> our Roadmap conference in November, Om asked GE CMO Beth Comstock why a young person would want to work at GE when there are hot companies like Facebook out there. If there’s one compelling reason why anyone would want to, it’s because a company like GE can invest in R&D and tackle some of these massive, basic science, infrastructure and manufacturing problems. Because who’s more likely to deliver a jetpack, GE or Facebook?
It's a little sad that this even needs to be pointed out. VC's don't have either the money or the time to fund huge, capital-intensive projects. If you think about the really big projects going on right now, most are backed by the government or big public corporations. Google's self-driving cars. Tesla's electric cars,[1] Space X's low-cost launch vehicles,[2] etc.
The article mentions that VC works when Moore's law is in operation. Ironically, the actual operation of Moore's law at semi-conductor fabs isn't funded by VC, but massive capital investment from Intel, IBM, TSMC, etc. To put things into perspective: Intel's annual capital expenditures of $10-12 billion alone are roughly comparable to all VC investment in Silicon Valley combined (which is 40% of all VC investment in the U.S. as a whole).
[1] While Tesla paid off its federal loans early, it took $465 million for four years in order to build the factory that produces that Model S, arguably one of the most crucial turning points in the development of the company. Also, Tesla benefits enormously from the electric vehicle tax credit.
[2] Subsidized both by historical NASA research and NASA prepayments on launch contracts amounting to about half the money Space X has thus far raised.
Important folks are leaving places like GE and joining places like Tesla, SpaceX, and my own company LightSail, precisely because they are disenchanted with the proclivity of GE to actually decide to invest in real new inventions.
Despite the fact that GE could invest in new technologies, they are not. The man promoted to the head of technology for power and water for GE instead joined us, for, he says, exactly this frustration.
Hopefully, this will be a possible thing for them to change -- or they are going to keep losing their best people.
This is undoubtedly a win for science, but I wish we could push the envelope even further. For example, which company is willing to fund Dr. Bussard's IEC Fusion research [1]? The government is currently the only interested customer, and just barely at that.
While Tesla benefits from electric vehicle tax credits, so do their established competitors. If tax incentives for wind power did not exist, GE would not be in that business - I suspect almost everything GE is involved in has significant tax credits.
The unfortunate reality is that it frequently takes a change of policies to make projects even semi-profitable - electric cars rely on tax credits and tax incentives for charger installations. Comparatively high gasoline taxes help as well.
Government can take on financing risk and regulatory risk. VCs only want the financing risk, any deck that involves a point "And then we're going to lobby the government to raise taxes on competitive industry and/or give us and our clients a tax break" is not likely to go well.
VCs do seem to be able to raise larger sums, if not for long durations. Look at all the late investors into Facebook. The challenge isn't dollars, it's the need to return money in 7 years.
I'm not sure this author knows what Moore's law is. They seem to think it has to do with widespread adoption, which is a side effect of the real Moore's law.
Java and Javascript were invented in industry (Python came from academia). I don't see how you can deny the GPU computing and mapreduce were invented by industry, since GPU computing evolved out of the game industry, and mapreduce was invented by Google.
That leaves "commodity clusters" or "mobile computing".
Are these things to vague to be classified as inventions, as you imply? I think they are not. The founders of Google came up with the idea of using a large cluster of relatively unreliable servers, which was optimized to deal with machine failure. Other people then followed suit. Similarly there were several big steps (e.g. the iPhone) in putting a much larger amount of computational power into mobile devices than had previously been done.
In fact, your idea that innovation must come in discrete inventions, is a limitation that is typical of academia. In academia, you are asked to show how your work is qualitatively novel and interesting. In industry, objective improvements are adopted no matter how mundane or incremental the change is.
I seriously doubt the concept of map-reduce or using GPUs for computation was created by industry, although it was probably popularized there.
Computing on clusters whose components may fail was not invented by Google; perhaps they made it good. Industry did not come up with the idea of mobile computers even if they made improvements that made widespread use possible.
We can quibble over what words should be used to describe incremental improvement and sui generis invention but it's clear that they are different and the latter comes predominantly from state sponsorship.
