Tumblr Icon RSS Icon

VC Khosla blows his credibility dissing plug-ins

By Joe Romm on December 26, 2007 at 9:54 am

"VC Khosla blows his credibility dissing plug-ins"


google plus icon

He may be a “venture-capital star” who is now putting a lot of money into biofuels — but he is no clean tech expert, as he proved during a keynote address at ThinkEquity Partners’ ThinkGreen conference in San Francisco. In remarks that should worry anybody relying on his judgment, Khosla said:

Forget plug-ins. They are nice toys. But they will not be material to climate change.

Very, very wrong. Plug ins are likely to be a central strategy for dealing with climate change, as readers of Climate Progress know (see below). I hope cellulosic ethanol will be, but that still remains to be seen.

Contrary to what Khosla says, affordable cellulosic ethanol probably requires a major breakthrough — and to play a major role in climate change, it needs a major new infrastructure investment. I recently test-drove a prototype plug in that avoids the need for a battery breakthrough (more on that after the Detroit auto show). And plug ins don’t need a major new infrastructure investment. They are not toys.

Khosla should stick with what he knows.

Related Posts:

The conservative war on Christmas

The Best of Climate Progress ” Spring 2007

27 Responses to VC Khosla blows his credibility dissing plug-ins

  1. jcwinnie says:

    I happen to believe that he is incorrect, also. Still, I am neither a clean tech expert nor venture-capital star.

    Khosla would seem to know how to make money, so at present, plug-in hybrid may fail his profit picture.

    Khosla goes with status quo, e.g., willing to endorse corn-to-ethanol, even when less than a good idea.

    And, he has good at thinking out things. So, while plug-in hybrids may be unprofitable at present and he beats the drum for ethanol as an alternative transportation fuel, if the battery picture changes, then you still want a clean efficient range extender.

    Volvo has buses that use ethanol in a compression ignition (read diesel) engine rather than a spark ignition engine. So, an abundance of ethanol available and usable means less gasoline, which means less GHG and less dependence on foreign oil. All good things, as far as I can tell.

    MinMax, Maxie!

  2. Ron says:

    You mean here’s a guy investing private money, not relying on taxation and hand-outs for his vision?

    I applaud him.

  3. copper potts says:

    does this guy know that PHEVs use gasoline too? this is a comment you’d make about electric cars.

    “In particular, he pointed to battery problems.

    Some analysts and companies have said current lead-acid and nickel-metal hydride batteries can’t store enough energy for their size to drive far enough — on electricity alone — for mainstream drivers.”

    I saw this article when I was surfing the internet reading about PHEVs.

    “Malcolm Bricklin is a wild card, a high-flyer and
    a crash-lander. Now his vision is about PHEVs.
    The stories from USA Today and Reuters give the
    facts — and say he’s already working on
    prototypes for a series PHEV that could cost


  4. Michael says:

    Khosla is trying to defend what might be very risky investments. It is unfortunate that his tactic is to spread FUD, (Fear Uncertainty and Doubt) but it is an old Silicon Valley tradition. In this case, I think he doesn’t understand the ethical and long-term practical dimension of talking down a perfectly good potentially climate-saving product that may eventually complement his investment (flex fuel plug ins). It is lucky that his investment in Ausra is on more solid ground and doesn’t require this type of public stupidity.

  5. Earl Killian says:

    To illustrate Khosla’s mistake on plug-ins, consider the land area required to replace gasoline in the US with either biofuels or electricity:

    Stirling Energy solar electricity 3,607 mi^2
    Algae biodiesel (theoretical) 24,081 mi^2
    Wind electricity 29,696 mi^2 (but land may be used for another purpose as well)
    Algae biodiesel (current yield) 149,324 mi^2
    Switchgrass ethanol 224,794 mi^2
    Corn ethanol 438,061 mi^2
    (raw data mostly from Wikipedia, combined into a spreadsheet)

    For me, the implications of the above calculations are pretty clear. The best biofuel (algae biodiesel) takes 7 to 41 times as much land as using mirrors to generate electricity. Cellulosic ethanol is 62 times the land area, and corn ethanol (which Khosla has advocated in the past) is 121 times the land area. Also, mirrors don’t compete with food (they like to go in deserts, which aren’t the choicest cropland). Which would you choose?

