Claire and Dan unpack Moore's Law — the principle that technology improves and costs decline as production scales — and explore its powerful role in driving down the price of solar, wind, and other clean energy technologies. They connect the dots between semiconductors, renewable energy, policy incentives, and innovation curves. They tackle the big questions: how far can Moore's Law take us, where does it fall short, and what does it mean for the future of AI, energy efficiency, and fusion power?
NOTE This file was generated by Descript
00:00:03 --> 00:00:07 Today we're gonna be talking about Moore's Law and Energy Innovation.
00:00:07 --> 00:00:10 This is Energy Matters with Claire and
00:00:11 --> 00:00:11 Dan.
00:00:19 --> 00:00:21 All right, this is gonna be a fun one, fun juicy one.
00:00:21 --> 00:00:24 Tell me what you wanna talk about with Moore's Law today, Dan.
00:00:24 --> 00:00:24 No question.
00:00:24 --> 00:00:26 So first we should define our terms.
00:00:26 --> 00:00:26 Yep.
00:00:26 --> 00:00:32 So Moore's Law comes from Gordon Moore's observation in the 1960s and
00:00:32 --> 00:00:38 early seventies where he noted on a technical basis that the number of
00:00:38 --> 00:00:41 processing units on a computer chip.
00:00:42 --> 00:00:46 Doubles or gets twice as dense about every 18 months.
00:00:46 --> 00:00:51 And if you flip that around, it means the computing power goes up.
00:00:51 --> 00:00:56 And as the price of those processors goes down, what we observe and what
00:00:56 --> 00:01:01 we call it in the world of energy systems, is the learning curve.
00:01:01 --> 00:01:06 And very broadly what that means is that what we observe.
00:01:07 --> 00:01:15 Phenomenologically or observationally is that the price of energy systems
00:01:15 --> 00:01:19 like solar panels and wind turbines, but also things out of the energy
00:01:19 --> 00:01:26 world like Bic pens, calculator else, they drop in price by about 20%
00:01:26 --> 00:01:28 every time you double the number of.
00:01:28 --> 00:01:35 Widgets that are both manufactured and sold and all of that's important in
00:01:35 --> 00:01:38 this learning or cost decline curve.
00:01:38 --> 00:01:41 And for the true nerds like me in the audience, this is called
00:01:41 --> 00:01:45 a power law, and we'll get back to those later, but the idea.
00:01:45 --> 00:01:49 Stemming from Moore's original observation, we're getting more and more
00:01:49 --> 00:01:52 efficient, better and better at computing.
00:01:52 --> 00:01:57 And there is a world of discussion about, is this how we can
00:01:57 --> 00:01:59 quantify and track innovation?
00:01:59 --> 00:02:03 Is Moore's law ongoing forever or will it come to an end?
00:02:04 --> 00:02:07 And the death of Moore's law has been forecast many times
00:02:07 --> 00:02:09 over, and like Mark Twain said.
00:02:10 --> 00:02:13 Reports of my death are greatly exaggerated.
00:02:13 --> 00:02:14 I love that.
00:02:14 --> 00:02:15 I love that.
00:02:15 --> 00:02:20 So one of my favorite examples of Moore's Law as uh, probably not surprisingly,
00:02:20 --> 00:02:26 is how TV modules, photovoltaic modules, um, and their pricing has.
00:02:26 --> 00:02:29 Consistently outgrown their expectations.
00:02:29 --> 00:02:34 So every single year that we have predicted how low solar costs will
00:02:34 --> 00:02:38 end up being and how many solar installations will be installed,
00:02:38 --> 00:02:41 we have underestimated grossly.
00:02:41 --> 00:02:46 There's a great article in The Economist, which we will put on our website,
00:02:46 --> 00:02:51 energy Matters World, that goes through installations that have been, you
00:02:51 --> 00:02:55 know, always been more than three times higher than their five-year forecast.
00:02:55 --> 00:02:56 And I love that statistic.
00:02:56 --> 00:03:00 And I think that's so critically important because it's not just
00:03:00 --> 00:03:02 that this has persisted for years.
00:03:02 --> 00:03:06 So the Moore's Law, the learning curve for solar panels is the most
00:03:06 --> 00:03:11 famous, and everyone who gives a talk about energy shows a version of this
00:03:11 --> 00:03:13 from the International Energy Agency.
00:03:13 --> 00:03:13 That's right.
00:03:13 --> 00:03:18 So, so pardon me for describing a graph, which is probably, you know,
00:03:18 --> 00:03:22 a, a no no in the podcast world, but thinking about it, what it says
00:03:22 --> 00:03:24 is that if you look at the price.
00:03:25 --> 00:03:31 Of a solar panel versus not time, but versus the number of
00:03:31 --> 00:03:33 solar panels that get deployed.
00:03:33 --> 00:03:38 What we see is a steady fall in prices, and what we normally do
00:03:38 --> 00:03:43 is we plot this in power law or log log space, and it's a straight
00:03:43 --> 00:03:45 line and it's been a straight line.
