That's No Seed
How to understand the monster $1B "seed rounds" in AI.
Recently with the AI Summer, some startups are getting off the ground with $100M to $1B seed rounds. There is something odd about these rounds, and I want to explore why.
First, I don’t have a problem with the rounds themselves. If you are a generational talent, or at least can convince investors you might be, go for it. There are challenges ahead when you promise the world, but having a bigger war chest is obviously good.
But won’t someone think about the poor VCs and what they’ll need to win?
The names we give for funding rounds are actually just marketing. This is true for more down to earth rounds, especially as we add stages like preseed, seed, seed plus, pre-A, and so on. I advise founders to consider the material questions of a round:
How much will you raise?
Is it priced? On what terms?
Do your investors expect follow-on? How much? Do they have that money?
Is there one lead, or a party?
Will anyone join the board? Who controls it?
What milestones did you declare for the next round?
How will you spend the money?
None of these questions is subjective and squishy like the name we give a round.
It seems inappropriate to use the word “seed” just because a round happens to be the first.
What are the economics for the VCs?
If you’re a seed investor, you expect some of your companies will fail and others will win big. Later stage investing expects far smaller multiples. The benefit of late-stage investing, even with the smaller multiple, is that it’s a multiple on a much larger basis. It’s relatively easy to have high multiples on lower dollar amounts. This might not be intuitive because most people don’t have a problem of too much money to invest, and Brewster’s Millions is a reference you’re too young to understand.
If your seed round is $1B on say $4B post money, investors are buying 25% of the company. What kind of exit would you need to match the earlier multiples typical of seed? This is a good lesson for founders to understand the basic math of these rounds.
My first angel investment was in Cruise Automation, which returned 62X in 3 years with a $1B exit. Tango.vc’s current position in Lovable is worth at least 185X, and even more if a round were to be priced today. These returns happen when you invest at tens of millions and exit at one to ten billion. Some example math:
(ownership of $100K invested / $10M value) * 40% after dilution * $5B exit / $100K original = 200X
Now do some basic algebra (if you’re a VC, bear with me) to find what value you’d need on a $1B to get this multiple.
200X multiple * $1B original / (diluted ownership 70% * ($1B / $4B)) = $1.14 Trillion
Now it’s possible for a company to be worth $1T. But that target is higher than Berkshire Hathaway or Eli Lilly or JPMorgan. It’s approximately in the top 12 in the world. It’s not impossible, just almost certainly not going to happen. That $5B exit in the original example would be among the top 1,000 companies.
But the VCs are going to be fine, because they aren’t seed investors expecting such a multiple. If you’re deploying hundreds of millions in a single round, you are traditionally happy with a 3-5X. For example, that was roughly the expectation of Google’ venture arm CapitalG investing $150M into Stripe at $9.6B in 2016. The 10X value at Stripe’s 2023 round makes a stellar multiple for the late-stage category.
What would happen if an actual seed investor got involved in these rounds? If your check size remains static, it would be very difficult for the investment to matter. If some of your multiples are 50X to 200X, that 10X won’t be good in comparison. The company could be on an epic path and go public, but that return wouldn’t affect your results. This is the tyranny of powerlaws in venture.
For founders raising these rounds, once again, I wish you luck. Take the money, go build something awesome. You might think “who cares about my investor’s multiple? That’s their problem.” But go back to a material question on the round: board control. You might think you could sell for $3B and make yourself generational wealth. But your investors probably won’t be happy, and if they’re on your board they can force you to keep going.
There are many stories about this behavior, but you’ll rarely see it in press releases. Most of the lore is stuck within VC and founder circles who have no incentive to play it out publicly because they want access to future deals.
Sometimes the mask cracks, like when A16Z’s Marc Andreessen tagged a few outmaneuvered founders after a partner at Benchmark said their strategy is “all based on the founder”. The comment from Benchmark is not entirely true, and the incentive for Marc is to publicly call it out because A16Z directly competes with Benchmark. You’ll see the same with Founders Fund calling out other firms, with the name including “Founders” to highlight whom they favor. This is like Drake vs Kendrick for people that spend too much time on Twitter, a delicious blood sport.
