📶 Leo Polovets (GP @ Humba Ventures) on what makes a great LP, attributing luck to seed investing, and using transparency to your advantage

One big misconception is that young funds spend most of their time on investing. For the first 6-18 months, there's so much other stuff to work on: you need to actually raise the fund -- which often takes the majority of your time; you need to set up a bunch of back office systems and operations; you might need to spend time on recruiting founding partners/teammates; and you need to quickly build a brand and presence in the ecosystem. It's very hard to get inbound deal flow for a new fund without a strong brand, and developing a good reputation takes a lot of time and effort.

All of these things take a lot of work, and most new managers spend much less time than they expected on the fun part of the job: actually investing and working with founders.

I believe deep tech is the best startup category to invest in. It's less capital intensive than most people assume -- especially with the rapidly declining costs for primitives like hardware components and genome sequencing. The exit periods are good, the barriers to entry are high for competitors, and the moats are stronger. I think moats and barriers to entry are especially critical in the era of AI coding, given the speed at which software products can be built and cloned these days.

The VC market is slowly warming up to this POV in 2025, but it was not a common point of view a few years ago when Humba started.

I've written a bit about what makes deep tech such an attractive area for builders and investors here and here.

Founders prepare materials like decks and memos to get investors up to speed, and it seems only fair that VCs should do the same for founders. So we have a few pages on our site like this one that describe areas where we can help, what our investment process is like, and so on. I think this kind of transparency is a good thing. I'd rather be clear and opinionated and try to attract the kind of founders I hope to work with, than try to be everything to everyone by not standing for anything. I also hope that these fund one-pagers show that Humba cares about founders, and that we are willing to publicly commit to delivering a great experience.

Very important: high integrity, believes in your thesis and the future you're investing in. Ideally they are interested in growing the relationship and investing in several funds, instead of being a one-time LP.

Important: pleasant to work with OR very hands off :)

Nice to have: helpful in some way. This could mean they know some industry well, they have a great network, or they can share benchmarks and learnings across funds they work with.

All of this is similar to what most founders want from a VC: committed/aligned capital, nice to work with, and ideally helpful but at least not value-destructive.

A good portfolio allocation should cover many asset classes at any one time. If you think you can time the market and only be invested in the best asset class at any time, then 1) good luck and 2) you're wrong. So regardless of whether 2025 eventually turns out to be a great year or a terrible year for VC, sophisticated LPs should be allocating to VC and to other asset classes on an ongoing basis.

Diversifying within VC is important as well. If you want to only invest in ten SaaS funds or ten consumer funds, that's fine, and in some years those will be the best VC categories. But if you want to have good coverage of tech startups, you should invest in VC funds with varying stage and sector focuses.

On that note, I think Humba provides great coverage of deep tech, investing in everything from robotics and biotech to defense and energy. As a GP, I'm technical and I have been in venture for well over a decade, both of which are rare among early stage deep tech funds. These are the things that stand out most to LPs about Humba. Coincidentally they are also what stand out to founders: my background and my teammate, Anna's, background* are unusual, and my years of venture investing across many categories have given me a set of lessons and data points that very few deep tech VCs have.

* Anna previously created a great deep tech publication called Foundations and Frontiers, and she's very good at helping companies tell their stories and explain sophisticated technologies in a memorable and publicly accessible way.

I think luck and skill are probably 50/50 here.

On the skill side, you need to develop a good reputation so that founders want to talk to you, and want to take your money even when their rounds are very oversubscribed. This reputation helps you get good deal flow, and if you don't have good deal flow then picking ability doesn't matter. And picking ability is also a skill. If you look at great seed managers vs. poor seed managers, the great ones have a 10%+ hit rate from seed to eventual $1b+ outcome, while the poor ones have a ~0% hit rate. Picking and deal flow can be developed over time and are within a manager’s control.

But luck also plays a big role. Let's say you're a great investor, and for each of your seed investments there's a 10% chance it'll become a unicorn. If you make 20 investments over 5 years, you have a 1 in 8 shot of having at least four unicorns in the portfolio. That would be an incredible outcome and LPs would line up to invest in your future funds. But there's also a 1 in 8 shot of having zero winners, and then you'd have a very hard time raising future funds. You're the same investor and your picking skill hasn't changed, but luck played a huge factor.

Here's one specific example of luck: during my first decade in venture, I would write a few angel checks each year where I liked the founders and what they were working on, and where I thought I could accelerate my learning as an investor about a sector or business model or something else. Over this decade, my worst year was 6 angel investments with a 0.3x overall return. But during the subsequent year, I made 5 angel investments with a 37x overall return. My investing strategy didn't change much year to year, and the quality of companies I saw was similar in both years, but one year was literally 100x better than another, and that was just luck.

We keep reserves low. I think large reserves make more sense for $100m+ funds, where you can't easily keep growing your initial check size. E.g. if you are targeting 15% ownership, then you can't just double your next fund size and start targeting 30%, because seed rounds are simply not that large. So if you double your fund size, maybe you keep targeting 15% but you reserve more and more for follow-on financing as your funds grow. But if you're targeting 2% or 5% or 8%, it's better to try to double your initial ownership than to do a lot of follow-on. Follow-on investments let you deploy more cash into good companies, but they’re less attractive because 1) they’re hard to do well, 2) future rounds are much more expensive than the first round, and 3) your return multiple typically drops. Humba is a $40m fund, and we keep our reserves to a minimum. 75%+ of the fund is for initial checks.

It’s been increasingly popular to tell founders to target 10-15% dilution for a seed round. While it’s smart to be dilution sensitive, the biggest risk to your company’s success is existential risk and not a little bit of extra dilution. Companies that don’t raise enough capital often struggle later because seed timelines are uncertain and it’s hard to predict how long it’ll take to find product/market fit. That’s why having a bigger cash cushion is generally better.

I think the prudent thing to do is to raise whatever capital you need to maximize your chances of success. If that’s 10% dilution, great, and if it’s 30% dilution then that sucks but it’s still better than owning a larger piece of a company that fails. Once you have product/market fit, it’s much easier to optimize for dilution in future rounds.

A lot of huge multistage funds will shrink dramatically as software companies become much less capital intensive thanks to AI.

Too many to name, but a few close friends that I admire are Mike Annunziata at Also Capital, Joe Wilson at Undeterred Capital, Julian Shapiro at Julian Capital, and Lucas Vaz at Ravelin Capital.