Noa Khamallah grew up without a microwave or a car, dropped out of school, and served prison time before he ever wrote a check to a startup. When he wanted access to Yann LeCun, the researcher widely credited as the godfather of modern AI, he did not seek a warm introduction. Instead, he cold-emailed him with the subject line: “former convict, street kid, turned IPO, turned VC.” LeCun replied within hours: “You are a statistical anomaly. Welcome in.”

Khamallah used the same approach when reaching out to rapper Fetty Wap and former NFL player Penny Hart, who is a supporter of his fund.

That unlikely background now sits behind a $10 million fund with concrete results. Khamallah’s Don’t Quit Ventures announced the second close of Fund I, Fortune learned exclusively, bringing the fund to 2.5 times investor capital. Three of its 17 investments are now valued above $1 billion, and nearly one in five of its bets have become unicorns, compared with roughly one in 70 for the venture industry overall.

Sources familiar with the fund said its cap table also includes a stake in Mira Murati’s Thinking Machine Lab, alongside disclosed positions in AMI Labs and Replit. The fund has already returned cash to limited partners through an early sale of part of its stake in one company, just 14 months after its first close, at 1.5 times its cost.

Don’t Quit Ventures also reflects a broader market dynamic that continues to favor smaller funds over larger ones, especially as venture capital has concentrated at the top and first-time launches have become harder to come by. Carta’s Q4 2025 data shows that funds under $10 million are still outperforming funds over $100 million in both cash returned to investors and growth in paper value. iCapital research found that smaller funds, defined as those under $275 million, returned 36%, compared with 24% for larger funds. In practical terms, a $10 million investment that grows tenfold barely affects a $1 billion fund, but in a $10 million fund it can transform the outcome.

At the same time, new funds have become increasingly rare. Only 101 first-time funds launched in 2025, the fewest in 14 years. Meanwhile, the 10 largest funds now capture one-third of all money raised, more than double their share five years ago.

Michael Ströck, whose firm Allocator One anchored Don’t Quit Ventures’ first close, said his firm reviews more than 800 funds each year and selects fewer than 1%. What convinced him to back Khamallah was not the pitch deck, but the combination of dealmaking instinct and operating experience.

“The quality of relationship building and being extremely relationship-driven and extremely commercial at the same time is very hard,” Ströck said.

He also rejected the idea that solo-run funds are inherently riskier than firms with co-founders, arguing that the risk math does not support that assumption. A single manager becoming incapacitated during a fund’s life is statistically rare, while split decisions among multiple partners are common. “I never understood the concern with solo GPs,” he said. “It makes no sense.”

The challenge, however, is durability. Smaller funds can swing more sharply between major wins and total losses, so if Khamallah raises a larger Fund II, the math that makes the model work will be harder to repeat.

Lily Mae Lazarus can be found on X at @LilyMaeLazarus. She can also be reached at lily.lazarus@fortune.com. Joey Abrams curated the deals section of today’s newsletter. This story was originally featured on Fortune.com.