What Do You Invest in?

Spiky founders building category-defining companies 12-36 months before it would appear on a market map

I finally figured out the simplest framework to what I’m looking for at seed.

As a generalist investor, this question has always been difficult to answer because I can’t use sector or layer of the stack to answer. If you look back at the publicly-announced investments I’ve led at Susa from nuclear microreactors to fish-processing robots to AI-native fund admin software, it’s very hard to draw a clear throughline.

After too many fumbling conversations where I hemmed and hawed my way through an answer when other VCs, LPs, or founders asked me this question, I’ve finally synthesized my answer to a simple framework:

Spiky founders building category-defining companies 12-36 months before it would appear on a market map

Let’s break down the 3 pillars.

Spiky Founders

People talk about exceptional founders, generational founders, etc. but it’s somewhat unclear what they mean.

Spiky is a word we use all the time at Susa: there is just something about the founder that is different and SPIKES. They are top 1 percentile on some trait or characteristic, it’s something that separates them from their peer class of founders: intelligence, drive, domain expertise, grit through humble or difficult beginnings, etc.

A well-rounded person is a mediocre person.

It’s easy to feel comfortable investing in someone whose credentials check the boxes, seems “generally smart/competent,” or has some level of “founder-market fit.”¹ Those rarely turn out to be good investments. Even if the founder gets lucky with an early insight/product or market pull, their return on luck tends to be low: they don’t convert that early piece of luck into a generational company.

ChatGPT’s visualization of a spiky founder

The great founders can turn early luck or unique insights into a transformative company. If a person has never done something outside the traditional path, it’s hard to underwrite them to becoming a person with maverick mentality who changes an industry. Listen to any Founders Podcast and you’ll be struck by how different these great people are from normal.

Other traits I personally like, some borrowed from The Gigascale Founder and Founders of Paradox:

• Technical, but also commercial

• Chip on their shoulder

• First-principles thinker

• This is their life’s work (hedgehog)

• Insanely big vision and ambition

• Relentless

• Adaptable learning machines

You may be wondering why I keep writing articles or refining my point-of-view about founders, but it’s important to remember to keep the main thing the main thing. And at seed especially,² the main thing is, and always has been, founders. Continually challenging your assumptions and fine-tuning what you’re looking for and how to develop taste in this area is the foundation of being a great investor.

Category-defining company

The best companies define a category and are usually the only one left standing a decade in.

That’s often hard to know or see at seed, so what can we look for? My framework is that there should be something about either 1) the genetics of the company or 2) the end market dynamics that lends itself to disproportionate economics or market share for the #1 player in the space.

Examples of things you may be able to squint and see an eye for at seed:

• Potential network effects³

• Proprietary technology

• Proprietary and compounding data moat that solves problems with big business value

• Innate switching costs, eg: system-of-record or database companies

• Very sticky customer base

• Economies of scale

Other business powers like brand are possible and extremely important long-term, but maybe harder to predict at seed.

12-36 months before a market map

Category-defining companies usually need to be early to the category⁴ in order to define it. But I would argue timing is even more important at seed.

Seed is a difficult place to invest for many reasons: 1) it can feel random 2) you’re competing with multistages who usually view seed rounds as option checks and blow up price or round size and 3) you often only see whispers of what the future market may look like.

For this reason, it’s often tempting to want to build out a market map⁵ and look for companies playing in existing known categories or ones that fill a logical whitespace. This is in part why the SF VC memeplex is so powerful:

While this may make sense at Series A or Series B when categories have started to become clearer and you’re often trying to pick the future category winner, I think this is a mistake at seed.

If it’s so well-understood by the Series A, but especially the Series B, market, that they’re making market maps about it, it’s likely too late to find a new winner in the space unless there’s something truly differentiated AND contrarian technology-wise about the company. The reason is that if it’s an exciting category that everyone’s looking at, multistage funds are probably going to beat or outbid you for any potentially new good companies that are popping up, even at seed.

For example, if you’re a seed investor and you’re trying to invest in voice AI agents right now, you’re probably NGMI.⁶

Two weeks ago, three different Tier-1 multistages put out their market maps for the voice AI agent space literally within a few days of each other and a couple days later, I spoke to a fourth Tier-1 venture fund that’s also building a thesis there.

So what do you have to do at seed? You have to go further out on the risk curve, to something that’s weirder, different, or too early for a later-stage investor. Early, but not too early: hence the “12-36 months” phrasing.⁷ If you go too early and either the technology or market is not ready, the company may run out of money before you hit an inflection point. Finding an idea whose time has finally come with a multi-decade tailwind driving your potential is critical to building a generational venture-backed company.

A good litmus test I’ve been using recently is: “If I asked someone who works at a middle-of-the-road Series B/C fund about this company, would they react with skepticism or excitement?” If it’s the latter, I worry that the market is already defined and obvious and it’s too late for me.

If it’s the former, there might actually be something there.

---

So there you have it. My new stock answer for “what do you invest in?” This can span sector, category, or layer, and hopefully gives a bit more color to the specific signature of founder, company, or market I’m looking for at seed.

---

---

Would love to hear feedback from you if you disagree or have more color to add on this topic. It’s important to note that is MY personal opinion and current investing framework, not necessarily the opinion of Susa. Part of what makes a great team is people with different points of view that sharpen one another. I’m often not right!

And of course, if you’re a spiky founder building a category-defining company 12-36 months before it would appear on a market map, get in touch at pratyush [at] susaventures [dot] com

1. While singularly unique domain expertise can make someone spike, I personally find the checkbox of “founder-market fit” overused by VCs. It’s often somewhat of a tautology with low signal-to-noise and thus somewhat useless as a framework: it’s very rare that a decent VC invests in someone with ZERO domain experience or understanding. And most of those companies still fail. Funnily enough, NOT being an expert or insider is sometimes helpful to upending or reinventing an industry like we saw with Palantir, SpaceX, Plaid, Tesla, Airbnb, Uber, Oscar, etc.

2. Important to note that whenever I talk to a great late-stage investor, they still say founders are the most important thing. Great founders are the ones who can make the company bigger than what an Excel model or market sizing exercise can predict.

3. Overused by VCs, but nonetheless there are some businesses that actually have it

4. Or invent it! (Eg: SpaceX, Anduril, OpenAI, etc.)

5. Or borrow Sequoia’s, which is what most people do.

6. Again, timing is everything. If you did a seed in some of the voice AI companies 6-12 months ago, you might be sitting very pretty as venture funds clamor over your companies.

7. I’m not wedded to exact timing here: cycles are moving even faster these days in the hyper-networked and competitive world of VC. The key is that it should not be the obvious thing that all the multistage funds are excited about right now.