
45 segments available
How will we feed the 100s of GWs of extra energy demand that AI will create over the next decade? On this episode, Casey Handmer (Caltech PhD, former NASA JPL, founder & CEO of Terraform Industries) walks me through how we can pull it off, and why he thinks a major part of this energy singularity will be powered by solar. His views are contrarian, but he came armed to defend them. 𝐄𝐏𝐈𝐒𝐎𝐃𝐄 𝐋𝐈𝐍𝐊𝐒 * Transcript: https://www.dwarkesh.com/p/casey-handmer * Apple Podcasts: https://podcasts.apple.com/us/podcast/china-is-killing-the-us-on-energy-does-that-mean/id1516093381?i=1000722101753 * Spotify: https://open.spotify.com/episode/3vV7Vki2qJhO2AyodIaTMo?si=Gv0aEF9mSSu5cmPMksndRg 𝐒𝐏𝐎𝐍𝐒𝐎𝐑𝐒 * Lighthouse helps frontier technology companies like Cursor and Physical Intelligence navigate the U.S. immigration system and hire top talent from around the world. Lighthouse handles everything for you, maximizing the probability of visa approval while minimizing the work you have to do. Learn more at https://www.lighthousehq.com/employers To sponsor a future episode, visit https://www.dwarkesh.com/advertise 𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒 00:00:00 – Why doesn’t China win by default? 00:08:28 – Why hyperscalers choose natural gas over solar 00:18:01 – Solar's astonishing learning rates 00:27:02 – How to build 50,000 acre solar-powered data centers 00:40:24 – Environmental regulations blocking clean energy 00:44:04 – Batteries replacing the grid 00:49:14 – GDP is broken, AGI's true value must be measured in total energy use 00:58:45 – Silicon wafers in space with one mind each
Casey Handmer discusses the competitive landscape of AI development, emphasizing the importance of manufacturing solar panels, batteries, and other critical components. He highlights China's dominance in solar manufacturing and questions whether the U.S. can compete effectively in this industrial race. The conversation explores the implications of capital allocation and the business environment in both countries.
"Today I'm interviewing Casey Handmer. Casey has worked on a bunch of cool things. Caltech PhD on some gravitational wave black hole gimmick stuff, then Hyperloop, then the Jet Propulsion Laborato..."
Handmer explains China's energy production challenges, noting that only a third of its energy use comes from electricity. He discusses the potential of synthetic fuels to transform this dynamic, allowing China to leverage its electricity production more effectively. The segment highlights the significance of energy in the context of AI and industrial growth.
"Do you think that China is better at capital allocation than the United States? Do you think the Chinese business environment is better for business than in the United States? I feel you can make ..."
The discussion shifts to the socio-economic landscape in China, where Handmer argues that despite its autocratic governance, China has made significant industrial advancements. He cautions against underestimating China's capabilities while also emphasizing the importance of recognizing the U.S.'s potential to compete in the energy and AI sectors.
"Synthetic fuels have been around for 100 years. There are projects in China right now working on synthetic fuels. It would not surprise me if they were thinking pretty seriously about this. Just t..."
Handmer elaborates on the U.S. strategy of export controlling chips to maintain its AI lead and discusses how energy is a critical input in this race. He speculates on the potential consequences if China were to impose tariffs on solar and battery exports, highlighting the interconnectedness of the U.S. and Chinese economies.
"Like the Indian middle class is larger than the US middle class. But also it's nowhere near as wealthy. Whereas there are parts of China which are humongous, which are actually as wealthy as the ..."
In this segment, Handmer argues that the U.S. has the capacity to ramp up solar manufacturing significantly. He challenges the mainstream narrative that China has an insurmountable advantage due to cheaper labor and less regulation, asserting that the U.S. can compete effectively with automation and abundant resources.
"If we decided we want to produce 100 gigawatts of solar capacity every single year… We’re already on track to do that. Is it going to be as cheap as it is to do in China? My views on this are so..."