    Note also that cars are in use only 5-10% of the day, and parked 90-95%. They can charge anytime power is available in that 90-95%, so they have tremendous flexibility to take advantage of energy that is available when the wind blows and the sun shines.

    People always put-down plug-ins saying that the battery technology isn’t mature yet. That doesn’t pass muster with me. Toyota designed and built a battery vehicle in the 1990s. The leased them to fleets (e.g. at Southern California Edison), where they built a solid track record. In 2002 they offered them for sale to the general public, and sold every one they built. My wife bought one of those, and has since put 76K trouble-free miles on it. The battery pack continues to work well, and according to SCE’s data, probably has many more years of service ahead of it. Today’s batteries are better than the 1990s technology Toyota used, and should deliver even better results. I cringe whenever I hear “the batteries are not ready”.

  6. Earl Killian says:

    I should add to the above that the calculations were done to be quite generous to biofuels. For example, this is 2005 vehicle miles traveled data from the BTS, and then the new CAFE standards for MPG are assumed, even though those won’t go into effect until 2020. The data for plug-ins is however based on old vehicles, not assumptions of future improvements. To use current MPGs, increase the land area requirements for biofuels by 56%.

  7. Michael says:

    I’ve done calculations for corn ethanol based on information contained on the Iowa corn growers and other advocacy sites (I got 40800 MJ/acre) and it comes out to 789,000 square miles of corn to substitute just for pump gasoline in the US using an extremely favorable EROI of 2.0 that is in reality closer to 1.2 to 1.3. 789,000 square miles is 22% of the area of the US including Alaska. We don’t know yet what the EROI of cellolusic ethanol will be…right now, from my understanding it’s worse.

  8. john says:

    “eventually complement his investment (flex fuel plug ins).”

    the only way to use is to use them as the fuel in a plug-in if the PHEV had great electric range and didn’t need much fuel both in the car and if widely used in response to peak oil. it’s best hope is as a back-up to a good PHEV. it probably doesn’t have as much of a future w/o a good PHEV.

  9. Paul K says:

    You are one of an extremely small number who were allowed to purchase the battery Rav4. Do you know Toyota’s reason for stopping production? They, like GM with the Volt, destroyed all the vehicles that were leased or in the Cal. program.

  10. Roy T says:

    Khosla has been completely debunked on the theoildrum.com by Robert Rapier, et al. The coup-de-grace for many was Khosla bold declaration that EROEI (energy return on energy invested) was not relevant

    But one example discussion

  11. Michael says:

    I’m sorry, I was looking at the gross and not the net production of corn ethanol: It is 40800 GROSS MJ/acre so with an energy returned on invested of 2.0 that means NET 20400 MJ/acre. Meaning we double the area to 1.5 million square miles or 44% of the area of the US including Alaska would need to be planted in corn with the productivity of Iowa. The more realistic EROIs mean areas close to the entire area of the US, just to substitute for gasoline usage, not other types of fuel.

    I guess my natural impulse is not to humiliate my opponents but in this case, the facts speak for themselves.

  12. John Mashey says:

    1) If anyone actually wants to make progress, commenting away on blogs isn’t going to have much effect.

    2) People should look at Khosla Ventures’ website and decide whether they think Vinod is:

    a) The Great Enemy.
    b) Only interested in making money and touting his portfolio
    c) An ally, and potential all, but sometimes dead wrong

    or some combination thereof.

    Vinod is for a sure a smart guy who is well-known. He is a VC, and VCs take risks, and make mistakes. VCs always tout their portfolios (some), and they are in business to make money … but for what it’s worth, there are much easier ways to make lots of money than the way Silicon Valley VCs invest.

    3) If you think that c) is part of this, then I’d suggest that Joe touch base with him and offer to help him understand why his biofuel investments need PHEV/FF to happen. Vinod communicated with Rapier, so I’d guess he’d even more likely communicate with Joe. As it stands, it’s hard to see how a VC invests in PHEVs, but if Vinod gets the idea they’re good, maybe things will pop up.

    4) If everybody thinks Vinod is a), and he should be demonized at every chance, I’m sure ExxonMobil will cheer you on.

    [Note: I don't know Vinod personally, and I have no money in KV.]

  13. Michael says:

    You are contradicting your #1 statement in making a thoughtful contribution here. There is the hope that people of influence will read what is said here and do something ever so slightly differently. Sometimes, if you approach these people directly they think you are trying sell them something. If you write it on the web, it may seem to some more disinterested and therefore more believable.