00:03:45 --> 00:03:49 Downward to lower prices since the 1970s.
00:03:49 --> 00:03:53 And so if we just put that in context, solar power.
00:03:53 --> 00:03:55 Back in the seventies when really the only installations were on
00:03:55 --> 00:04:02 Sky Lab and some remote locations, it was $50 per watt or more.
00:04:02 --> 00:04:09 Today, 2025 solar is as low as 50 cents.
00:04:10 --> 00:04:12 Per kilowatt hour, so it's fallen.
00:04:12 --> 00:04:12 Oh, cheaper way.
00:04:12 --> 00:04:12 Yeah.
00:04:12 --> 00:04:15 I, I'm just trying to, trying to get a, a, a sort of a benchmark.
00:04:15 --> 00:04:21 So prices have fallen dramatically, orders of magnitude, and I think
00:04:21 --> 00:04:25 that the current record, but Claire is gonna hopefully correct it.
00:04:25 --> 00:04:31 Is that the cheapest solar project I've seen today is a large scale project
00:04:31 --> 00:04:37 in the desert of Dubai and the all in price, the solar panel, the installation,
00:04:37 --> 00:04:41 the land, the everything else came in at something like 60 cents for.
00:04:42 --> 00:04:44 Everything per watt generated.
00:04:44 --> 00:04:44 Yeah.
00:04:44 --> 00:04:46 That's hard to beat though.
00:04:46 --> 00:04:49 I will say at Sunrock distributed generation where we're financing
00:04:49 --> 00:04:52 solar installations, I was just pricing a project that
00:04:52 --> 00:04:54 was a at a buck 98 in Maryland.
00:04:54 --> 00:04:57 Yeah, so these are amazing numbers when you think how great the solar
00:04:57 --> 00:05:01 installation is in the middle of the Dubai desert versus here, but the broader.
00:05:02 --> 00:05:07 Point about these learning curves or these experience curves is that
00:05:08 --> 00:05:12 it's not just a function of doing the research and development, it's also a
00:05:12 --> 00:05:18 function of getting the product into different modules or built into solar
00:05:18 --> 00:05:23 panels that are building integrated on rooftops or low cost arrays in
00:05:23 --> 00:05:25 your backyard or in some big system.
00:05:26 --> 00:05:31 And it's a function of the policy environment, the learning, what are
00:05:31 --> 00:05:33 the barriers to put it in place.
00:05:33 --> 00:05:36 And so this is not a law like Eagles mc squared.
00:05:36 --> 00:05:38 It's not a basic law of physics.
00:05:39 --> 00:05:44 It's a observational, or what we call in the physics world, phenomenological.
00:05:44 --> 00:05:47 We observe it out there and we explain it after the fact.
00:05:47 --> 00:05:53 But what it means is that we have so many decades of its experience working.
00:05:54 --> 00:05:56 And we'll talk about when it doesn't work.
00:05:56 --> 00:06:00 Hint, hint systems that you can't mass produce and then mass deploy.
00:06:00 --> 00:06:00 Yep.
00:06:01 --> 00:06:06 But if you can count on it, which we've done for decades in solar panels,
00:06:06 --> 00:06:11 electric vehicles, wind turbines, what it means is that if you do what economists
00:06:11 --> 00:06:17 call forward price or subsidize, when the technology is new and expensive.
00:06:18 --> 00:06:22 Remember when you've only deployed a few solar panels or a few electric vehicles,
00:06:23 --> 00:06:25 it's easy to double the number out there.
00:06:25 --> 00:06:25 Right?
00:06:25 --> 00:06:28 It's not so easy when you're way mass commercial.
00:06:28 --> 00:06:29 Scale down the, the curve.
00:06:29 --> 00:06:29 Yeah.
00:06:29 --> 00:06:34 But what it means though is that, um, well placed and.
00:06:34 --> 00:06:38 Ideally, temporary subsidies in the beginning can do a great deal
00:06:38 --> 00:06:40 to drop the price for everyone.
00:06:40 --> 00:06:45 As an example, so when I started Sun Edison, we were getting a self generation
00:06:45 --> 00:06:52 incentive program or SIP rebate of $4 and 50 cents back in 2005, and we
00:06:52 --> 00:06:54 were selling projects for $9 a watt.
00:06:55 --> 00:06:55 You.
00:06:55 --> 00:07:00 At this point in 2025, you would be a fool to spend, you know, more
00:07:00 --> 00:07:04 than somewhere between two 30 and two 50 a watt for an installation
00:07:04 --> 00:07:05 pretty much all over the country.
00:07:06 --> 00:07:09 Of course, it depends on labor costs and wage rates and you know,
00:07:09 --> 00:07:12 et cetera, et cetera, and whether it's a carport or ground mount.
