It’s tempting to think that you can tell which strategy an investor is playing by the price that they pay. Fairway investors pay up. Rough investors and silver miners buy cheap. That’s the logical assumption.

But I don’t think that simple heuristic is quite right, and it betrays a misconception about how the early-stage market actually works.

Dead and Alive

When I think about seed valuations, I keep coming back to one of my favorite paradoxes in physics: Schrödinger’s cat. In the thought experiment, a cat is sealed in a box with a vial of poison that breaks if a radioactive atom happens to decay. Taken literally, quantum mechanics says that until the box is opened, the cat is in a superposition, neither definitely alive nor definitely dead. Schrödinger meant this as a jab at how absurd the theory sounds when applied to everyday things. But the image stuck because it captures something real: some situations can’t be resolved until they are observed.

Seed-stage companies are kind of like that cat. Every one of them is simultaneously massively overvalued and massively undervalued. It sounds absurd, but until the box opens, both are true.

These companies are overvalued because most of them fail. If you look at the returns of the venture industry overall, they’re pretty underwhelming, and the returns that do exist are carried by a small number of companies. That suggests that, on average, we’re all paying too much for all early-stage companies.

But every seed company is also massively undervalued.  This is because each one is an option on an enormous outcome. At seed, this is the lowest price anyone will ever pay for the asset. If a company turns into a mega-compounder, nobody looks back and thinks the seed price was too high. They feel like they massively underpaid for the privilege of being part of the journey.  In my own experience, every successful company felt terribly over-priced initially, and then stupidly underpriced in the end.

That’s the reality of early-stage investing. Until the box is opened, both things are true.  The cat is both alive and dead. The company is both massively over-valued and undervalued.

Early Stage Pricing is Oddly Efficient

In late 2025, AngelList published some data that supported something that we’ve intuitively seen around seed pricing. The data showed that the odds of an outlier seed outcome is about the same across the valuation spectrum, thus suggesting that pricing is rather efficient overall.  This is kind of hard to believe, but it makes logical sense. Lower priced companies are probably much riskier, but the cost basis is much, much lower. The most expensive companies probably are more valuable on average, but you are paying such a high price that you need that much larger of an exit to generate a venture scale return.

Here’s a summary of the data below – it looks at more than 6,000 first pre-seed and seed rounds from 2015 to 2021, sorted into valuation deciles within each year. The chart below shows the share of deals in each decile that have hit a 10x.

There is almost no pattern. The cheapest decile comes in around 5%. The most expensive is around 6%. The best deciles are the 5th and the 9th, at roughly 8%, and the worst is the 4th, at roughly 3%. That feels more like noise than an actual trend. Whether you pay a very high price or a very low one, neither end of the market is meaningfully more likely to produce an outlier.

Before anyone gets too worked up, yes, this data is very imperfect. A lot of these are paper marks, and the valuations of today are very different from the 2015-2021 period. But even still, I found this to be incredibly surprising, and it simply points to a market that has become more efficient at pricing early stage risk than I ever expected.

There is some evidence that the pricier companies are less likely to fail. In the same dataset, about 29% of bottom-decile deals are currently marked up, versus about 42% of top-decile deals.

This seems to be a modest trend, but in an upside-driven business, I think it’s small enough to mostly ignore. Seed funds don’t win by having fewer companies go sideways. They win by owning the few that become enormous, and the odds of that look similar no matter what you paid. The AngelList author landed in a similar place, describing consensus rounds as “priced almost correctly.”

Diamonds in the Rough Are Not Defined by Price

So back to the misconception. People tend to assume that searching for diamonds in the rough means paying a low price. I don’t think that’s true.

Companies get priced on what the market can see: how capable the founders are, how attractive the market is, how obviously large the company’s ambition is, and how much traction exists. That’s more or less the whole pricing model. And the market does a pretty good job taking this into account on average.

That means a diamond in the fairway doesn’t always come with a $100 million price tag. Some founders are obviously great and everyone loves them. But if they are super early in their careers and what they are building is very speculative with zero traction, they won’t command the premium of someone like Bret Taylor starting Sierra, but both are diamonds in the fairway, just priced at different levels.

Similarly, a diamond in the rough isn’t necessarily a cheap company. It’s a company that might have some attributes of obvious quality, but probably some that are total head scratchers. Put those together, and you might have something that is still priced at the median of seed or higher, but would be priced 2-3X that if everything was right down the middle of the fairway. The edge then comes from seeing what the price doesn’t capture, not from the price itself.

To put it differently, investing in diamonds in the rough doesn’t mean investing in the lower end of the valuation distribution, assuming that the relationship between price and quality is inefficient. It’s finding companies priced across the distribution, but ones that have much higher quality than their price would seem to suggest. These can live across the valuation curve, and are the companies that have significant potential that most people miss.

Two Approaches to Finding Diamonds in the Rough

If you accept that the market is mostly efficient, I think there are two ways to approach it.

The first is to be very strict on ownership targets. Usually, this is governed by fund size, which in turn implies a narrow range of check sizes and entry valuations. Say I want to write a $3 million check for 10%. By definition, I’m only going to look at companies raising at around a $30 million post. If the market is efficient, that’s a perfectly reasonable thing to do. Your odds of a 10x in that band are about the same as anywhere else. But you end up ignoring both ends of the market. You’ll obviously miss some great companies that just happen to be priced outside your box. But counter-intuitively, you’ll also get so used to this style of investing that you may be unable to build conviction about companies priced well below this band. This is why larger seed funds often struggle to get their heads around pre-seed opportunities. As a result, you’ll have a much smaller universe of companies to try to pick from, but hopefully better calibration for quality within your price range. It’s like staking out a very specific corner of the rough and just circling very thoroughly with a magnifying glass to find your diamonds.

The second approach is to invest across the whole spectrum, but size the check to fit your fund. If the odds of a 10x are roughly the same everywhere, then price isn’t what drives returns. What drives returns is whether you can get access to the right companies, whether you can pick the right ones, and whether your check size makes sense for your model. The good news is that you have a much larger pool of companies to evaluate. But it’s a tougher universe to cover and it’s harder to evaluate when a true diamond exists. This is like walking around a vast field looking for diamonds and doing so with one person and no magnifying glass.

I personally lean toward the second approach. The first feels a bit like the tail wagging the dog for no real purpose. The second runs into limitations of scale and judgement. But those seem potentially like problems that might be solvable with a handy little thing called artificial super intelligence. In either case, the main point is that diamonds in the rough can exist across the valuation spectrum. The job of the VC is to find and polish them and help them achieve their potential. If you do that, you’ll have a very happy living cat.