Whenever we talk to investors about backing Y Combinator companies before Demo Day, the same objection comes up first: the valuations are too high. A few years ago you could get into a YC company at a couple million dollars. Now the best ones are priced several times higher before they even pitch. Doesn't that kill the returns?
It's the right question to ask. It's also a question with a precise, quantitative answer — and the answer is more encouraging than most people expect. The math of venture returns isn't about entry price in isolation. It's about entry price relative to outcome, multiplied across a portfolio governed by the power law. Here's how to think about it.
Start With the YC Benchmark
Y Combinator itself invests at roughly a $2M valuation and has historically returned on the order of 150–160x on that equity across its program. That's the reference point. It's a publicly discussed figure and it anchors everything that follows.
Now suppose you invest in the same companies, but later — after the batch has started, when traction is visible and the company is priced higher. Say you enter at 10 to 15 times YC's entry valuation. The intuition most people have is that paying 10–15x more destroys the return. But that's not quite how the arithmetic works.
If YC earns ~160x at a $2M entry, and you enter at 10–15x that price on the same underlying companies, your expected multiple is approximately 160x divided by your entry premium — roughly 10–16x. That is still a venture-scale fund return. The entry premium compresses the multiple; it does not eliminate it. You are paying more for materially more information and materially less risk.
The Power Law Does the Heavy Lifting
The entry-premium math gives you the headline. The power law tells you where the return actually comes from — and it's never the average company. It's the extreme tail.
Walk the funnel. Start with a portfolio of, say, 200 seed-stage YC companies. A meaningful fraction raise a Series A. A fraction of those reach Series B, then Series C. By the time you get to the end of that funnel, a small number — on the order of 10% of the portfolio — reach unicorn status. That's the power law in action: most of the portfolio returns little, and a handful return almost everything.
Here's the part people underestimate: the median YC unicorn doesn't exit at $1B. It overshoots — the median outcome for companies that reach unicorn status is closer to $3B, and a meaningful share become decacorns ($10B+). When you run a $3B exit against a seed-stage entry, a single position can return many multiples of the entire fund. You don't need most of the portfolio to work. You need a few of the right ones.
This is why portfolio construction in early-stage venture is fundamentally different from picking stocks. You are not trying to be right on average. You are buying a broad enough basket of credible power-law candidates that you are statistically likely to hold one or two of the extreme winners — and those winners pay for everything, including the companies that don't make it.
Why the J-Curve Is Outrunning the Price
Now back to the valuation objection. Yes, entry prices have risen. But three things have changed alongside them, and together they more than offset the higher entry.
First, the J-curve is faster. The time from founding to a Series A has compressed dramatically for top performers — from a year and a half or two years down to roughly nine to twelve months. Companies are reaching $1M in annual recurring revenue faster than ever. A higher entry valuation that reflects genuinely faster value creation isn't inflation — it's the market correctly pricing a steeper growth curve.
Second, larger checks buy pro-rata rights. Writing a slightly larger check early can secure the right to invest again in later rounds. That lets an investor concentrate more capital into the companies that are clearly working — maintaining ownership through the exact value inflection where the power-law winners separate from the pack. The follow-on dollar is often the highest-conviction, best-informed dollar in the whole portfolio.
Third, the AGI resiliency test. We now stress-test every company against one question: when far more capable AI arrives, does this company's value increase or evaporate? It's a filter that pushes a portfolio toward defensible positions — companies whose moats deepen as AI improves rather than businesses that a better model simply replaces. In a world where capabilities are advancing this fast, durability of the moat matters as much as the entry price.
What This Means in Practice
Put it together and the picture is straightforward. The entry-premium math says a disciplined later-but-still-early entry can still target a venture-scale multiple. The power law says a few extreme winners — not the average company — drive that return, so you build a broad, credible basket. And the faster J-curve plus pro-rata rights plus a moat-durability filter mean today's higher entry prices are buying faster-growing, better-understood, more defensible companies.
Entry price matters. It just isn't the whole story — and treating it as the whole story is the single most common mistake we see investors make when they evaluate YC-focused strategies. The discipline that actually drives returns is getting in early enough to capture the power-law winners before they're obvious, on companies durable enough to survive the next wave of AI, with the structure to keep buying the ones that work.
That, in one paragraph, is the entire case for investing before Demo Day rather than at it.