The conversation focuses on the current decisions made by hyperscalers regarding energy sources for data centers. Handmer critiques their preference for natural gas over solar, questioning the long-term viability of this choice and emphasizing the importance of solar energy in future data center operations.
"do that in two years or less if you started today. It's currently 11 o'clock. So we're going to start cutting checks by noon. You could ramp up pretty quickly. A lot of technology already exists h..."
Handmer discusses the constraints faced by energy producers, particularly in relation to natural gas availability and turbine manufacturing rates. He highlights the challenges of scaling up energy production to meet the demands of AI and other industries, emphasizing the need for strategic planning and investment.
"That's actually something that Elon and his companies are great at. It's figuring out this mass production, semi-automated mass production. They've got this facility in Texas which is making the ..."
Drawing parallels to historical industrial efforts, Handmer reflects on the rapid scaling of production during World War II. He suggests that similar levels of motivation and coordination could enable the U.S. to significantly increase its energy production capabilities in response to growing demands.
"our ability to go and rent gas turbines?" No, they're not, because there was enough available once, maybe twice. But at a certain point, you realize as you grow, you start to touch all these addi..."
In this segment, Handmer explains the complexities and costs associated with delivering energy to consumers. He discusses the challenges faced by utilities in maintaining infrastructure and the implications for energy pricing, particularly in the context of renewable energy sources.
"Not to reach prematurely for analogies, but Henry Kaiser set up the shipyard in Richmond, just down the road here in San Francisco near Berkeley. He was initially making ships for the British an..."
Handmer posits that large industrial operations, such as AI data centers, may need to build their own power plants to ensure reliable energy supply. He discusses the efficiency and economic rationale behind this approach, suggesting that it may become a standard practice in the industry.
"The problem that we see—and the reason that PG&E here in California, for example, is perpetually on the brink of bankruptcy—is that even though the cost of an additional solar panel or additiona..."
The segment addresses the future of turbine production and its implications for energy generation. Handmer discusses the current demand for turbines and the potential for scaling up production to meet future energy needs, particularly in the context of AI and data centers.
"Okay, AI might be a special case. But big picture question. Across different kinds of ISOs, from Texas to Pennsylvania to whatever, people are building data centers which will not be online for ..."
In this segment, Handmer breaks down the mechanics of conventional power generation, explaining the Brayton cycle and how it converts heat into electricity. He emphasizes the inherent costs associated with building and operating gas turbines, setting the stage for a discussion on the economic implications of energy sources for AI-driven data centers.
"of these turbines. Here's one thing you have to grapple with sooner or later. Conventional power generation is a steam engine. You have some kind of chemical that you find inside the earth that ..."
Handmer explores the economic value of AI services compared to electricity costs. He highlights that for hyperscalers, the cost of electricity is a minor fraction of the overall value generated by AI services, leading to a counterintuitive perspective on energy expenses in the AI industry.
"Okay. What is the cost of… GE makes these 100 megawatt gas turbines, right? I don't actually know what the retail price is. I would suspect that if their price is flexible, it would have gone up a..."
This segment discusses the anticipated shift towards solar energy for powering data centers by 2032. Handmer argues that the supply of natural gas turbines is limited and that solar energy's learning rates and cost reductions will drive its adoption, making it a more viable option for future energy demands.
"At the same time, what is the economic value to you of using Claude or Grok or whatever you use on a monthly basis? A lot. It's obviously much more than the subscription, but is it 10 times more ..."
Handmer delves into the impressive learning rates of solar energy production, citing Wright's Law and the rapid cost reductions associated with increased production. He contrasts this with the slower learning rates of natural gas, arguing that solar will continue to dominate the energy landscape for data centers.
"In 2032, we're going to have hundreds of gigawatts of extra demand for data centers and at that point, most of it is coming from solar? Why is that? There aren't enough turbines being manufacture..."