    As I said in my original post KV is to be praised for investing in Ausra…and consequently VK doesn’t have to do as much damage control as for his riskier investments.

  14. Earl Killian says:

    Michael, you’ll find the data I used cited in my spreadsheet at
    (or .xls or .html)
    In particular, the value I got for corn ethanol from Wikipedia was 3,900 L/ha (http://en.wikipedia.org/wiki/Ethanol_fuel gives a range of 3100-3900). At 23.5 MJ/L, this comes out to 91,650 MJ/ha. If the figure is 40,800 MJ/ha, then indeed the land area is over twice what I posted. Wikipedia got its data from a Nature “business feature” (not a journal article): http://www.nature.com/nature/journal/v444/n7120/pdf/444670a.pdf
    It is unclear where the Nature writer got her data. Please cite your source for comparison.

    I did not try to put in a EROI factor. To do that, one would deduct from 23.5 MJ/L the energy for the corn to ethanol conversion. In
    I find they use 21.2 MJ/L for ethanol, with 7.4 MJ/L of energy for corn farming, 12.3 MJ/L for ethanol production, -4.3 for the co-product credit, giving a net energy balance of 5.7 MJ/L. 5.7 MJ/L multiplied by 3,900 L/ha gives 22,230 MJ/ha. That is close to the number you used. (If one uses the Pimentel column, the result would be negative!) However, this is making the assumption that the energy to convert corn to ethanol comes from ethanol itself, which is probably not appropriate.

    Again, I was trying to be quite generous to biofuels, and even under generous assumptions, they look enormously inferior to using sunlight to charge car batteries.

  15. Earl Killian says:

    Paul K writes “You are one of an extremely small number who were allowed to purchase the battery Rav4. Do you know Toyota’s reason for stopping production?” I have heard that Toyota set aside enough parts for a fixed number of RAV4-EVs long before they brought them to market. If true, that suggests they calculated how many credits they needed under the California ZEV program, and built accordingly. It appears they didn’t want to market the vehicle; they were simply responding to the ZEV mandate. When California’s Air Resources Board folded in the GM/Toyota lawsuit against the mandate, Toyota no longer had a need to do another production run. Also, because of a patent dispute with Chevron, Toyota could no longer use the Panasonic NiMH batteries that have worked so well in the RAV4-EV.

    By the way, as far as my wife being “one of an extremely small number,” we owe that honor, in a way, to GM. I sat on the EV1 waiting list for a long time (driving my planetkiller gasoline vehicle in the meantime). When GM finally wrote me to tell me I wasn’t going to get one, I got a Solectria Force instead. Within a year, Toyota offered the RAV4-EV for sale, and my wife snapped it up (I still drive my Force). According to http://en.wikipedia.org/wiki/Toyota_RAV4_EV my wife was one of 147 people to buy them in 2002.

    Paul K also wrote, “They, like GM with the Volt, destroyed all the vehicles that were leased or in the Cal. program.” That is not true. They initially planned to crush the vehicles as their leases expired, like GM. However, a protest effort by DontCrush got Toyota to reverse their decision (and also stopped Ford from crushing their Th!nks). After this success, DontCrush morphed into PlugInAmerica.

  16. Earl Killian says:

    John Mashey perhaps had my comments in mind when he chided us for blog comments. The purpose of discussion is to both teach and to learn. I agree that blog comments usually don’t measure up to that standard, but this time something useful came of it. Consider’s at Michael’s response to my post. If he follows up by citing his Iowa data, I may end up with a better source for the calculations I’ve been doing, so the “learn” aspect may have be operative.

  17. Michael says:

    I got the corn productivity data here:
    I got 512 gallons ETOH/acre but this can vary. I used Iowa figures to be favorable to ETOH.

    Energy balance here:

    I took the highest number of 2.0 but 1.3 seems closer to what most researchers report. I did not use Pimentel, not because he’s wrong but because he’s considered an opponent by the corn lobby.

    Energy conversion factor here for both :

    I used the LHV value though it doesn’t make a huge difference given the staggering number of acres required.

    I took the value of 9.2 million barrels a day (x 365) of gasoline for 2006 from the US DOE EIA site.

    None of this data comes from a corn ethanol opponent.

