Vinod Khosla at Milken Institute: Part I
Thanks to a reader for bringing this to my attention. Vinod Khosla (VK) just did a lengthy interview at the Milken Institute 2009 Global Conference. The interview was conducted by Elizabeth Corcoran (EC) of Forbes. You can see the video of the interview here:
I am going to listen to the interview, transcribe it, and comment on interesting/controversial exchanges. (If a transcript of this interview exists already, I haven’t seen it). I will strive to be as accurate as possible, but may paraphrase lengthy questions and answers. I will also try to provide links to all of the companies VK mentions. Any comments I make will be preceded by “RR” and will be in italics. I will also note the time into the video of each question so you can listen for yourself if you like.
In this first installment, VK discusses the role of government money, argues that these businesses need not be capital intensive if you make someone else will pay the capital (he explains his low-capital strategy, in which he has managed to outsmart everyone else in the energy business), and then discusses some of his solar investments.
First question from Elizabeth Corcoran (1:40 into the video): (RR: She first mentions that VK “hates the Prius” and promises to get into that later). The essence of this green revolution is built around a very interesting tension of private and government money. Whose money is really going to transform energy and transform this industry?
VK: I don’t think it’s built on government money. The only thing it can be built on is private money. The scale of the energy problem is so large that we cannot solve it with government money; there is not enough government money available. The only way we solve the energy problem and the climate problem is with technologies that achieve unsubsidized market competitiveness. There is not enough money in the world to subsidize oil replacement or coal replacement, or replacing automobiles. Besides, most of the most interesting markets in the world, the fast growing markets – India and China – don’t have these subsidies. The fact that we have government money maybe compensates for the fact that its competitors, fossil fuels, have had lots of subsidies too – and maybe continue to get larger subsidies than renewables get today. But it doesn’t matter. All those things will go away; will disappear at scale. So I challenge your premise to begin with.
EC (3:45): At the same time, let’s talk about how much money you have put into these technologies, and how much money they still need to achieve any kind of viable production scale. You have invested in quite a number of companies, and some of them are doing well; some of them are not doing so well. How much money does it take at this point to even conceive of having a production level cellulosic ethanol plant?
VK: I love these questions, because I can challenge her every single time. I disagree with the basic premise that it takes lots of money. If you look across our portfolio, 80% of our portfolio doesn’t need any more money that the typical venture start-up. (EC: Which is how much?) Somewhere between $30 million and $100 million to get going, whether it’s a chip start-up, semiconductor start-up, enterprise software start-up, a computer systems start-up – it’s about that range. These are very typical. Most of them are not as expensive as biotechnology which end up needing a billion dollars. And it’s not as cheap as a Web 2.0 start-up where two 23-year olds start something. But they are right smack in the middle of what venture capital has been doing. Among the remaining 20%, at least 15 of the 20 have strategies that make it very capital efficient; that means they don’t need much capital to build a plant; to be in the fuels business. And I can explain how. Then there’s a few that are very venture capital intensive. So yes, there’s a few, but that’s also true in the traditional venture business. So I don’t think there’s a difference, and this is a very common misperception.
EC (5:40): Just to cite one; perhaps one of the few, but I couldn’t help it – you have the CEO of Coskata saying it is going to take $400 million to get that plant up to production level.
VK: The question is, does Coskata need to build their own plants. (EC: They want to). They may want to – and they will – if the capital is available. If not, what do they do? They are using a Westinghouse gasifier. They let Westinghouse build the gasifier. They will just build the fermentation tank. All I am saying is strategies exist for people who are doing that kind of thing to have relatively low-capital outcomes.
(RR: I don’t understand this answer at all. Just because Westinghouse is building the gasifier doesn’t mean Coskata isn’t paying for it. It is a part of the capital cost of building their plant).
EC (6:40): These companies have said that unless they are producing on the order of 10 million gallons of fuel a year they aren’t really viable.
VK: Sure. And what do you think they will say when they ask for government money at low interest rates? Tell them “No, we don’t need the capital; we have a low cost strategy?” Look, to build demonstration plants you absolutely need money. What I am saying is most of these companies have strategies where – if the markets are great they absolutely can supercharge their growth with high amounts of capital – or they can take a lower capital strategy. You have seen that in biotechnology too. Biotech companies license some drugs and keep others for themselves. Exactly the same strategy is possible here.
(RR: Presumes you have something that has value for a licensee. A very heavy capital intensive venture with a marginal energy return isn’t going to be any more attractive to build for a licensee than for the company trying to license the technology).
EC (7:40): One more question on the capital side. The rumor in Silicon Valley is that you have been out pounding the doors looking for a billion dollar fund. Are you doing that?
VK: Actually, I can’t comment on that.
EC (7:53): And the rumor has been that you are not going to get there, because investors are scared. They are worried that they are not going to make a return when you have got companies that are doing a Series B round that is plus $100 million.
VK: Well, we will see what we want to do and where we get to. (EC: That’s hardly an answer). Let me put it this way. I don’t think I have attempted something that I have been unable to do in the past, and I don’t expect it to be different in the future. (RR: Big grin toward the audience).