In this segment, Handmer challenges the conventional wisdom that solar demand will plateau. He argues that solar adoption is accelerating, and the increasing efficiency and decreasing costs will continue to drive demand, contrary to popular belief.
"So it’s roughly 15-20% per year. Then just as a result of that price reduction, demand skyrockets by probably six times more than that additional marginal production capacity increase. This is on..."
Handmer discusses the uncertainties surrounding the production of natural gas turbines and the risks associated with investing in their expansion. He highlights the potential for solar energy to outpace natural gas in the long term, given the current market dynamics and technological advancements.
"We’re still in the Apple II computer era of solar. Backing up, if the story is that the reason solar is getting cheaper is because there's a lot of demand for more solar, and that demand can sustai..."
This segment focuses on the projected energy consumption of data centers and the transition from natural gas to solar energy. Handmer predicts that by 2040, solar will dominate the energy supply for data centers, driven by the retirement of older coal plants and the increasing demand for energy.
"If I was looking at the same charts as they're looking at right now, I'd say, "What are the odds that in 25 years' time we can produce gas turbines at a price that is relevant in a world where ..."
Handmer addresses the misconceptions about land availability for solar energy production. He argues that there is ample land for solar farms, particularly in regions like Texas and Nevada, and discusses the logistical challenges of permitting and land acquisition for large-scale solar projects.
"If a gas plant is still making money, people will keep operating it. But at a certain point… It is the case right now that operating a coal plant costs more than building a new solar plant. So it..."
In this segment, Handmer outlines the considerations for building solar-powered data centers, including the need for large contiguous land areas and the importance of ensuring reliable energy supply. He emphasizes the advantages of solar over traditional energy sources in terms of sustainability and cost.
"10x more H100 equivalents in the world. About a kilowatt each, something like that. Okay, so that's like 100 gigawatts. That sounds roughly right. You're not the first person to give me a call and..."
Handmer discusses the critical role of solar energy in supporting the infrastructure needed for AI development. He highlights the importance of achieving high uptime for data centers and how solar energy can provide a reliable and sustainable power source for the growing demands of AI technologies.
"And you don't need a massive amount of land. The problem with the solar approach is that there's no two ways about it. It's a farming operation. You need a huge amount of land. The total amount of..."
Casey Handmer discusses the potential for solar energy deployment across various terrains, emphasizing that solar can be effectively utilized in diverse locations, including Europe. He argues that overbuilding solar capacity is not a negative, as it can provide excess energy for local utilities, creating a symbiotic relationship between data centers and energy production.
"No. Doesn't matter. Do trees grow on mountain slopes? So it doesn't matter. For reference, Nevada is something like 80 million acres. Just Nevada, which is like 90% federal land, is 80 million ac..."
Handmer outlines the infrastructure needed for solar-powered data centers, including the importance of battery storage and the spatial arrangement of solar arrays. He highlights the increasing power density of computing racks and the necessity for robust energy solutions to ensure high uptime for AI operations.
"you need to have a lot of solar overbuilt. Is solar overbuilt a bad thing? No. Is the fact that we produce 40% more food than we need a bad thing? No. It's much better than producing 40% less tha..."
In this segment, Handmer explains how solar arrays can be integrated with data centers and the grid. He discusses the logistics of connecting these systems, including the use of optical fibers and the potential for off-grid solutions in remote areas, emphasizing the scalability of solar energy.
"That's where all of US GDP growth is occurring right now. I guess you didn’t answer the question of , yes, theoretically we could do this, but is it going to be possible to get the permitting to ..."
Handmer provides a detailed analysis of the land requirements for large-scale solar installations, comparing them to historical land uses for nuclear projects. He discusses the feasibility of deploying extensive solar farms and the economic implications of such projects in relation to energy demands.
"on private land somewhere in the backwoods of Texas where no one lives and no one will ever live because it's completely inhospitable to humans. In terms of the ratios, one trend that was impress..."
