2. The Challenge
This is the second post in a multi-part series. In Part 1, available here, I explained the idea of Moore’s Law, which states that computer technology gets about twice as good, for the same price, every two years. In this Part, I explain that it will deceptively difficult to get to 80% clean energy in 2035. Feedback is welcome to Jason.firstname.lastname@example.org.
In last month’s State of the Union, President Obama told Americans that the “future is ours to win.” He then listed a number of incredible American innovations, reminding us that “we’re the nation that put cars in driveways and computers in offices; the nation of Edison and the Wright brothers; of Google and Facebook.” Calling this our “Sputnik moment,” his paean to innovation ended with a bold green technology initiative: his goal is that “by 2035, 80 percent of America’s electricity will come from clean energy sources.” Cue thunderous applause.
President Obama properly noted that innovation is inherently unpredictable, and so he offered little detail about how we go about revolutionizing our energy use. “Some folks want wind and solar,” he said. “Others want nuclear, clean coal and natural gas. To meet this goal, we will need them all.” But although he’s right that no one can predict the exact shape of the next great innovation, the rate of improvement of some technologies can show remarkable stability, as demonstrated in Part 1 of this series. In the computer industry, that rate of change has been remarkably steep. But the green revolution will not be like the digital computer revolution. Change does not and cannot happen in this sector nearly as quickly. Accordingly, barring a remarkable technological breakthrough — always a possibility, of course — we will not hit 80% clean energy by 2035, and we should instead focus on crafting a policy that takes better account of medium-term technological realities. This post explains why in some level of detail.
Where We Are Now
First, we have to understand the scope of the challenge. How far are we from 80% right now? In a fact sheet released after the State of the Union, the Administration claims we are halfway there because our electricity grid is at 40% “clean” energy. Doubling our “clean energy” in 25 years — that makes the task seem quite manageable.
But things are not as they seem. For one thing, it’s impossible to come up with 40% by looking at the actual numbers compiled by the Energy Information Administration. Really. If you take a look at the latest data from the EIA below, you’ll only be able to find 30.6% “clean” energy — that’s the total from nuclear, hydroelectric, and “other renewables,” which essentially means wind and solar.
What’s going on? It turns out that the Administration hasn’t publicly said anything about this that I can find. But those in the industry understand that what’s happening is that the Administration is counting natural gas as “50%” clean — so the 20% share for natural gas becomes the additional 10% “clean” energy output. As I explain below, being able to include natural gas in “clean” energy won’t really make much of a difference in our ability to get to 80%, but let me briefly digress to say that this is a dubious move. The day before the State of the Union, a coalition of environmental leaders published an open letter noting that although gas may “burn ‘cleaner’ than coal on the short term . . . just like coal, natural gas is unsustainable and damaging to the climate and the environment.” (Curiously, despite the fact that the President of Public Citizen signed that letter condemning the government support of natural gas, the organization a week later called the President’s proposal “totally doable and relatively modest.”)
But even though giving natural gas half-credit as “clean” energy is not exactly a wonderful policy move at the outset, let’s assume that will be the case and dig somewhat deeper into the data. The key data point in the graph above is the tiny sliver of pie for the “other renewables” — that is, renewable sources like solar, wind, and geothermal, but not hydroelectric or nuclear power. The latest data shows that in the 12 months ended in October 2010, America generated about 162 million megawatthours of electricity from those “other renewables” — out of 4 billion megawatthours overall. That’s about 4% of our total energy output from those sources. In 2000, that number was about 80 million megawatthours, for about 2% of our total energy output. In other words, in a decade, we’ve been able to about double the amount of energy we generate in America from the “innovative” sources that President Obama emphasized in his talk.
Even counting natural gas as 50% clean, those “other renewables” must be the key to his plan. That’s because we’ve essentially maxed out all the hydroelectricity we can generate, and we actually produce about 100 million megawatthours less hydro now than we did 15 years ago. The other major clean energy category, nuclear, remains controversial, and building new nuclear power sources is difficult and takes a very long time. This amazing graph, from a recent DoE report about new energy sources, shows about all you have to know about the prospects of nuclear being a major part of the solution: [editor’s note: due to internet archive issue, the graph disappeared.]
