Andy Grove thinks Intel should start making car batteries. Electric car batteries. Is he crazy or what?
Before we get into that, here’s my take on the whole global warming/energy consumption issue, partly to show you why Grove is so excited about batteries, mostly because there is so much nonsense out there on the subject:
1) While the science of global warming is still in its infancy and is often practiced by grotesquely biased practioners, it is nonetheless prudent to assume that at some point in time, increased CO2 levels in the atmosphere will cause significant change. So we should not indefinitely depend on carbon-based energy, even if we make the rather big assumption that supply will not be a big problem. This is especially so given that world energy consumption will jump throughout the 21st century as big chunks of the non-rich world stop being poor and start using energy like the rich world: improving diet, using much more of electricity, buying many more cars.
2) Because most big energy consumers by the latter half of the 21st century aren’t going to be all that rich, cost matters. It will be the ex-poor, not the rich, who will decide what the world does about global warming. There is no point to the Western world converting to extremely expensive non-carbon based energy, since that will just leave more relatively cheap carbon-based energy to be used by everyone else. Alternative energy prices must be at least in the same ballpark as carbon-based energy sources for widespread use, and realistically, it needs to be as cheap or cheaper. Understand that for most people, the issue won’t be paying more to be green. The issue is whether they can afford to run a car or air conditioner or whatever or not.
3) Any serious effort to greatly reduce carbon emissions must eliminate the internal combusion engine by electrifying the car and powering it through non-carbon-based power sources. That means batteries. No, not ethanol, no, not natural gas. They are not clean; they are just less dirty. Corn-based ethanol produces 78% of the carbon emissions of oil. Only cellulosic ethanol might theoretically qualify as a “clean” fuel source in a seriously green world, but we don’t have the technology yet, and as recent events have indicated for corn, an ethanol-driven world could be a “drive or eat” one. Natural gas produces about 70% of the carbon emissions of oil. Hybrids are conceptually ridiculous. Instead of using an internal combustion engine to drive the car, you add 500-1,000 pound of batteries to the weight of the car, then use an internal combusion engine to drive the car plus the batteries around. Yes, the engine would be smaller in a hybrid, but you’d be better off just using a tiny engine to begin with. It’s just as ridiculous to have an all-electric car that gets its power from an oil- or coal-burning power plant, all you do then is remotely emit carbon, and probably more of it, given transmission losses.
4) Electrifying our transportation system means we must dramatically increase our electricity-generating capacity. The only sure way we can do so in the short- to medium-term (i.e., 10-20 years) and at the same time reduce carbon emissions is to go nuclear. While carbon sequestration could be a decent interim palliative, especially for existing coal-fired plants, it is an unproven technologies, and certainly will reduce power output from a plant significantly. While wind power is economically competitive, it can only supplement, not substitute for base electrical power. Solar power isn’t economically competitive, and even if you’re an optimist, won’t be in a position to become a mainstream energy source until around 2030. It might well end up being a big part of the very long-term “solution” to oil and coal (maybe along with fusion, after all, the sun does go down fairly regularly), but we’ll be a lot older or dead by the time that happens.
5) There is no quick way to significantly reduce carbon emissions. Period. We will be very fortunate to reduce them by half of current levels by 2050. Anyone who tells you otherwise is naive, ignorant, stupid, dishonest, or some combination of the above. Part of this is because we really don’t have the technology quite yet to do so economically, much more of it is because we’ll have a lot more users, and you’re not going to be able to stop them unless you plan on conquering countries whenever their people drive or run the A/C too much. The biggest part of the reason, though, is that changing the picture, on both a societal and individual level, takes a lot of time and money. You won’t replace your TV or air conditioner or washing machine or furnace or car or house tomorrow for more efficient models; you just can’t afford to in short order. You may have a real problem if you find you have to pay a good deal more for using less energy, or find that the energy miser doesn’t work quite as well. Your friendly electric company may not be able to afford shutting down its current plants and building nicer ones, probably because you won’t want to pay them extra to construct and run them. Your friendly gas company will probably not be too big on replacing its pumps with plugs, and anybody else who might be more enthusiastic about it faces a chicken or egg problem: people won’t buy electric cars without places to plug them in, and people won’t buy and install places with plugs until there are cars that will use them. None of this makes conversion impossible, after all, we had to make even bigger changes to go from horses to cars. It just won’t happen quickly, and it won’t happen much until it is to the financial benefit of those who’ll make it happen.
If you happen to think we’re doomed if we don’t get a lot better right away, then we’re doomed. Fortunately for all involved, if you happen to think that, you’re wrong. The good news is that while the world won’t go on a strict carbon diet soon, by about 2100, we probably won’t be using much at all. We’ll gradually shift away from carbon-based to other energy sources over the course of the next half-century or a bit more, as technology, economics and a few governmental prods allow. The earth will probably warm up a little, no big deal; this planet has gotten a lot warmer and a lot colder than it is now or will be in 2070. If cavemen could survive Ice Ages, we’ll certainly survive one or two extra degrees of temperature. If you’re especially nervous, I heard there are great deals on northern Canadian and Siberian real estate.
With all that being said, let’s get back to Andy Grove and his idea. We can now see that creating much better and cheaper car batteries are a pretty big chunk to the energy solution in the long run, and Mr. Grove is certainly right in that in saying that somebody gets that right is pretty likely to make a ton of money in a couple decades or more. He’s certainly not crazy there.
What should be questioned is whether it’s a good idea for Intel to try to be that somebody.
What is Intel good at?
Intel is good at researching and developing, then efficently producing nanodevices, devices with structures that operate practically on the atomic level. So if you could build a better battery by building it like a CPU, taking advantages of effects that only work using nanostructures, well, yes, Intel would definitely have an edge that a Panasonic or Sanyo wouldn’t have.
But who is to say that it does? Maybe it does, maybe it doesn’t. For Intel to find out whether or not it has any competitive advantage, it’s going to have to do a lot of blue-sky research first, a lot of expensive blue-sky research with no guarantee of success. But even assuming they do, then they have to then somehow translate nanofabbing to what will be big bulky devices at a nanocost.
Let’s assume for argument’s sake that Intel comes up with something that can do with a hundred pounds of nanoteched battery what it takes a thousand pounds worth of lithium-ion batteries today. This would be a stupendous breakthrough, but then Intel has to fab that hundred pounds of nano-based battery for a selling price of say, $5,000. Do you know how many CPU cores you can make out of a hundred pounds of material, and more importantly, how much money Intel could get for them? What kind of fabbing capacity would be needed to make these things? You might say that Intel wouldn’t necessarily need to build batteries like a CPU, but if that’s the case, what particular advantage would they have over the battery companies?
We could go on and on about this, but given that we’re talking about building imaginary parts, let’s not. What can be certainly said about such a venture is that it would be a huge effort and huge distraction from what Intel does now. It’s the kind of venture where it would be bad if Intel failed and even worse if it succeeded. Intel would be in great danger of becoming a battery company that made CPUs on the side. That’s bad enough, but let’s face it, the track record shows Intel is very good at making CPUs and chipsets to run them, but little if anything else. Up to now, that hasn’t hurt them much because (outside of flash memory) their ventures outside of their core competencies have been small and no one ever thought any of them could possibly overshadow the main business. This would. The potential rewards would be so great that it would suck in resources long after it should have been written off. Remember Merced.
So far, the official Intel response to Grove’s idea has been a very polite version of “Yeah, yeah, sure we’ll look into it, Andy.” I think they’re right. Sometimes a good idea is a bad idea for a particular company.