This segment addresses the regulatory challenges faced by solar energy projects in the U.S. Handmer critiques existing environmental regulations that hinder solar deployment, arguing for a more balanced approach that allows for renewable energy growth while still protecting the environment.
"So you've got about 10 acres of solar. So 10 acres of solar, six truckloads of batteries, one truckload of data center, and some cooling stuff. For how big of a data center? One megawatt. That's j..."
Handmer discusses the economic considerations of solar energy compared to traditional gas turbines. He highlights the capital expenditures involved in solar projects and the strategic decisions made by tech giants regarding energy sourcing for AI infrastructure.
"For people who do optimization problems for fun, this is how you do it. You start off with a bunch of NREL data on what your solar abundance is in this particular part of the world, and then you ..."
In this segment, Handmer predicts the future of electricity prices in light of regulatory challenges and the growing demand for energy. He argues that current pricing reflects regulatory inefficiencies and emphasizes the need for innovation in energy technology to meet future demands.
"How much does a megawatt of solar cost? If you go and ask the usual suspects, they'll tell you a million dollars. But this is one of the things that breaks my brain at Terraform, which is my day j..."
Handmer concludes by discussing the critical role of solar energy in transitioning away from fossil fuels. He emphasizes the urgency of adopting renewable technologies to avoid economic decline and environmental disasters, advocating for solar as a key solution.
"firehosing energy at a given problem because it rains down from the sky. Between the fact that maybe solar prices will go down and the fact that demand is going to go up. "
Casey Handmer discusses the urgent need for solar synthetics and sulfur injection to combat climate change and prevent economic decline. He highlights the risks of flooding coastal cities and the stagnation of transmission line growth, emphasizing the importance of transitioning to solar energy to avoid a similar fate as the UK during its coal depletion.
"we'll get poor the same way the UK did, because they ran out of coal, basically. The second thing is we'll get poor because we'll flood our coastal cities than Florida underneath climate change. ..."
In this segment, Handmer explains how batteries can replace traditional grid functions by performing temporal arbitrage, storing energy for later use. He argues that while the grid is expensive and difficult to expand, batteries offer a more flexible and efficient solution for energy storage and distribution.
"That's a really great question. You and I had a conversation along these lines almost two years ago when we first met. It caused me to go and write a blog post. This is a good way of thinking abou..."
Handmer elaborates on the increasing per capita allocation of lithium-ion batteries, predicting a future where batteries are ubiquitous, from homes to substations. He discusses the efficiency of batteries in energy storage compared to traditional grid assets, highlighting their role in managing solar energy's variability.
"one side. We know why the grid is expensive. It's a lot of wires strung up in hard to reach places that are hard to maintain, especially as the workforce ages, with regulations and all the rest a..."
This segment focuses on the predictability of solar energy generation and the challenges posed by weather variability. Handmer discusses strategies for optimizing battery usage to mitigate curtailment and ensure a steady energy supply, emphasizing the importance of planning for low solar days.
"had of meaningfully storing energy, storing electricity on the grid, was pumped hydro. That only works in a handful of places and with limited capacity. It doesn't work all that well either. The ..."
Handmer speculates on the future energy demands driven by AI, discussing the potential for civilization to evolve towards a model where energy use is a primary measure of economic value. He highlights the need for significant energy resources to support AI advancements and the implications for society.
"That trend is only going to continue. We've got batteries that are performing this temporal arbitrage. The sun comes up every day, right? So the power swings from midday—you're otherwise curtailin..."
In this segment, Handmer explores the economic implications of achieving AGI, arguing that its value will be tied to energy consumption rather than traditional GDP metrics. He contrasts the current economic contributions of AI with the potential future value of human-level intelligence.
"costs are increasing year after year. So it's just very clear that the average distance the electron is going to travel between generation and consumption is going to decrease in the future prett..."