00:07:12 --> 00:07:17 But you know, we're seeing prices anywhere from a buck 80 to four bucks a watt,
00:07:17 --> 00:07:19 depending on how complicated projects are.
00:07:19 --> 00:07:22 I mentioned this record project in In the Middle East, where admittedly
00:07:22 --> 00:07:27 they got the land at some incredibly low cost 'cause it's desert of
00:07:27 --> 00:07:30 basically 60 cents for all in.
00:07:30 --> 00:07:36 But what it says to our thinking is that anyone who wants to forecast where
00:07:36 --> 00:07:40 energy prices could go, needs to really look hard at these learning curves.
00:07:40 --> 00:07:43 And so what Claire highlighted is from the.
00:07:44 --> 00:07:48 Regular reporting by the International Energy Agency in Paris where she worked.
00:07:49 --> 00:07:49 Yep.
00:07:49 --> 00:07:53 At one point, what we observe is that utility executives,
00:07:53 --> 00:07:55 people in different fields.
00:07:56 --> 00:08:03 Are consistently underestimating how powerful this law is and the forecast
00:08:03 --> 00:08:08 year after year don't do justice to the actual cost declines and the
00:08:08 --> 00:08:12 actual increase in the amount of these technologies being sold, right?
00:08:12 --> 00:08:13 You can use that.
00:08:13 --> 00:08:18 In your models, in your business case forecast, in whatever else to say.
00:08:18 --> 00:08:22 If we keep building the market as the US was doing and may or
00:08:22 --> 00:08:24 may not be stalling now, right?
00:08:24 --> 00:08:30 But China and others are doing really strongly, you can then influence.
00:08:30 --> 00:08:35 The continuation of this, so the prices will fall as a function of what you do.
00:08:35 --> 00:08:41 And so there's a virtuous or a positive feedback cycle of taking advantage of
00:08:41 --> 00:08:48 Moore's Law, getting more technology in the field, seeing low prices a result, and
00:08:48 --> 00:08:50 then using that to continue the process.
00:08:50 --> 00:08:50 Right?
00:08:51 --> 00:08:54 The only caveat I would add to Moore's law is regulatory uncertainty.
00:08:55 --> 00:08:55 Right.
00:08:55 --> 00:09:00 So if you change laws, which is happening every day these days, then
00:09:00 --> 00:09:01 you know things can get outta whack.
00:09:01 --> 00:09:04 And it's not just on the federal level, it's also on the state level.
00:09:04 --> 00:09:08 You might look at the Massachusetts SMART 3.0 program, which is changing how
00:09:08 --> 00:09:10 things are going in the state of Maine.
00:09:11 --> 00:09:14 They're retroactively changing some of the rules, so you might be.
00:09:14 --> 00:09:16 In construction on a project and the rules are changing.
00:09:16 --> 00:09:21 So, you know, the, on the technology perspective, Dan, I totally agree with
00:09:21 --> 00:09:25 you with Moore's Law, but then, you know, people come in and muck everything up.
00:09:25 --> 00:09:28 And just to highlight how, I mean, Claire said the amazing word.
00:09:28 --> 00:09:31 There are some of these changes regulatory in nature.
00:09:31 --> 00:09:34 That are retroactive or what I would say punitive.
00:09:34 --> 00:09:34 Absolutely.
00:09:34 --> 00:09:39 And so Nevada may take the cake for a place that enabled solar
00:09:39 --> 00:09:42 because there were federal subsidies and programs out there.
00:09:42 --> 00:09:47 And then after the fact, the Nevada Public Utilities Commission that regulates this
00:09:47 --> 00:09:52 decided they wanted to claw back those benefits that had already been given out
00:09:52 --> 00:09:54 to homeowners that put solar on the roof.
00:09:54 --> 00:09:54 So, yep.
00:09:55 --> 00:09:58 That's goes way beyond kind of regulatory confusion.
00:09:58 --> 00:10:04 That's just outright hostility to the new leading edge or the clean energy leader.
00:10:04 --> 00:10:06 That is gross,
00:10:06 --> 00:10:07 which makes me sad.
00:10:07 --> 00:10:10 So let's get back to Moore's Law and Technology, right?
00:10:10 --> 00:10:13 So I think one of, you know, another point is energy
00:10:13 --> 00:10:15 efficiency between 2010 and 2018.
00:10:16 --> 00:10:21 Global data processed by data centers rose by six, a factor of six, but
00:10:21 --> 00:10:25 energy usage only increased by about 6%.
00:10:25 --> 00:10:29 But let's talk about data centers for a second because the, you know, and
00:10:29 --> 00:10:33 this is a question back to you, is, is Moore's law hurting us because there's new
00:10:33 --> 00:10:36 massive energy demand from data centers.
00:10:36 --> 00:10:41 And AI and crypto mining, is that, you know, is that making things
00:10:41 --> 00:10:42 worth for worse for the climate?
