Cendana Capital's Michael Kim on 16 years of seed fund investing: pattern recognition, the Yale model, and the arms race for young founders

Cendana Capital's Michael Kim on 16 years of seed fund investing

Michael Kim runs Cendana Capital, a fund-of-funds with roughly $3B under management that invests exclusively in seed and pre-seed stage venture funds. His LP base is largely US endowments and foundations. The firm acts as lead investor in over 80% of its funds and was built on a single early conviction: that super angels institutionalizing in the late 2000s were becoming the real engine of early-stage venture, not the traditional Series A firms.

Pattern recognition over 16 years

Kim started raising Cendana's first fund in 2010 and closed it in 2012 after 18 months of work. The core thesis has held, but he admits the scope of his ambition didn't. He passed on Josh Kushner's first fund and Micky Malka's Ribbit Capital, both of which he now describes as roughly 50x funds, because he wanted managers to stay inside the seed sandbox. In hindsight, he was too disciplined.

The firm has never invested in a first-time investor, but has backed many first-time funds. Kim argues that domain expertise and networks have a shelf life — being a VP at Google a decade ago doesn't travel as far as it once did — and what he's really underwriting is hustle, intellectual curiosity, and access to exceptional founders.

“The thesis was that these super angels were starting to institutionalize and bring outside capital in. Today, we have almost 3,000,000,000 under management. Our LP base is largely US endowments and foundations... We are convinced that early stage investing is still the way to go in this world of AI.”

The arms race for young founders

The most active thread in Kim's current thinking is what he calls an arms race for younger founders. AI-native companies can generate substantial revenue with small teams, which means their next funding round often looks more like a growth round than a traditional Series A. That compresses the window between seed and scale, and puts a premium on investors who are embedded in the places where young technical talent concentrates.

Neo, backed by Ali Partovi, is Kim's clearest example. Partovi built a sourcing engine through the Neo Scholars Program, recruiting on campuses at Harvard, MIT, Stanford, and CMU. The result: first checks into Cursor, Kalshi, Cognition, Ramp, and others. Cendana has backed Neo since fund one.

Kim also points to Nova, a fund run by Carlo and Coco (ages 25 and 23, with a 19-year-old third partner, Henry), and Corey Levy at Z Fellows, as examples of funds winning through campus proximity rather than traditional institutional pedigree. Josh Browder, a Thiel Fellow and DoNotPay founder, sits on the Thiel Fellowship selection committee, giving him a direct pipeline to early-stage technical talent.

Access versus picking ability

Kim is candid that he weights access and sourcing infrastructure heavily, partly because a fund with 20 to 30 portfolio companies has more shots on goal than a single angel bet. But he doesn't dismiss picking — he's currently backing an unnamed manager who worked alongside Nat Friedman and Daniel Gross, and whose main credential is having seen what great looks like at close range and being in the right rooms now.

On the question of whether former founders make better seed investors, Kim's answer is nuanced. At the earliest stages, most of Cendana's fund managers are ex-operators or founders, because early-stage companies actually want hands-on help. Later-stage investing skews toward ex-bankers, consultants, and lawyers. Immad Akhund, CEO and founder of Mercury, is one example of a current operator running a seed fund — Kim argues that perch gives him founder credibility that translates directly into deal access.

Portfolio construction and marking discipline

When diligencing a new fund manager, Cendana runs a founder-centric process: Kim's team calls every founder the manager has ever backed to assess whether the investor actually added value, and whether the fund's stated portfolio construction plan — say, 20 companies at $1M each — matches what founders actually received.

On valuation discipline, Kim is pointed about the "SaaSpocalypse." SaaS companies that raised at 50x revenue between 2016 and 2022 — the cohort he calls the "messy middle" — are now growing at 10% with $100M in revenue, in a market where software multiples have compressed to roughly 3.5x. Those companies are worth $300M to $400M at best today. He's seen fund managers proactively mark those positions down; others haven't, and Cendana knows the difference.

The Yale model and LP narrative

On how endowments came to allocate heavily to venture, Kim credits David Swensen at Yale for pioneering the endowment model. Yale's portfolio runs roughly 60% in private markets, with around 25% in venture alone. Kim's view is that institutional LP interest in venture is narrative-driven by design — power law returns mean a handful of companies generate most of the gains, and those stories pull capital in. He cites one undisclosed endowment that put $4M into Takashi, now worth $1.5B, as the kind of outcome that makes the asset class legible to institutional allocators.

The regrets

Kim is in Founders Fund II, which he says is approaching a 300x return, and Blockchain Capital II, currently at 157x and potentially heading toward 250x. He knows Chris Sacca but wasn't in Lowercase One, which returned 204x. Not having more in those funds is, in his words, the regret.

The consistent tension in Cendana's model is the same one facing any fund-of-funds: you give up the ability to go all-in on the single breakout bet in exchange for a more systematic exposure to the asset class. Kim has made peace with that trade-off, but the anti-portfolio is a real number.