KPCB very nearly owned the lion's share of Tesla, and missed out due to a mere $20MM in delta in valuation (VantagePoint offered $70MM, Kleiner offered $20MM). They then missed out again simply because of political struggles meant John Doerr couldn't take a board seat.
They were less than a board seat or $20MM in conviction away from being proven right -- now everyone thinks they are proven wrong. Just goes to show you the risks in the game.
They didn't get Tesla because Musk wanted Kleiner on the board not the partner who had done all the work. So Tesla took a $70M valuation with a board member they didn't care about rather than a $50M valuation with the board member they did want, which Musk says was a mistake. Still there were other missteps.
This article came out a day after this one, which is more in-depth about all the problems they have faced.
There are really a lot of potential legitimate reasons for wanting one partner on your board rather than another; particularly Doerr at KPCB vs Ray Lane, now emeritus.
Have they made it? I think there's going to be an interesting tug of war between tax credits for green tech and tax credits for what amounts to toys for the wealthy.
Right but if your strategy was to invest in computer hardware in the 80s and 90s you presumably did pretty well. If your strategy was to invest in cleantech in the last 5-10 years you probably did pretty poorly.
If VCs were able to see today's distribution of cleantech returns without seeing exactly who the winners and losers were I doubt they would have allocated nearly as much capital to that industry as they did.
So yes, high risk, high reward. But that means sometimes you get stuffed and in general those who invested heavily in cleantech got stuffed.
I'm wondering if this is true, and if so, why? It's the same winner take most dynamics. In most fields, like disk storage, it was extremely competitive and most profits disappeared.
VC in general has poor returns for funds outside of the top 10 or 20.
It is not the same winner take most dynamic: Most companies in a sector are not competing against one another, often they provide complementary goods and services. Also, small pieces of a big market can still be large.
Putting aside the specific example of computer hardware it is clear that industry and sector macro trends exist:
Consider the DotCom boom when tech VCs made superlative returns.
Yes. Cleantech was bad for Kleiner Perkins and a lot of other VCs that bet on it. But it also didn't work out for large companies, governments or research groups investing in it either.
Many of the arguments made here could equally well be applied to life sciences, but life science and biotech VCs have seen reasonable returns in investments and their LPs are often large pharma companies.
Also the author seems to be clueless as to the fact Moore’s Law is nothing to do with the speed of company growth but rather to do with the performance of computing power.
While their might be certain products that require an investment structure that VCs can't provide, this article fails to make any kind of coherent argument for it.
One good thing has come out of the VC's move on cleantech and that is bigger mindshare. More people are now aware of the possibilities and in particular, aware that concerted effort and investment towards a goal can pay off. It just takes longer than the VCs expected but note that one response to the VC activities has been that other companies are increasing their R&D spending.
It will be interesting to see how the stars of the digital age respond. We already have Google investing in various technology initiatives outside of the mainstream software and networking area. But what will Amazon and Twitter and Facebook do? Will one of them invest in solar breakthroughs? Or high speed rail?
It's a little sad that this even needs to be pointed out. VC's don't have either the money or the time to fund huge, capital-intensive projects. If you think about the really big projects going on right now, most are backed by the government or big public corporations. Google's self-driving cars. Tesla's electric cars,[1] Space X's low-cost launch vehicles,[2] etc.
The article mentions that VC works when Moore's law is in operation. Ironically, the actual operation of Moore's law at semi-conductor fabs isn't funded by VC, but massive capital investment from Intel, IBM, TSMC, etc. To put things into perspective: Intel's annual capital expenditures of $10-12 billion alone are roughly comparable to all VC investment in Silicon Valley combined (which is 40% of all VC investment in the U.S. as a whole).
[1] While Tesla paid off its federal loans early, it took $465 million for four years in order to build the factory that produces that Model S, arguably one of the most crucial turning points in the development of the company. Also, Tesla benefits enormously from the electric vehicle tax credit.
[2] Subsidized both by historical NASA research and NASA prepayments on launch contracts amounting to about half the money Space X has thus far raised.