Episode Summary
Executive Summary: David Roberts and Shale Khan compare their biggest open clean-energy questions for the next 5–10 years: whether autonomous vehicles will change urban form and emissions, whether data-center demand will force off-grid power, whether software-enabled homes and cars will become exploitative platforms, whether electrification can survive high power prices, whether recycling will become a strategic mineral source, and whether geoengineering will move from theory to real-world testing.
Main Topics: Autonomous vehicles and urban form (Priority: 5/5): They debate whether self-driving cars will reduce fatalities and pollution while also encouraging more vehicle miles traveled, sprawl, and lower-density development by making long commutes easier and more comfortable. Data-center growth and off-grid power (Priority: 5/5): Khan argues the key question is how much large load will go fully off-grid as AI/data-center demand strains the grid; they discuss bridge power, microgrids, gas vs storage, and whether demand forecasts are overblown. Softwareification and “inshittification” of homes and cars (Priority: 4/5): Roberts worries that as homes, chargers, thermostats, and vehicles become software-managed platforms, users may face lock-in, ads, variable pricing, privacy loss, and exploitative subscription models. Electrification under high electricity prices (Priority: 5/5): They question whether industrial electrification can continue if power prices keep rising and siting new loads gets harder, since cheaper electricity is central to the economics of decarbonization. Recycling as critical-mineral supply (Priority: 4/5): Both see battery, solar, and rare-earth recycling as moving from an environmental nice-to-have to a strategic and potentially cheapest source of minerals, though Khan notes recycling lags demand growth. Geoengineering governance and moral hazard (Priority: 3/5): Khan raises whether a meaningful solar-radiation-management demo will happen and whether billionaire-led experiments create unacceptable moral hazard, even as the technology may be cheap enough to become tempting. Permissionless DERs and balcony solar (Priority: 3/5): A spare question centers on plug-in distributed energy devices that can be installed without utility approval, and whether they could create a DIY solar culture and broader public engagement.
Key Arguments: AVs may lower deaths, noise, and tailpipe pollution, but by making car travel easier they could increase congestion and sprawl, undermining density and possibly raising emissions indirectly. Because all current autonomous fleets are electric, AV expansion may accelerate vehicle electrification, partly offsetting the extra vehicle miles traveled. The major open question for data centers is not only how much demand exists, but how much large load will abandon the grid entirely or rely on long bridge-power periods and microgrids. If grid interconnection remains too slow or constrained, off-grid data centers may push innovation in distributed generation, storage, and modular infrastructure. Homes and vehicles becoming software-managed could improve efficiency and flexibility, but without interoperability and privacy rules they may become locked-in platforms that exploit users. Clear interoperability, transparency, and privacy rules could prevent the worst forms of platform lock-in and reduce the risk of shittified home-energy products. Electrification depends on cheap electricity; if electricity prices rise and siting becomes harder, industrial electrification loses some of its core economic advantage. Even if data-center demand is overstated, the broader electrification of transport, heat, cooling, and industry still implies the need for much more generation and a stronger grid. Recycling should be treated as a strategic source of critical minerals, not just waste management, because it can reduce dependence on new mining and imported supply chains. Khan argues recycling will remain behind current demand during rapid growth, but it can still become cheaper and more important over time, especially for mature materials like copper and aluminum. Solar-radiation management is likely to be pursued by someone eventually because it is relatively cheap and potentially impactful; the challenge is preventing uncontrolled billionaire or state experimentation. Once solar geoengineering becomes common knowledge, it may be hard to contain, making governance and international coordination urgent but difficult. Permissionless DERs like balcony solar could have outsized social effects even if their direct megawatt contribution is modest, by changing public participation and awareness. Cheap Chinese solar exports are already reshaping developing-world grids and politics, potentially creating rapid, unplanned renewable buildouts that stress grid stability and governance.
Data Points: Crossover episode date: February 25th, 2026 - Episode intro Previous crossover timing: November 2023 - Roberts references the earlier Volts-Catalyst crossover AV adoption horizon: 5 to 10 years - Both hosts use this timeframe for widespread autonomous-vehicle rollout and clearer answers AV presence in San Francisco: 1 out of every 3 cars - Khan describes Waymo saturation in the Bay Area city Khan’s son’s driver’s-license bet: 12 years - He predicts his 4-year-old son may never need to drive when licensed US electricity share of final energy: 25% - Khan notes the US is still far from full electrification Non-electric share of final energy: 75% - Khan emphasizes how much energy use remains to electrify Current data-center/AI load concern horizon: next 5 to 10 years - Khan frames the period when grid strain and off-grid behavior may emerge Off-grid reliability tradeoff: Lower grid reliability / fewer 'nines' - Discussed as the main downside of leaving the grid Battery recycling demand lag: ~10 years behind demand - Khan explains recycling supplies reflect old demand, not current growth Solar-radiation-management cooling estimate: about 0.5°C - Khan cites rough estimates for a possible planetary cooling effect Solar-radiation-management cost estimate: a couple billion dollars - Khan suggests a billionaire could fund meaningful cooling at relatively low cost Rare-earth elements: 16 total; 4 heavily used - Khan explains why recycling can be cheaper than virgin mining due to reduced separation needs Balcony solar legalization: 25 states proposed or announced - Roberts cites US legal momentum for plug-in DERs Balcony solar installed base in Germany: ~4 million systems - Roberts cites Germany as evidence of strong consumer uptake China solar exports to Pakistan: 40% of total load equivalent imported in 2–3 years - Roberts describes rapid distributed solar growth in the developing world
Pivotal Quotes: "How much large load goes off grid in the next five to 10 years?" — Shale Kahn: Khan reframes the data-center boom around the possibility of off-grid power and microgrids "Do we want intrusive, ad-based, ad-supported subscription nagging, different tiers, you know, real-time variable pricing, all these sort of exploitative things that we're running into? Do we want that in our cars and in our house?" — David Roberts: Roberts expresses concern that home and vehicle software will replicate platform exploitation "Is electrification dead or will electrification be dead as a pathway for industrial emissions reduction?" — David Roberts: Roberts asks whether expensive electricity and data-center competition could undermine industrial decarbonization
Implications: The conversation suggests clean-energy winners will be shaped by grid bottlenecks, platform governance, mineral circularity, and political affordability. The next decade may determine whether electrification accelerates or stalls, and whether new technologies empower users or lock them into exploitative systems.