At a Seattle climate event a couple weeks back, Bill Gates reiterated his long-standing argument for increased government support for clean energy innovation and R&D (transcript and video). Key quote:
In my view, you need a portfolio of investments, but the thing I think that’s the most underinvested in is basic R&D. That’s something that only the government, really, is going to do. Once you get further downstream, there are a lot of opportunities [for companies] … That upstream part is part of what makes me optimistic, even though we seem to have a political road-jam, at least temporarily, on some of the issues like carbon pricing. The innovation piece really is so important, and I do see good things happening, but the federal government should more than double what it’s spending on that piece.
This is a point he’s been making on his own and as a part of the elite American Energy Innovation Council, and it’s one we largely agree with: you’ve got to invest in innovation to make cleantech work. If you only worry about pushing the existing menu of clean technologies out the door, and largely ignore developing the affordable technologies you need, like many climate activists do, you’re not going to adequately address climate change, and just end up raising costs and pushing emissions overseas. Cleantech is actually a basket of different technologies with different challenges and pathways to competitiveness. All will benefit from innovation to reduce costs, so policy should promote that innovation wherever and whenever possible, whether it’s during the deployment phase or not.Anyway. The downside is that whenever Gates makes his argument, many folks immediately jump to comparisons between clean energy and ICT, and use them to argue that Gates is wrong and doesn’t understand the history of his own industry, and that deployment is actually the single most important thing you can do. Case in point is Grist columnist David Roberts’ reaction from a couple weeks back. Roberts argues that Gates misses two key points due to “blind spots.” First, that deployment can change the world, and second, that distribution accelerates innovation. Quoting Roberts:
Rather famously, Bill Gates and his early Microsoft crew didn’t invent anything new or groundbreaking. Personal computers already existed (remember Tandy?) and the graphic interface idea (“windows”) was already featured in Apple products. What Microsoft did is scale those ideas up — way up — through deployment…But for some reason, when it comes to energy, Gates has no faith in deployment. He thinks we spend too much money encouraging deployment of existing energy tech. Why? Why couldn’t deployment do for energy what it did for computing?
Roberts goes on to argue that, because developing large-scale cleantech is difficult and expensive, we should instead shift away from large power sources and focus on small, decentralized energy, because it’s cheaper, easier and more conducive to decentralized innovation. On how distribution accelerates innovation, he says this:
If anyone should understand this, it should be the guy who masterminded the personal computing revolution. Computers used to be gigantic, centralized mainframes run by small groups of super-geeks in big IBM labs. Those early computer guys were completely dismissive toward the very idea of personal computing….Of course, once computing power was decentralized, widely dispersed into the hands of ordinary people, the world saw an explosion of innovation and economic productivity the likes of which hadn’t been seen since the post-WWII boom.
Roberts deserves a lot of credit for being a tireless climate advocate, but unfortunately he seems to be on the wrong track here.The biggest objection I have is over his assumption about where innovation actually comes from, and he’s not the first to make these arguments. He correctly observes that the private sector—including big companies, small entrepreneurs, and geeks in their garage—accounted for big portions of the innovation found during the ICT revolution. But they were not the only sources. The ICT industry very much owed its creation to public sector investment. Public funding—especially the military—contributed to early technological developments in the sector, and provided early markets, and something like half of all academic computer science funding came from the public sector up through the ‘80s. And even later in the revolution, after the private sector had fully taken up the innovation and deployment mantle, government still had a role to play in facilitating and driving it through initiatives like SEMATECH, research funding, and other ways. What does this show us? It was a mixed public-private approach that led to success at different stages of the ICT revolution. Yes, a decentralized system was and is an essential element for promoting innovation, but so is coordination, and incentives in the private sector are often at odds with this goal. And we should also realize that each innovation revolution is unique. I’d argue it doesn’t just need to be the same mixed, public-private story in energy; it needs to be several versions of the same story. In the IT revolution, Microsoft, Apple, and Dell all worked on different versions of the personal computer. In the cleantech arena, we effectively need the equivalent of, say, ten different personal computers: one “family” of machines to capture wind power, one family to capture solar power, one family to store energy in the grid and its cousins to power electric cars, one to convert biofuels, one to split (or fuse) the atom, plus all the control technologies, grid upgrades, and so on. This is a bigger problem than computers, and with a ticking clock to boot thanks to a changing climate. Just as there is a diversity of technology needs, there is a diversity of sources for innovation. Can demand, deployment, and learning-by-doing be a source of innovation? Of course – the ICT revolution demonstrated that, as have other industries. But lifting the innovation story from ICT and applying it to energy means looking at the whole story. And that story doesn’t start with Microsoft or Apple – it starts with the University of Pennsylvania and the Army Ballistics Research Laboratory in the 1940s, before Gates was even born, and encompasses far more than the conventional narrative. Deployment changed the world, as Roberts says – but not just deployment. This angle also gets somewhat at Roberts’ argument for distributed generation over centralized sources. The energy space has historically been about centralization—big plants producing lots of affordable power for legions of businesses and consumers. The inherent size of power plants means longer project development times and higher capital costs, thus making developing big new power sources a fundamentally different animal than IT, which is much more decentralized and user-scale in many ways. As Stanford President John Hennessy said at the Energy Frontier Research Center Summit just yesterday, this difference has presented a challenge for startups attracting VC funding: “This is not the Internet, this is not Google. It’s a much bigger problem.” He went on to argue that many in the VC community need to start thinking differently about energy. Roberts’ suggestion to get around this problem by emphasizing distributed generation means focusing on that which in some ways is more like IT and computing than traditional, big energy. It’s decentralized, doesn’t have the same capital costs as big biorefineries or solar farms, and in theory could be more accessible for many more small players to get involved, which could lead to more competition and innovation. I’d argue we’ll probably end up needing both centralized and distributed, though I can understand the attraction to the latter. This definitely isn’t meant to be an argument against it. Shoot, if it turns out to be the best, most affordable low-carbon approach, I’d be all in favor of a system that uses distributed, user-scale energy sources and nothing else. Of course, centralized power also has its advantages. But the point is that even if we ignored centralized power entirely, the best bet for small-scale power is probably still a mixed approach that facilitates innovation throughout research, demonstration, and deployment, with appropriate policies to deal with different technologies at different development stages and with a proper role for government somewhere in the system, just as it was for IT. It’s never been about leaving it to the geeks in their garages. Truth be told it’s not like there’s a hard-and-fast line on this stuff. Plenty of scientists and engineers in the public and private sectors are working on improved catalysts for biofuels conversion, or materials for hyper-efficient solar cells, or advanced batteries, and there’s not much reason to think these new technologies won’t be relevant for centralized and distributed energy sources alike. So perhaps it’s not as much of an either-or situation anyway. Still, a balanced approach that embraces both centralized and distributed generation, and understands the complexity of the innovation story, seems wisest.