00:10:43 --> 00:10:46 So there's both a direct answer and my direct answer is gonna be, no,
00:10:46 --> 00:10:48 it's actually making things better.
00:10:48 --> 00:10:52 But let me back up, because this idea that as you make something cheaper.
00:10:54 --> 00:10:56 You do more and more of it has a long history.
00:10:56 --> 00:10:59 It goes back to something called Jevons Paradox.
00:10:59 --> 00:10:59 Yep.
00:10:59 --> 00:11:03 It was something that came out of the 1780s in England where people forecast
00:11:03 --> 00:11:09 that because England was getting better and better at mining coal, the prices
00:11:09 --> 00:11:10 would fall so people would use more of it.
00:11:12 --> 00:11:16 And we have a current version, which I would say has been totally debunked, but I
00:11:16 --> 00:11:21 think there's still probably some holdouts out there that used to claim there was a
00:11:21 --> 00:11:26 so-called rebound effect, right against energy efficiency, where if you made
00:11:26 --> 00:11:28 energy cheaper, people used more of it.
00:11:28 --> 00:11:31 And this was a. Negative spiral.
00:11:31 --> 00:11:36 And I would argue that the analytics have said that these kind of
00:11:36 --> 00:11:39 innovation pessimists, or whatever you want to call them, were wrong.
00:11:39 --> 00:11:41 I think they would still say they weren't.
00:11:41 --> 00:11:46 But I think the data shows that we actually don't want to just
00:11:46 --> 00:11:47 use things for its own sake.
00:11:47 --> 00:11:51 So for example, if we make driving cheaper by moving to electric
00:11:51 --> 00:11:54 vehicles, doesn't mean that we'll spend our whole lives in the car.
00:11:54 --> 00:11:55 Right.
00:11:55 --> 00:11:58 We might spend our whole lives in the car because of traffic, but we don't
00:11:58 --> 00:12:02 spend our whole lives in the car because the price of driving has dropped.
00:12:02 --> 00:12:07 And that is kind of this perverse look at learning curve.
00:12:07 --> 00:12:07 So.
00:12:07 --> 00:12:07 Right.
00:12:07 --> 00:12:11 This question that you asked is a little more focused than that,
00:12:11 --> 00:12:12 I think, for sure.
00:12:12 --> 00:12:16 But I think there's a lot of discussion about AI right now and the Jevons paradox
00:12:16 --> 00:12:18 and what AI is going to do to all of that.
00:12:18 --> 00:12:18 Yeah.
00:12:18 --> 00:12:21 So I guess, you know, there's no single answer and we will
00:12:21 --> 00:12:22 be doing multiple episodes.
00:12:22 --> 00:12:26 You can be sure on AI and data centers and machine learning, but I think
00:12:26 --> 00:12:28 that the simplest version from my.
00:12:28 --> 00:12:34 Look at the data today is that whatever worries or benefits or upsides you have
00:12:34 --> 00:12:39 around, um, AI usage and I have a friend I just talked to yesterday who said
00:12:40 --> 00:12:47 he talks to chat GPT so much that he has renamed it chat and when he hasn't
00:12:47 --> 00:12:49 been on for a while, chat asks him.
00:12:49 --> 00:12:50 Is something going on?
00:12:50 --> 00:12:51 Are you okay in your life?
00:12:51 --> 00:12:53 Why aren't you talking to me regularly?
00:12:53 --> 00:12:53 Yeah.
00:12:53 --> 00:12:57 So leaving aside the, a little spooky, the, the kind of, the spooky aspect of
00:12:57 --> 00:13:04 it, I would say that just like calculators in the 1970s when teachers said, well,
00:13:04 --> 00:13:09 because kids use calculators, their math skills will deteriorate, um, and other
00:13:09 --> 00:13:13 kind of technology crises, if we use ai.
00:13:13 --> 00:13:20 To enable productivity and not to replace our thinking, then we actually
00:13:20 --> 00:13:25 want to use more and more energy because we want to do more AI and
00:13:25 --> 00:13:26 machine learning based projects.
00:13:26 --> 00:13:32 And because of Moore's law with the falling price of clean energy,
00:13:32 --> 00:13:34 that's actually a wonderful issue.
00:13:34 --> 00:13:38 'cause if we can do what people in the field call electrify everything.
00:13:38 --> 00:13:42 Move not only from a home that's partially electric and partially
00:13:42 --> 00:13:44 gas to an all electric house.
00:13:44 --> 00:13:49 Add solar panels, switch our driving, have battery storage, battery
00:13:49 --> 00:13:53 storage, switch our driving from petroleum cars into electric vehicles.
00:13:53 --> 00:13:55 That's part of this electrify all.
00:13:55 --> 00:13:57 And because it is true today.
00:13:57 --> 00:14:02 Thanks to, um, to Moore's Law and the learning curve that it's now
00:14:02 --> 00:14:09 cheaper to build with new money, a clean energy power plant than it is
00:14:09 --> 00:14:12 to operate an existing fossil plant.