(RR: I do expect it to be different, because I think VK has vastly underestimated the level of difficulty here. He believes his experience from the computer industry is applicable to technologies which in some cases have already been around for almost 100 years. But this isn’t the computer business and Moore’s Law is not in effect. As Geoffrey Styles argued in his latest column “If there is a Moore’s Law for energy, it has yet to be discerned, let alone quantified. In the early phases of any new technology, “experience curve” effects can emulate Moore’s Law-style improvements for a while. Then, as cumulative output grows the rate of change slows dramatically.”)
EC (8:25): Let’s go back to some of the specific technologies. You have made this teasing comment now that there are companies out there that are smaller – modest sized investments – that presumably you believe will have a big impact. Let’s talk about a couple of those. What are the ones that to your eye look the most promising, starting at the small level and going back up to the big guys.
VK: Let’s pick a company. We have a little company in our portfolio called PVT. (RR: See VK’s portfolio here). They can sit behind Sunpower solar panels or First Solar solar panels and triple the energy efficiency of the panel – without changing the panel – by utilizing the waste heat from the panel. My bet is that they will look more like a Web 2.0 start-up. For $15 million they will be a profitable company. That’s pretty rare in the venture business.
A lighting company – somebody like Soraa (RR: I can’t find a website for them) – LED lighting, massive market. Probably could be in the market for relatively small amounts of money. Pick a number – $25 million or less. So let me challenge you and tell you; in Cleantech, the phenomenon is, for most opportunities, the capital investment is the same as traditional venture capital, but the markets are 10 times larger and we have half the competition. Now, wouldn’t you prefer those markets? (RR: Sounds like an advertisement for the billion dollar fund he couldn’t talk about earlier).
(RR: VK smiles and looks at the audience, EC starts to ask another question, and then VK starts in again). How many markets do you see in the chip business, in the semiconductor business, or enterprise software, where the smaller end of the market is $5-$10 billion like in lighting? (EC: Not too many). Very few. Most enterprise software companies, or wifi companies – are going after markets that are hundreds of millions, maybe a billion dollars. Yet they need the same amount of capital. Guess which one is the better market to pick?
(RR: Depends on the margins, doesn’t it? Refining is a >$100 billion business in the U.S., but the margins are often very poor. This is partially what’s got so many ethanol producers in a bind. The energy markets are cyclical, and can suffer poor margins for extended periods. It doesn’t really matter how big the market is if margins in that market tend to be poor. Likewise, if the market is large but the barriers to entry are low – also a problem for the corn ethanol industry – then it will be hard to keep margins high before competitors show up in force.)
EC (10:33): Let’s talk about solar for a minute. You have a couple of solar investments. PVT you mentioned; Ausra.
VK: Yes. We have three solar companies. PVT is one I mentioned. Stion is another great example because everyone thinks photovoltaics are very capital-intensive. Yes and no. If you do it the dumb way, they are capital intensive. If you start manufacturing and redesigning your manufacturing line equipment, it will be. We chose to do two things differently. One, we are not touching the manufacturing process, so we can use other people’s equipment. For $6 million we were able to set up a pilot line because we bought 15-year-old equipment. We only innovated in the materials. This dramatically reduces the capital-intensity.
(RR: Seems to me that the potential flaw in the thinking that they will force others to do the capital-intensive stuff is the presumption that they will actually do it. If some of these technologies are so capital-intensive that VK won’t touch them, why would anyone else sink lots of money into them over the long-term?)
The second thing we did is said: There’s plenty of people in Nanosolar, MiaSolé and Konarka going after First Solar. Among the photovolaic market is Sunpower; high efficiency and high cost; First Solar; low efficiency and low cost; all of the start-ups competing with First Solar. Nobody wants to be high cost, low efficiency. So there is another quadrant, which is high efficiency, low cost. We decided to go after that. (RR: Wow, why didn’t someone else think of that. Instead of paying high cost for high efficiency, pay low cost for high efficiency. It has the sheer genius of 7-minute abs). And frankly, for $25 million dollars we have a pilot line running. We have very little risk, with very little capital. That gives us lots of options. Now you can say that photovoltaics are expensive, but you can pick the right strategy if you are knowledgeable about the trade-out you are making; if you make a change in manufacturing equipment you are going to have an expensive start-up.
Solar thermal is our 3rd start-up. Here is another great example of capital efficiency. They make these solar panels; these big mirrors that take concentrated light and turn it into steam. Building a power plant is $600 million, very capital intensive. That is an option for them. Last fall we decided the market wasn’t going to be good for project financing. They can sell $10-$15 million worth of gear, add two lines of solar to an existing coal plant, make it a little more green. And that’s what they are doing. They are doing equipment sales in $10-$15 million chunks to add to existing industrial processes where they don’t need power; or to existing power plants; whether they are coal-fired plants or natural gas plants. In small chunks, this becomes much lower capital-intensive. Looks just like Sun Microsystems manufacturing.
EC (13:40): (RR: Will pick up here in the next installment).