Let me get to the bottom-line. There are four categories of “clean” energy that we have to worry about: nuclear, hydro, natural gas, and “other renewables.” Nuclear is at 20% and is realistically not going anywhere from that number (just ask the EIA, which predicts 17% nuclear in 2035 in its 2011 Energy Outlook). Hydro is at 7% and dropping. Natural gas is growing quickly, and half of it counts as clean. But if you do the math, you’ll see that the most natural gas can hit is “true” 40% of generation (for an adjusted “clean” proportion of 20%) because then there’s only 60% of our electricity left and that remainder would have to be entirely clean to get to a total of 80% clean overall. So there’s a natural limit to how much we can rely on gas even assuming we could turn off every coal plant in America and generate just from natural gas and other clean sources, and we are more than halfway to that limit for natural gas. (Moreover, as I’ll address next week, there’s no way we can turn off every coal plant in America by 2035, but that’s beside the point for now.)
And then there’s those “other renewables,” currently at a tiny 4% of generation. Assuming nuclear and hydro power hold steady (at 20% and 7% respectively) and that natural gas replaces every coal plant that it can, that would get us to 45-50% “clean.” So the “other renewables” category will have to jump from 4% of our 2010 electricity to 30% of our 2035 electricity. That would mean a jump from 160 million megawatthours to 1.5 billion megawatthours. Unless something major happens that changes America’s attitude toward nuclear energy or there is a major breakthrough in carbon capture technology such that existing coal plants can be retrofitted, that is the minimum amount of electricity we’ll need to be generating in 2035 as I see it: 1.5 billion megawatthours, or ten times the amount we’re now generating.
Oh, and one more thing you may not have heard, since it went unreported by the media as far as I can tell: 2010 saw “the largest build since 1985″ of capacity of new coal-powered power plants. (See slide 10 here). Not exactly an auspicious start.
The Slow Pace of Change
So we have a long way to go with renewable energy. But I admit that having a long way to go is not an insurmountable problem, because 2035 years is a long time from now. A lot can happen in 25 years. The real problem is how quickly we are innovating with these innovative sources. The pace of change of those critical “other renewables” — wind and solar, in the main — is too slow. The curves do not look anything like the escalating curve of Moore’s Law, where prices quickly plummet as quality rapidly increases in the computer sector. The curve is linear, and not exponential.
Let me give an illustration with solar, the most important of the innovative sources. A report from Berkeley Labs, run by the Department of Energy, revealed that the cost of installed, grid-connected solar systems was about $10.50 per Watt in 1998 and about $7.50 per Watt in 2007. That’s a nice savings, but we should recognize that in a decade, the cost-per-Watt did not even halve. Solar thus remains much more expensive than coal, by a factor of ten or more.
But now contrast that with how quickly costs drop in the computer industry. For instance, as I mentioned in Part 1, one gigabyte of hard disk storage was over $40.00 in 2000, and it is less than a nickel today. It is almost impossible to imagine how cheap storage space will be in 2020. Yet in ten years, solar cells will only be a little bit more efficient.
I am not the only one who thinks this. The technologist Kevin Kelly has a post containing more data, which shows that the improvement in batteries and in solar panels “has been increasing steadily for two decades” but “nowhere near the rate of computer chips.” Bill Gates isn’t optimistic about finding a Moore’s Law-like curve either, telling people last summer that we’ve been “fooled” by Moore’s Law, but in fact “there are things that don’t move forward.” As examples, he noted that “batteries haven’t improved hardly at all” and that “nuclear energy stopped [improving] in the 1970’s.” That’s why Gates is hoping for an “energy miracle” — business as usual just won’t be good enough. Bruce Everett, a Professor at Tufts’ Fletcher School, believes the situation is even worse than the numbers indicate, as in his view the recent cost decreases in solar are not due to significantly better technology but due to cheaper Chinese manufacturing. Labor costs surely do not follow Moore’s Law.
What has happened for the last few decades is not good enough to make solar cheap enough by 2035 so that it makes sense for it to be 30% of our electricity output (wind energy does not follow Moore’s Law but rather cost depends mostly on how good the location is, and we’ve already taken some of the best, cheapest places to put wind farms). This is the case even if the President gives Berkeley Labs and MIT and lots of other smart people more money for R&D. That influx of money won’t change physics, and I believe Moore’s Law is much more strongly related to the underlying nature of the technologies than to the amount of money or brain power we throw at the problem. (I recognize that this is a controversial claim, and I hope to explain it in more detail in a few weeks).
So that’s a first pass at the the challenge — but I’m sorry to say that it gets worse still. In my next post, I’ll explain two other, related factors that work against the 80% goal.fas