00:14:12 --> 00:14:17 Electrifying all and greening all will push us down the learning
00:14:17 --> 00:14:22 curve and will benefit everything, including the use of ai.
00:14:22 --> 00:14:24 So I view AI as a good.
00:14:25 --> 00:14:28 Steady customer for clean energy, not some worry.
00:14:28 --> 00:14:29 Yep.
00:14:29 --> 00:14:35 Another question which I hear from my dad and and naysayers of which I put my dad
00:14:35 --> 00:14:42 in that is, is the rapid thank you dad, is the rapid obsolescence or e-waste cycle.
00:14:43 --> 00:14:45 How does Moore's law factor in that?
00:14:46 --> 00:14:46 Oh boy.
00:14:46 --> 00:14:47 So I
00:14:47 --> 00:14:49 know that's a tough one, but you know, the question is, you know,
00:14:49 --> 00:14:51 can you recycle solar panels?
00:14:51 --> 00:14:51 Yeah.
00:14:51 --> 00:14:52 Yes you can.
00:14:52 --> 00:14:56 What are we gonna do with all these wind turbines, you know, when they're obsolete?
00:14:56 --> 00:15:00 So the, I, there's so many aspects and you who do the business
00:15:00 --> 00:15:02 side of this should really be answering your own question here.
00:15:02 --> 00:15:06 I know, but I would say one feature is that the lifetime of a solar panel
00:15:07 --> 00:15:10 is somewhere between 25 and 40 years.
00:15:10 --> 00:15:12 So their lifetimes are long.
00:15:12 --> 00:15:19 And while humans are not very good at recycling today because of laziness
00:15:19 --> 00:15:24 and lack of policies, I'm actually gonna predict that 10 years from
00:15:24 --> 00:15:29 now, the fastest growing new research institutes and the new departments
00:15:29 --> 00:15:34 that will be launched at Colleges of Engineering around the world will be on.
00:15:34 --> 00:15:40 Recycling, upscaling, reusing materials because we could be
00:15:40 --> 00:15:46 building almost everything to make the process of disassembly
00:15:46 --> 00:15:48 and recycling easier and easier.
00:15:48 --> 00:15:49 Absolutely.
00:15:49 --> 00:15:50 And we've been very, very lax.
00:15:50 --> 00:15:52 Lax on doing so.
00:15:52 --> 00:15:57 But even if we weren't good at it, solar panels last multiple decades,
00:15:57 --> 00:16:03 wind turbines last couple decades, and these are all high value materials that.
00:16:04 --> 00:16:09 If we're even halfway intelligent, we can build markets in to recycle.
00:16:09 --> 00:16:13 And let me give you one example that goes totally under the radar screen.
00:16:13 --> 00:16:15 Anyone who bought a laptop.
00:16:16 --> 00:16:24 Or a cell phone in California, almost certainly unwittingly paid a $14
00:16:24 --> 00:16:31 rough, roughly $14 cost per cell phone and about a $29 cost per laptop.
00:16:31 --> 00:16:32 You didn't notice it.
00:16:32 --> 00:16:37 Nothing changed in your behavior, but those monies go into a process
00:16:38 --> 00:16:43 to assist companies to recycle, and should we be doing this for everything?
00:16:43 --> 00:16:44 Yes.
00:16:44 --> 00:16:46 Should the cost be higher?
00:16:46 --> 00:16:47 Probably also, yes.
00:16:48 --> 00:16:53 But this was put in place based on some ideas that BMW and others launched.
00:16:53 --> 00:16:59 Initially when they said, we can't guarantee that every physical atom
00:16:59 --> 00:17:05 and every vehicle that we make will be recycled, but if we use a hundred
00:17:05 --> 00:17:10 tons of platinum, we commit to recycling a hundred tons of platinum
00:17:11 --> 00:17:12 from the international market.
00:17:12 --> 00:17:18 Now whether you think this is actually something devious where they are
00:17:18 --> 00:17:21 making a commitment that looks good and they're just buying it up and other
00:17:21 --> 00:17:26 people have to deal with the problem is a fair complaint, but the idea that.
00:17:27 --> 00:17:34 Each ton of material that is used by a company or by a person is going to get
00:17:34 --> 00:17:40 recycled through an active policy like this that's not so hard to envision.
00:17:40 --> 00:17:42 We would do it across the whole economy.
00:17:42 --> 00:17:42 Absolutely.
00:17:43 --> 00:17:43 Yeah.
00:17:43 --> 00:17:48 I wanna get back to a point about, which I often hear again from naysayers of.
00:17:48 --> 00:17:49 Well, you have a lot of
00:17:49 --> 00:17:50 naysayer friends, I have to say.
00:17:50 --> 00:17:51 I sure do.
00:17:51 --> 00:17:52 I sure
00:17:52 --> 00:17:52 do.
