Rob Day | July 22, 2008 at 4:21 PM 5 Comments

The seed stage capital gap (or: Why I learned to stop worrying and love the national labs)

One thing entrepreneurs commonly get frustrated about when dealing with early stage VCs, particularly in cleantech, is when the VCs tell them “we like the idea, but it’s still too early.”

After all, it’s “early stage” venture capital, so how can an entrepreneur’s idea be “too early”?

What gets wrapped up in this, as well, is the fact that many entrepreneurs and researchers are working on some truly impressive ideas, but with long development paths ahead of them.  Fusion, geothermal, even cold fusion (yes, cold fusion) are areas where I myself have seen some really intriguing entrepreneurial efforts—efforts that, as an environmentalist, I could get excited about.  But as an investor, I cannot back at this stage.

There are two major reasons why venture capitalists can’t go into very early stage efforts very easily.

The first is structural (apologies in advance for oversimplifying a lot in this section, just trying to move the prose along…).  Venture capital funds are amongst the most illiquid of all asset classes LPs might put their money into—but they still want returns sooner or later.  Most VC funds are structured with a 10 year limit from time of launch.  But of course, that means you can’t invest even in year 1 with an expectation of 10 years to an exit, because what if it slipped a year or two?  The time from initial investment to exit typically has to be 5-7 years at MOST—preferably much less.

Factor into the equation that an exit is most likely only going to come once a company has significant and growing revenues, and not when the technology is simply brought to market, and very quickly the VC’s decision-making starts to be clear.  For example, at @Ventures, when we say “early stage cleantech venture capital”, we’re typically looking for companies that are zero to two years away from commercial revenue (and preferably on the shorter end of that range).

Now, structural challenges could be addressed with creative thinking.  But the reason for the 10 year fixed life of VC funds is not only because of the LPs’ needs for liquidity, but also because of the time value of money.  Discount rates (not that they’re often used in the industry, but still…) are really high for venture capital investments.  That reflects the high risks associated with launching any new business, along with the high expected returns of the asset class.  Let’s paint the picture with some numbers:

• Click here for a chart from a thesis showing IRR hurdle rates for venture capital investments at various stages—80% for seed, 60% for “start up”, 40-60% for “other early stage/ expansion.”
• Click here for a great column by long-time early stage IT investor Fred Wilson, who shows that he’s beaten the market with a 35% success rate.

Put those two together, and you’ve got a picture where a VC who knows that the technology in question should work (there’s rarely any “science risk” associated with an internet startup, after all, just market and execution risks) expects to achieve a minimum 40% IRR at least 35% of the time.

What I just explained is where the oft-mentioned “5x” (or in other words, “we look for investment opportunities that we think will at least grow to 5x our initial investment”) comes from in venture capital.  Because if you return at least 40% IRR over 5 years on a million dollars, you’ve turned it into $5+mm.  If you do that with the kinds of success rates Fred Wilson talks about (let’s say 35% 5x investments, 35% 1x investments, and 30% wipeouts, to vastly oversimplify), you will have returned 17% per annum, not including the management fees, etc.  (And, obviously, also not including follow-on investments, and assuming all initial investments are the same size, etc., etc., etc.).

Still with me?  Sorry, but it gets even worse from here…

The point is that if you achieve 40% IRRs on your wins, and your “wins” happen about a third of the time, then you’re delivering what your LPs are typically expecting when they put money into an early stage VC fund.  A lot of SWAGs in the above analysis, true, but I’m just trying to get a point across so…

It’s hard enough to achieve a 5x in five years.  But really, the investors are hoping for 10x, not 5x, to bake in some of the risks, etc.  So a 10x in five years would equal an IRR of almost 60%.

