December 2008

*(I originally wrote this at the request of a company producing a report about entrepreneurship.

Unfortunately after reading it they decided it was too controversial to include.)*

VC funding will probably dry up somewhat during the present recession, like it usually does in bad times.

But this time the result may be different.

This time the number of new startups may not decrease.

And that could be dangerous for VCs.

When VC funding dried up after the Internet Bubble, startups dried up too.

There were not a lot of new startups being founded in 2003.

But startups aren't tied to VC the way they were 10 years ago.

It's now possible for VCs and startups to diverge.

And if they do, they may not reconverge once the economy gets better.

(I wrote this for a report on entrepreneurship; they found it too controversial to include.)

VC funding will probably dry up in this recession. But this time the number of new startups may not, and that could be dangerous for VCs. After the Bubble, when VC dried up, startups dried up too. But they're no longer tied to VC the way they were. The two can now diverge.

VC funding will dry up in this recession, but this time the number of startups may not fall with it. Startups are no longer tied to VC, so the two can diverge and may never reconverge.

The reason startups no longer depend so much on VCs is one that everyone in the startup business knows by now: it has gotten much cheaper to start a startup.

There are four main reasons: Moore's law has made hardware cheap; open source has made software free; the web has made marketing and distribution free; and more powerful programming languages mean development teams can be smaller.

These changes have pushed the cost of starting a startup down into the noise.

In a lot of startups—probaby most startups funded by Y Combinator—the biggest expense is simply the founders' living expenses.

We've had startups that were profitable on revenues of $3000 a month.

Startups depend less on VCs because it's gotten cheaper to start one: cheap hardware, free software, free distribution, smaller teams. The biggest expense is now the founders' living costs. We've had startups profitable on $3000 a month.

Startups no longer depend on VCs because it has gotten much cheaper to start one. The biggest expense is now just the founders' living costs.

$3000 is insignificant as revenues go.

Why should anyone care about a startup making $3000 a month?

Because, although insignificant as revenue, this amount of money can change a startup's funding situation completely.

Someone running a startup is always calculating in the back of their mind how much "runway" they have—how long they have till the money in the bank runs out and they either have to be profitable, raise more money, or go out of business.

Once you cross the threshold of profitability, however low, your runway becomes infinite.

It's a qualitative change, like the stars turning into lines and disappearing when the Enterprise accelerates to warp speed.

Once you're profitable you don't need investors' money.

And because Internet startups have become so cheap to run, the threshold of profitability can be trivially low.

Which means many Internet startups don't need VC-scale investments anymore.

For many startups, VC funding has, in the language of VCs, gone from a must-have to a nice-to-have.

$3000 is insignificant as revenue. So why care? Because it can change a startup's funding situation completely.

A startup always tracks its "runway"—how long till the money runs out. Cross into profitability, however low, and runway becomes infinite—a qualitative change, like the stars turning to lines when the Enterprise jumps to warp.

Because Internet startups are so cheap to run, that threshold can be trivially low. For many, VC has gone from must-have to nice-to-have.

$3000 a month is nothing as revenue, but crossing into profitability—however low—makes runway infinite. Once profitable you don't need investors, so for many startups VC has gone from must-have to nice-to-have.

This change happened while no one was looking, and its effects have been largely masked so far.

It was during the trough after the Internet Bubble that it became trivially cheap to start a startup, but few realized it because startups were so out of fashion.

When startups came back into fashion, around 2005, investors were starting to write checks again.

And while founders may not have needed VC money the way they used to, they were willing to take it if offered—partly because there was a tradition of startups taking VC money, and partly because startups, like dogs, tend to eat when given the opportunity.

As long as VCs were writing checks, founders were never forced to explore the limits of how little they needed them.

There were a few startups who hit these limits accidentally because of their unusual circumstances—most famously 37signals, which hit the limit because they crossed into startup land from the other direction: they started as a consulting firm, so they had revenue before they had a product.

This change happened while no one was looking. Starting a startup got cheap during the post-Bubble trough, but few noticed, startups being out of fashion. When they returned around 2005, investors wrote checks.

Founders no longer needed the money but took it if offered—partly tradition, partly because startups, like dogs, tend to eat when given the opportunity. So they never explored how little they needed VCs. A few hit it by accident—most famously 37signals, consultants who had revenue before a product.

This shift happened while no one was looking, masked because VCs kept writing checks and founders kept taking the money. As long as checks flowed, founders never had to find out how little they needed VCs.

VCs and founders are like two components that used to be bolted together.

Around 2000 the bolt was removed.

Because the components have so far been subjected to the same forces, they still seem to be joined together, but really one is just resting on the other.

A sharp impact would make them fly apart.

And the present recession could be that impact.

VCs and founders are like two components that used to be bolted together. Around 2000 the bolt was removed. They still seem joined, but one is just resting on the other. A sharp impact would make them fly apart—and this recession could be it.

VCs and founders are like two components that used to be bolted together. Around 2000 the bolt was removed; a sharp impact would make them fly apart, and this recession could be it.

Because of Y Combinator's position at the extreme end of the spectrum, we'd be the first to see signs of a separation between founders and investors, and we are in fact seeing it.

For example, though the stock market crash does seem to have made investors more cautious, it doesn't seem to have had any effect on the number of people who want to start startups.