00:17:52 --> 00:17:53 And family.
00:17:53 --> 00:17:53 Yeah.
00:17:53 --> 00:17:58 But, um, you know, the, the argument that like, oh, I shouldn't install
00:17:58 --> 00:18:01 solar right now 'cause I'm gonna wait until the technology gets better.
00:18:01 --> 00:18:03 Or, oh, I'm not gonna, you know, buy a newer.
00:18:03 --> 00:18:05 Refrigerator until the technology gets better.
00:18:06 --> 00:18:08 That seems like such a lame argument to me because, you know,
00:18:08 --> 00:18:12 for example, in 2008, grow Solar.
00:18:12 --> 00:18:13 Thank you Jeff Wolf.
00:18:13 --> 00:18:18 If you're listening, um, installed eight Evergreen Solar modules on my
00:18:18 --> 00:18:22 roof, uh, which were, I think 180 watts.
00:18:22 --> 00:18:25 And that's, you know, a silly size right now.
00:18:25 --> 00:18:28 And I paid $11 a watt for that installation.
00:18:28 --> 00:18:33 If I had waited, sure I would have better modules on my roof right now,
00:18:33 --> 00:18:38 but I wouldn't have saved a third of my energy costs every year for the last,
00:18:38 --> 00:18:40 you know, 15 years if I just waited.
00:18:41 --> 00:18:45 And you wouldn't have contributed to the learning curve by being an earlier adopter
00:18:45 --> 00:18:46 that then helped the prices go down.
00:18:46 --> 00:18:50 So I think for this one, we have to go to that high end source,
00:18:50 --> 00:18:51 and that is the peanuts cartoons.
00:18:52 --> 00:18:53 And in that.
00:18:53 --> 00:18:58 Um, Lucy talks about Arbor Day and what she highlights.
00:18:58 --> 00:19:00 She, first of all, she thought Arbor Day was harbor day, so she
00:19:00 --> 00:19:04 went down to the docks and that didn't pan out well for her.
00:19:04 --> 00:19:10 But what she did discover was that the best day to plant a tree was
00:19:10 --> 00:19:14 20 years ago, and the second best day to plant the tree is today.
00:19:14 --> 00:19:17 And that fits into this learning curve mentality.
00:19:17 --> 00:19:22 But it also says that if a new product gives you better services and notice that
00:19:23 --> 00:19:29 things that are purchased, there's already this inherent issue that we're making.
00:19:30 --> 00:19:32 The process of buying solar for your roof easy.
00:19:32 --> 00:19:36 If you own a roof, and if you are a renter, low income, it's much harder.
00:19:36 --> 00:19:42 So there is a justice inequality aspect in this story that a few policies
00:19:42 --> 00:19:43 have tried to address, but not many.
00:19:44 --> 00:19:45 That's a diversion for now.
00:19:45 --> 00:19:51 We'll come back to it, but the issue here is really if you can get the
00:19:51 --> 00:19:56 benefit of that service, even at a slightly higher cost, you and all
00:19:56 --> 00:19:57 other customers will benefit, right?
00:19:57 --> 00:20:00 By moving down the learning curve.
00:20:00 --> 00:20:02 Now, there are cases, early cases, early adopter as an
00:20:02 --> 00:20:03 early adopter in the process.
00:20:03 --> 00:20:09 Now, there are example technologies where learning curves don't hold.
00:20:09 --> 00:20:12 And one should be upfront about them because in some ways the
00:20:12 --> 00:20:14 exception defines the rule.
00:20:14 --> 00:20:22 So we observe strong learning behavior in modular technologies where the so-called
00:20:22 --> 00:20:29 just-in-time delivery of parts where the market can be amenable to uptake it, where
00:20:29 --> 00:20:31 the companies can grow to put it in place.
00:20:31 --> 00:20:33 All of that contributes positively.
00:20:33 --> 00:20:37 To these learning curves, meaning lower and lower and lower prices.
00:20:37 --> 00:20:42 Where we don't see learning is also instructive technologies that are
00:20:42 --> 00:20:47 already fully mature or technologies where you can't mass produce like
00:20:47 --> 00:20:52 building aircraft carriers or interestingly and controversially
00:20:52 --> 00:20:53 nuclear power plants, right?
00:20:53 --> 00:20:56 We see prices rising over time now.
00:20:56 --> 00:21:00 Nuclear proponents are gonna raise a flag right away and say, wait a second.
00:21:00 --> 00:21:04 That's due to regulatory issues or risk perception, whatever else.
00:21:05 --> 00:21:09 But the basic numbers, whether you're pro or against.
00:21:10 --> 00:21:13 Solar or nuclear or aircraft carriers or whatever else is that
00:21:13 --> 00:21:16 we observe this data very clearly.
00:21:16 --> 00:21:20 Things that can be standardized and mass produced and mass distributed
00:21:21 --> 00:21:24 exhibit learning behavior, and things that are more one-off
00:21:24 --> 00:21:26 and unique for whatever reason.