Looking at it that way, with such a high hurdle rate, it starts to be understandable why the holding period for institutional venture capital investments can’t be 10 years to be successful.  If you put $1mm into a new technology development effort expecting it to turn into $10mm in 10 years, you’re really targeting a 26% IRR, a big drop from the 60% IRR for 10x in 5 years.  If that’s your upside scenario, your overall portfolio results won’t look very good…

Or, to put it another way, if you want me to put $1mm into your seed-stage company with expectations that it’ll exit in 10 years, based on typical returns hurdles, I need to be expecting that it’ll be worth over $100mm at that time.  Not hoping, expecting.  And that’s even if my LPs don’t begin to hate me anyway, for locking up their capital for that long.

A nice little 100x, right?  We’ve witnessed a few of those in the history of venture capital (I’ve even had the privilege of working with an investor who sourced one…).  So you can hope for it to happen.  But it’s not an event you can credibly plan an investment in expectation of.  So clearly, the “time value of money” means that it’s very tough to achieve venture capital type returns when your investment holding period is going to be more than 7 years or so.

In cleantech, it can be even harder to justify going that early.  Because there can be significant science and development risk still associated with some of these technologies at that stage.  We’ve already seen some high-profile investments taking seven years (and counting) to achieve commercialization, much less an exit.  Quite often, if you’re looking at that long of a holding period, you’re looking at some kind of R&D effort on the front end, and that adds additional risk.*

Let’s take an example—cold fusion, which if it works would be a great thing with huge market potential.  InTrade (which as a trained economist, I love) says that the most high profile current cold fusion effort (Dr. Arata’s, linked to from above) has a 15% chance of being replicable—of “working” in a lab.  Now, even if it’s replicable, then it would take a long time to commercialize, because there’s a big difference between finding excess heat and generating power, not to mention designing the equipment, building the devices at cost-effective levels, etc.  The last fusion-related plan I looked at was talking about 10 years to commercialization (and that was the entrepreneur talking, who’s probably optimistic), from a much more progressed development stage than Dr. Arata’s lab experiments.

So if I need to believe a $1mm investment in something at that stage would turn into over $100mm by the end of 10 years… but then I also have to discount further to account for the 85% chance the thing just plain doesn’t work… I need to believe my $1mm will turn into AT LEAST $733mm over those 10 years, to justify the investment.

Why did we just walk through that boring hypothetical numerical journey?  What I’m trying to illustrate is why VCs can’t be counted on to consistently go out and find the ideas that will take longer than 3-5 years to commercialize, and give them the financial resources to come to market.  (And, in fact, the emergence of more dedicated cleantech “growth stage” funds illustrates how VCs are moving even more toward shorter holding period strategies these days—but I digress).

Early stage cleantech venture investors, to be effective, need to be disciplined and pragmatic…

There are seed stage funds who specialize in going in early, but as the above hypothetical exercise illustrates, they’re going to have to be ruthlessly selective, and thus a lot of promising technologies won’t attract their capital.

There are angels who might be willing to use their individual bank accounts to back these kinds of efforts, with more patience or at least lower returns expectations.  But that’s even more hit-and-miss than the seed stage industry.

By and large, what I’ve described above is exactly why the government plays a vital role in sponsoring early stage research, when it’s in a sector (such as renewable energy) where we care about it a lot as a society.  Because there’s an inevitable capital gap between where the wild-eyed scientist has a brilliant but long-path-to-commercialization idea, and the stage where the institutional VCs can step in to help them build a company.

Whether at the national or the regional level, we need to see strong government support to fill that gap.

*Regular readers will note that I have previously argued against those who say that cleantech necessarily involves more capital and longer development periods than other sectors.  And my arguments on this point are still true.  There are plenty of investment areas across the widely varied cleantech sectors that will resemble the relatively quick development/ commercialization paths of IT, software, internet, services, etc.  And even the longer-development challenges in other cleantech sectors are most often reminiscent of similar challenges in semiconductors, biotech and other popular VC sectors.  So please don’t take this column for anything more than a hypothetical exercise talking about the difficulties in backing visionary, breakthrough technology development efforts in ANY market—cleantech or otherwise.