  18. John Mashey says:

    No, I didn’t mean Earl’s comments at all, but maybe I wasn’t precise enough.

    There are lots of comments & blogs in blogs that are perfectly useful, but may not be sufficient. It is perfectly useful to discuss various topics for the readers of a blog, and I’m very happy to see this blog getting more attention (at least, it seems that way to me at first glance).

    It’s simply that in many blogs, there may well be plenty of agreement about something, and everyone is happy to agree, like on:

    a) Science reporting is bad, especially in newspapers.
    b) There are lots of hardcore climate change denialists, and look at what {Inhofe, Monckton, etc} have done now.
    c) VK is wrong.

    and the net effect in the real world is ~0 unless people take appropriate action, usually over a long period, and usually with many unsuccessful attempts. Joe has certainly been trying for years.

    For example, regarding a), we had a long discussion of this over in Deltoid:

    about positive measures to take with regard to science reporting.

    In the particular case of VK, unlike b) and c), he is one person, and hence far more likely possible to reason with, unlike b) (where it’s almost impossible), and c) where it’s a dispersed, long-term educational contest.

    As always, one must pick one’s wars. I just think, in this case, that it’s a whole lot easier to get the ear of somebody like VK than a lot of others, and people do listen to him, right or wrong. If I knew VK personally, I’d happily introduce him & Joe, but I don’t.

    If Joe’s going to be out here [Silicon Valley] some time, I might be able to arrange a meeting with another serious VC who invests in this turf [and Earl knows] and who knows VK well. Maybe, sooner or later, that would lead to something, or maybe not. Again, Rapier was able to establish a dialog of sorts with VK, and I think Joe has even more real-world credibility, so one never knows.

    Michael: if one can get VK to read Climateprogress, that would be good, but if he were reading it, he would probably already have posted something. With all due respect to blogs, not everyone reads everything relevant, and VCs are *busy*. Most don’t read blogs at all….

  19. Earl Killian says:

    Michael, thanks for the references. Here some FYI sort of comments/observations.

    From your Iowa page, I got 500 gallons/acre, not 512. That converts to 4677 liter/ha, which is higher than the 3900 I got from Wikipedia, but only by 20%. Even 500 seems a little high compared to 2.7 to 2.8 gallons/bushel * 176 bushels/acre = 475 to 493 gallons/acre (or 4443 to 4610 liter/ha). The 176 comes from 2.2e9 bushels / 12.5e6 acres, numbers from
    The 2.8 gallons/bushel comes from your page, but it elsewhere implies 2.7 (1.7 billion gallons / 630 million bushels and 4.9 billion gallons / 1.8 billion bushels).

    So far this is all within 20%.

    The most interesting reference you posted in the USDA/ANL paper on energy balance. First, it presents data on 9 states, rather than just Iowa, and obviously Iowa cannot fulfill all transportation needs, so the 9 state averages are much more relevant. For example, the 139.34 bushels/acre is a lot lower than the 176 calculated above from the Iowa page (even the Iowa entry in the paper is lower at 152.06). 139.34 converts to 3500 L/ha, which is the middle of the Wikipedia range. 156.02 converts to 3840 L/ha, or pretty close to the high-end of the Wikipedia range.

    What is more interesting about Table 1 is that it gives you most of what you need to substitute bioenergy for fossil energy on a case by case basis, instead of just using a single factor for EROI. For diesel substitute biodiesel (e.g. from rape) land use, for gasoline substitute ethanol land use, for electricity use biomass land area (e.g. 18.1 GJ/ha from http://www.sciencemag.org/cgi/reprint/314/5805/1598.pdf), for LPG and NG use the biomass gross energy land area (e.g. 68.1 GJ/ha from the same reference). (Since LPG/NG are probably used for heating, and burning biomass for heat is very efficient, the last one probably makes sense.)

    That leaves the issue of co-products. Notice that makes the difference in Tables 3 and 4 between EROI of 1.10 and 1.77 for dry milling. What a difference! USDA/ANL are correct to adjust the EROI for co-products for small-scale ethanol. However, at the same required to replace all gasoline, co-products become waste products: it is unlikely we have any use for that much DDGS in the U.S., and it will take energy to dispose of them (e.g. compost and return to the fields). With an EROI of 1.10, the multiple on land area is enormous! I’ll leave it here.