00:21:27 --> 00:21:28 Don't observe.
00:21:28 --> 00:21:29 We don't observe this behavior
00:21:29 --> 00:21:33 though you would think with nuclear there would be more modular design.
00:21:33 --> 00:21:37 So I think how in the last how many years, we've only installed one nuclear
00:21:37 --> 00:21:41 power plant and it was, the cost overruns were just extraordinary, right?
00:21:41 --> 00:21:41 That's right.
00:21:42 --> 00:21:45 So we'll be talking a whole episode, many episodes, but one coming up
00:21:45 --> 00:21:49 very soon about nuclear and actually my lab just published a paper
00:21:49 --> 00:21:51 comparing the costs of nuclear.
00:21:52 --> 00:21:57 In the United States where we have about a hundred nuclear plants, France,
00:21:57 --> 00:22:03 where there's about 58 and China, where there's around 25, but more than half
00:22:03 --> 00:22:08 of all new nuclear plants all worldwide that are planned to be built in the
00:22:08 --> 00:22:10 next decade are gonna be built in China.
00:22:10 --> 00:22:14 So what we observe is that the cost of nuclear in the US.
00:22:16 --> 00:22:20 Across different plant designs and different manufacturers, the cost has
00:22:20 --> 00:22:25 risen steadily to make it vastly more expensive today than installing the
00:22:25 --> 00:22:27 same amount of renewables plus storage.
00:22:28 --> 00:22:33 France also shows cost increases, although not quite as bad as the US.
00:22:33 --> 00:22:33 Well,
00:22:33 --> 00:22:36 and to interrupt you for just a minute, 75% of France is powered with nuclear.
00:22:37 --> 00:22:40 And the pro and the prices are rising, which is why France is
00:22:40 --> 00:22:42 diversifying into wind and solar,
00:22:42 --> 00:22:43 right?
00:22:43 --> 00:22:43 Um,
00:22:43 --> 00:22:46 versus Germany, which has already always done a lot of solar.
00:22:47 --> 00:22:50 Which, and of course Germany is interesting because like California,
00:22:50 --> 00:22:54 they have both said, and that's the third, that's the fourth and the
00:22:54 --> 00:22:56 fifth largest economies in the world.
00:22:56 --> 00:22:56 Yep.
00:22:56 --> 00:23:01 They have said we won't do any more nuclear because of risk and waste issues.
00:23:01 --> 00:23:03 Although they both cheat.
00:23:04 --> 00:23:08 Both California, they imports nuclear power from the four corners area of
00:23:08 --> 00:23:14 the southwest, nuclear by wire, and France sells nuclear to Germany.
00:23:14 --> 00:23:14 Right.
00:23:14 --> 00:23:17 So Germany doesn't have the plants, but they have the benefit
00:23:17 --> 00:23:19 of those nuclear electrons.
00:23:20 --> 00:23:20 Right.
00:23:20 --> 00:23:25 And so both places cheat, but they're able to say they're nuclear free.
00:23:25 --> 00:23:25 Yep.
00:23:25 --> 00:23:27 And back to Germany for just a second.
00:23:27 --> 00:23:30 They were an early adopter of the feed in tariffs and all of the concepts
00:23:30 --> 00:23:32 behind making solar cost effective.
00:23:32 --> 00:23:34 Yeah, and I think that's a really critical story because Yep.
00:23:35 --> 00:23:40 In Germany, um, Herman Sheer, who was essentially the architect of the solar
00:23:40 --> 00:23:47 revolution that made the pricing of solar super not only cheap, but also simple, and
00:23:47 --> 00:23:51 the way it worked in the feed and tariff in the early model, it's evolved in the
00:23:51 --> 00:23:53 many countries that have adopted it since.
00:23:53 --> 00:23:58 Was that the earlier that you commit to putting solar on your roof in a
00:23:58 --> 00:24:03 given period of time, say a one year block of solar deployment, you get.
00:24:03 --> 00:24:10 A 20 year commitment for the utility to buy your energy at a fixed price.
00:24:10 --> 00:24:10 Yep.
00:24:10 --> 00:24:13 And that price was often quite generous.
00:24:13 --> 00:24:14 Sometimes too generous,
00:24:14 --> 00:24:15 sometimes too generous.
00:24:15 --> 00:24:15 Yeah.
00:24:15 --> 00:24:21 And that led to learning in Moore's Law because there was a known set market.
00:24:21 --> 00:24:22 Yep.
00:24:22 --> 00:24:22 And.
00:24:23 --> 00:24:30 As we have evolved, uh, and developed a model that was adopted by all the
00:24:30 --> 00:24:32 utilities in the United States, it's not clear where things are going
00:24:32 --> 00:24:33 now, or not all of them, not all,
00:24:34 --> 00:24:37 but at, for example, the Massachusetts SMART Program, the Solar Massachusetts
00:24:37 --> 00:24:38 Renewable Target program.