We take applications for funding every 6 months.

Applications for the current funding cycle closed on October 17, well after the markets tanked, and even so we got a record number, up 40% from the same cycle a year before.

Maybe things will be different a year from now, if the economy continues to get worse, but so far there is zero slackening of interest among potential founders.

That's different from the way things felt in 2001.

Then there was a widespread feeling among potential founders that startups were over, and that one should just go to grad school.

That isn't happening this time, and part of the reason is that even in a bad economy it's not that hard to build something that makes $3000 a month.

If investors stop writing checks, who cares?

At the extreme end of the spectrum, YC would see the separation first, and we do. The crash made investors cautious but hasn't touched how many want to start startups. Applications closed after the markets tanked, yet hit a record—up 40%.

So far there's zero slackening of interest. That's unlike 2001, when founders felt startups were over and one should just go to grad school. Now $3000 a month isn't hard, even in a bad economy. If investors stop writing checks, who cares?

At the extreme end of the spectrum, YC would see a separation first, and does. The crash made investors cautious but didn't dent founder interest—applications hit a record, up 40%—unlike 2001.

We also see signs of a divergence between founders and investors in the attitudes of existing startups we've funded.

I was talking to one recently that had a round fall through at the last minute over the sort of trifle that breaks deals when investors feel they have the upper hand—over an uncertainty about whether the founders had correctly filed their 83(b) forms, if you can believe that.

And yet this startup is obviously going to succeed: their traffic and revenue graphs look like a jet taking off.

So I asked them if they wanted me to introduce them to more investors.

To my surprise, they said no—that they'd just spent four months dealing with investors, and they were actually a lot happier now that they didn't have to.

There was a friend they wanted to hire with the investor money, and now they'd have to postpone that.

But otherwise they felt they had enough in the bank to make it to profitability.

To make sure, they were moving to a cheaper apartment.

And in this economy I bet they got a good deal on it.

I've detected this "investors aren't worth the trouble" vibe from several YC founders I've talked to recently.

At least one startup from the most recent (summer) cycle may not even raise angel money, let alone VC. Ticketstumbler made it to profitability on Y Combinator's $15,000 investment and they hope not to need more.

This surprised even us.

Although YC is based on the idea of it being cheap to start a startup, we never anticipated that founders would grow successful startups on nothing more than YC funding.

We see the same in startups we've funded. One had a round fall through over a trifle—their 83(b) forms—yet is obviously going to succeed: their graphs look like a jet taking off. When I offered more introductions they said no; four months of investors had left them happier.

They had enough to reach profitability, and to be sure were moving to a cheaper apartment. I've caught this "investors aren't worth the trouble" vibe from several founders lately. Ticketstumbler reached profitability on YC's $15,000—surprising even us.

Existing YC startups show the same divergence. One whose graphs look like a jet taking off declined more intros, happier without investors. Ticketstumbler reached profitability on YC's $15,000 alone—surprising even us.

If founders decide VCs aren't worth the trouble, that could be bad for VCs.

When the economy bounces back in a few years and they're ready to write checks again, they may find that founders have moved on.

There is a founder community just as there's a VC community.

They all know one another, and techniques spread rapidly between them.

If one tries a new programming language or a new hosting provider and gets good results, 6 months later half of them are using it.

And the same is true for funding.

The current generation of founders want to raise money from VCs, and Sequoia specifically, because Larry and Sergey took money from VCs, and Sequoia specifically.

Imagine what it would do to the VC business if the next hot company didn't take VC at all.

If founders decide VCs aren't worth the trouble, that's bad for VCs. When they're ready to write checks again, founders may have moved on.

There's a founder community as there's a VC community, and techniques spread fast: try something that works, and in 6 months half use it. The same holds for funding—founders want Sequoia because Larry and Sergey took money from Sequoia. Imagine what it would do if the next hot company took no VC.

If founders decide VCs aren't worth the trouble, by the time checks flow again founders may have moved on. Techniques spread fast through the founder community—imagine if the next hot company took no VC at all.

VCs think they're playing a zero sum game.

In fact, it's not even that.

If you lose a deal to Benchmark, you lose that deal, but VC as an industry still wins.

If you lose a deal to None, all VCs lose.

VCs think they're playing a zero sum game. In fact it's not even that. Lose a deal to Benchmark and you lose that deal, but VC still wins. Lose a deal to None, and all VCs lose.

VCs think they're playing a zero-sum game; it's not even that. Lose a deal to Benchmark and VC still wins—but lose it to None, and all VCs lose.

This recession may be different from the one after the Internet Bubble.

This time founders may keep starting startups.

And if they do, VCs will have to keep writing checks, or they could become irrelevant.

This recession may differ from the last. This time founders may keep starting startups—and if they do, VCs must keep writing checks, or become irrelevant.

This recession may differ from the last: founders may keep starting startups regardless. If they do, VCs will have to keep writing checks or risk becoming irrelevant.

Thanks to Sam Altman, Trevor Blackwell, David Hornik, Jessica Livingston, Robert Morris, and Fred Wilson for reading drafts of this.

Thanks to Sam Altman, Trevor Blackwell, David Hornik, Jessica Livingston, Robert Morris, and Fred Wilson for reading drafts of this.

Thanks to the readers of drafts.