  20. Earl Killian says:

    Michael, another interesting report to consider:
    The co-products issue looms large.

  21. Michael says:

    This is not my field but in selecting a high EROI (2.0) I think I’ve allowed for a lot of extra efficiencies that may emerge in using co-products.

    Patzek, who is a (controversial) specialist in this area, tries to account for ecosystem services that biofuel production would need to re-supply to the soil to maintain its fertility. It gets complicated fast. He concludes that only certain types of tropical biofuel production may become net positive in terms of energy production after enough organic material is returned to the soil.


  22. Paul K says:

    This is not an endorsement, but it is worth noting that John McCain is the only presidential candidate of either party to oppose the awful and counterproductive ethanol subsidies.

  23. Ronald says:

    John McCain not supporting ethanol subsidies is something good. He represents a non agriculture state which allows him to do that. (Arizona)
    I imagine it hurts him in Iowa. To bad there is only one.

  24. Earl Killian says:

    Michael, I don’t think a factor of two covers it. Let’s take the numbers from your citation
    (USDA/ANL) in Tables 3 and 4. Take the dry milling process, which is superior to the wet milling process in terms of efficiency. With co-products they get an energy balance of 1.77, which means a factor of 2.3 is appropriate. Without co-products they get 1.10, which means a factor of 11.4 is required. I compute the factors as follows:

    A EB of 1.77 means that for each liter of ethanol, energy equivalent to 0.57 liters of ethanol when into its production (0.57 = 1/1.77). So I don’t need to produce one liter, but rather 1.57. But then to produce that extra 0.57 liter, I need 0.57*0.57 = 0.32 liters of input. But to produce that 0.32, I need an energy input of 0.32*0.57 = 0.18 liter. And so on. There is a standard formula for summing an infinite series of this form, and it is 1/(1-0.57) = 2.3. For an EB of 1.10, the same analysis is that to produce one liter, one needs 0.91 of energy inputs. The infinite sum is then 1/(1-0.91) = 11.4.

    Now, using ethanol as the energy input for producing ethanol may not be the best way to estimate land use. But if that is what you choose to do, then you need to get the calculation right, so you need to multiply by a factor of 2.3 if co-products are valuable, and a factor of 11.4 if co-products are waste. Multiplying by the energy balance number is not mathematically correct for using ethanol to produce ethanol.

  25. Michael says:

    Thank you for working this out. In my calculations of land area, I didn’t consider to figure in the infinite series which I appreciate you doing.

    The factor though, that you’ve arrived at means that one needs to multiply the area required for “net” ethanol production by that number? There should be a name for this factor to make it simpler to talk about. Maybe “Gross biofuel footprint factor” or something.

    I would add that there are competing uses for co-products, including animal feed, so it is difficult to know how much they will be re-used in ethanol production or sold for other uses.

    Also, none of these analyses takes into account the depletion of the soil, something that I am not competent to assign numbers to, but which Patzek has attempted to do in his analyses.

    One way or the other its, as a mass solution, unsustainable.

  26. Paul says:

    I don’t know if anyone will see this 8 months after the original post, but here goes:

    Regarding: “People always put-down plug-ins saying that the battery technology isn’t mature yet. That doesn’t pass muster with me. Toyota designed and built a battery vehicle in the 1990s…My wife bought one of those…I cringe whenever I hear “the batteries are not ready”.

    Earl, that’s rediculous. I guess you didn’t realize that Toyota intentionally sold this car at a loss. Toyotal lost about sixty thousand dollars per car. The RAV4-EV cost more than one hundred thousand dollars each to build. That’s wholesale manufacturing cost, not the MSRP that would be needed for Toyota to make money selling them. By the way, even at a huge loss, Toyota still charged around $42,000 for the -EV. A base model RAV4 was only $16,000. The cost was well over eighty thousand dollars more than a standard RAV4 before any kind of markup or profit.

    You paid $42,000 for a $16,000 car that cost over $100,000 to build, and that’s your version of “ten year old batteries are already ready for ‘prime-time’”? I don’t think so.

    Oh, and by the way, 76K miles and 5 or 6 years isn’t very long at all! What’s it going to cost to replace those batteries in a few more years? What will the environmental impact be of disposing of them?

  27. Paul says:

    Oops, that’s “ridiculous”, not “rediculous”.