00:24:38 --> 00:24:40 That is a feed and tariff, right?
00:24:40 --> 00:24:45 And but places like California have had this other model, which is an RPSA
00:24:45 --> 00:24:48 renewable portfolio standard that.
00:24:49 --> 00:24:54 Percentage requirement meant the utilities had to install at minimum a
00:24:54 --> 00:24:55 certain amount of these technologies.
00:24:56 --> 00:24:59 No one knew exactly what punishment would happen if they didn't get it.
00:24:59 --> 00:25:04 When New Jersey punished its utility for not making its target, they used the
00:25:04 --> 00:25:07 fine to then install solar streetlights.
00:25:07 --> 00:25:09 So they got some more solar into the system.
00:25:09 --> 00:25:12 So there are some interesting little side stories.
00:25:12 --> 00:25:18 But all of these are part of a feedback process where installing more of the
00:25:18 --> 00:25:23 technology that makes your cell phone cheaper, that makes your laptop cheaper,
00:25:24 --> 00:25:29 goes into this process of making the product cheaper over time and.
00:25:30 --> 00:25:34 We literally rely on this in utility planning, in energy
00:25:34 --> 00:25:36 planning over and over again.
00:25:37 --> 00:25:43 So at we at Energy managers are a fact-based entity, but I
00:25:43 --> 00:25:46 wanna ask you, Dan, about some predictions to close this out.
00:25:46 --> 00:25:55 What do you think, what technology or concept that A, in which Moore's Law.
00:25:55 --> 00:26:00 Is applied is going to be the most valuable for reducing climate change.
00:26:00 --> 00:26:06 So I'm gonna say something that I think 99% of our audience will disagree with
00:26:06 --> 00:26:07 as you do,
00:26:07 --> 00:26:08 as I tend to do.
00:26:08 --> 00:26:08 Yeah.
00:26:08 --> 00:26:14 So my forecast is that in the year 2070, which I unfortunately do not think I'm
00:26:14 --> 00:26:22 gonna get to experience, I think the world will be 70% powered by fusion.
00:26:22 --> 00:26:28 And Fusion has zero market share today, but I think that 70% of our energy mix
00:26:28 --> 00:26:34 in the year 2070 is going to be half terrestrial fusion, meaning plants we
00:26:34 --> 00:26:38 don't have today, we only have one working fusion system in Livermore, California,
00:26:39 --> 00:26:44 but the other half of that, so 35% of our total energy I think is gonna be solar.
00:26:45 --> 00:26:49 Because solar is fusion 93 million miles away, meaning the sun, right?
00:26:49 --> 00:26:55 So I am banking on Moore's Law to keep the cost declines going for
00:26:55 --> 00:27:01 solar and from the first fusion plants that are vastly expensive because
00:27:01 --> 00:27:03 they're just research facilities.
00:27:03 --> 00:27:07 I think starting at about a decade, a decade and a quarter from now.
00:27:08 --> 00:27:13 We're gonna start to see fusion plants being deployed on the grid as pilots,
00:27:13 --> 00:27:19 but by 2070 we will have discovered that using the universe's most abundant
00:27:19 --> 00:27:25 element, hydrogen and some heavier hydrogen deuterium, we are gonna see
00:27:25 --> 00:27:30 fusion takeoff because fusion is not only a source of energy that we can do
00:27:30 --> 00:27:33 in power plants, but fusion has another.
00:27:33 --> 00:27:37 Killer application to go back to our startup episode, and that is
00:27:37 --> 00:27:39 moving us around the solar system.
00:27:39 --> 00:27:39 Yep.
00:27:40 --> 00:27:40 I love it.
00:27:40 --> 00:27:41 I love it.
00:27:41 --> 00:27:42 And don't mess with this guy.
00:27:42 --> 00:27:48 In 1992, he predicted that Miami would be underwater in 2050, and
00:27:48 --> 00:27:49 that prediction's gonna be true.
00:27:50 --> 00:27:53 I wish it wouldn't be true, but I fear we're on a path for that.
00:27:53 --> 00:27:56 So we have raced through the learning curve, or we've
00:27:56 --> 00:27:57 raced down the learning curve.
00:27:57 --> 00:27:59 We're gonna come back to it a number of times.
00:28:00 --> 00:28:03 This is Energy Matters with Dan
00:28:03 --> 00:28:03 and Claire
00:28:04 --> 00:28:07 and you can follow us and you can interact both via the
00:28:07 --> 00:28:09 website, energy Matters Matters,
00:28:09 --> 00:28:10 world,
00:28:10 --> 00:28:10 world.
00:28:10 --> 00:28:15 Or you can go to the email address and send us your comments, critical
00:28:15 --> 00:28:21 noncritical, new topics, et cetera, and that's info at Energy Matters World.
00:28:21 --> 00:28:22 Thanks for listening.