are the biggest losers.
A 10-year fund cannot capture 30 years of compounding.
For companies in infinite markets, not a single share should be sold.