Handmer discusses the bottlenecks in AI development, particularly the relationship between hardware and software. He emphasizes the need for advanced hardware to unlock the full potential of AGI and the economic implications of automating human labor.
"In that asymptote, I want to get to the crazy nerd sci-fi…. What does our civilization look like? What is happening? Kardashev Level 1. Let's wait to get to turning the entire earth into an AI f..."
This segment critiques traditional economic metrics like GDP in the context of AI and energy use. Handmer argues for a shift towards measuring civilization's value based on energy consumption, highlighting the disconnect between energy's importance and its representation in GDP.
"Human labor generates on the order of $60 trillion of economic value. That's how much is paid out in wages to labor around the world. So that's what AGI can do. Even if you curtail it to just wh..."
Handmer examines how AI's integration into the economy could lead to deflationary pressures, potentially decreasing GDP while increasing overall value. He discusses the implications of AI complementing human labor and the emergence of new job roles tailored for AI capabilities.
"through the gas tanks of our cars and through our aircraft and in our grids and stuff like that. Right now, the AI revolution is about routing around cognitive constraints, that in some ways writ..."
In this thought-provoking segment, Handmer posits that the future of civilization may be better understood through energy consumption rather than GDP. He discusses the potential for AI to drastically reduce operational costs, reshaping economic structures and societal values.
"But if you don't have oil, then you have these oil shocks, which cause double digit decreases in GDP. So the elasticity of demand often matters more than its raw fraction contribution to GDP. Anyw..."
Handmer speculates on the future infrastructure needed to support AI's energy demands, envisioning a world where solar energy and advanced computing coexist. He discusses the minimal requirements for deploying AI systems and the potential for innovative energy solutions.
"Again, GDP will see this huge deflation because the variable cost of running AI will just be pretty cheap as compared to paying humans wages. At the point where you've got a mixed economy with an..."
In this segment, Handmer explores the role of silicon in future energy and computation systems. He discusses the potential for integrated solar technology and the implications for space-based energy generation, envisioning a future where silicon is central to both energy and AI.
"Exactly, if we think that the value of cognition is going to be unbounded, and the way to derive cognition—to the extent you think solar will eventually win—you can derive it from how much land ..."
Handmer concludes with a visionary perspective on the future of energy and AI, discussing the concept of solar sails and integrated systems for computation. He imagines a world where energy generation and AI capabilities are seamlessly integrated, paving the way for advanced technological evolution.
"You can actually make do with a buck converter or with relays or whatever to match the current output of your solar array with the charge state of your batteries and the power consumption of you..."
In this segment, Handmer explains the process of refining silicon from natural sources to meet the energy demands of future AI technologies. He discusses the multi-stage process of silicon production, emphasizing the potential for faster and cheaper methods using abundant solar power. This exploration of silicon's role in the energy singularity highlights the intersection of energy production and cognitive advancement.
"don't need to because they're immortal. Is this what the Dyson sphere will be made of, Casey? Is it just going to be computronium at the center of a solar cell? They can fly closer to the sun to g..."
Handmer presents a thought-provoking idea about the energy singularity, where the demand for energy and cognitive capabilities of AI converge. He reflects on the evolution of complexity in life forms and suggests that we may be witnessing a shift back towards simpler, more efficient systems that maximize energy use for cognition. This segment challenges listeners to consider the implications of energy on the future of intelligence.
"So then the constraint is, well, how quickly can you convert the crust into enough silicon to support silicon thought? What does the silicon ecosystem look like? Any thoughts? Well, it's pretty qu..."
In this closing segment, Casey Handmer invites listeners to consider career opportunities at Terraform Industries, the company he founded. He emphasizes the unique work environment focused on ambitious hardware projects and the chance to work with top talent in the field. Handmer shares his vision for the company's future and the exciting potential of their technology, including plans for extraterrestrial applications.
"We have fusion in stars and the inky blackness of space and that provides our temperature gradient. Then the most efficient way to convert that into usable cognition is silicon. Literally electron..."