Ep15: Inside Diablo Canyon Nuclear Power Plant | BG2 w/ Bill Gurley & Brad Gerstner

Episode Summary

Executive Summary: The episode is a pro-nuclear deep dive centered on Diablo Canyon’s role in California’s grid, arguing that existing nuclear plants should be kept online and expanded. The hosts contrast U.S. nuclear stagnation with China and South Korea’s faster, cheaper buildouts, blame regulation, litigation, and lost scale for U.S. cost overruns, and propose bipartisan, government-led reforms plus hyperscaler-backed projects and SMRs as the path forward.

Main Topics: Diablo Canyon as a case study for nuclear value (Priority: 5/5): The hosts and Maureen Zawalnik frame Diablo Canyon as a high-performing, low-cost, baseload clean-energy asset that supports California’s grid reliability and local economy, and argue it should not be decommissioned. U.S. nuclear decommissioning vs. extension (Priority: 5/5): They discuss how Diablo Canyon was slated for shutdown in 2016, how changing demand assumptions (AI, electrification, extreme weather) reversed the calculus, and how license renewal could extend operations to 2030 and beyond. China and South Korea’s cost and execution advantage (Priority: 5/5): A major segment compares delivered nuclear costs and timelines, arguing that China and South Korea build AP1000-class reactors far more cheaply and quickly due to scale, repeatability, and lighter regulatory burdens. Regulation, litigation, and institutional inertia in the U.S. (Priority: 5/5): The hosts argue that U.S. nuclear is burdened by dual state/federal permitting, fee-based NRC oversight, lawsuits, and red tape that dramatically slow projects and inflate costs. SMRs, Gen 4 reactors, and innovation pathways (Priority: 4/5): The discussion explores small modular reactors, advanced/Gen 4 designs, naval nuclear experience, and whether modular or factory-built systems can bypass some of the cost problems of large-scale fission. Energy, AI, and national security (Priority: 4/5): They connect nuclear power to AI data-center growth, arguing that cheap, always-on power is becoming a strategic advantage and that nuclear capability affects economic and geopolitical competition. Policy reform and open-source collaboration (Priority: 4/5): The hosts propose Manhattan Project/NASA-style coordination, open-source or IP-free standards, and public-private partnerships to accelerate reactor deployment and reduce risk for utilities and startups.

Key Arguments: Diablo Canyon is one of California’s most important clean-power assets because it provides dense, reliable baseload electricity at roughly 6 cents/kWh. The original shutdown decision made sense under 2015–2016 assumptions, but AI, electrification, drought, wildfire risk, and grid stress changed the system need. U.S. nuclear plants are technically capable of much longer lives; the real constraint is regulatory and political, not physical obsolescence. China and South Korea can build AP1000-class plants at roughly one-quarter the U.S. cost because they have repeatable scale, fewer delays, and more streamlined execution. The U.S. nuclear cost problem is driven by a mix of permitting, litigation, fee-based regulation, labor/material costs, and the loss of a continuous build pipeline. Keeping existing U.S. reactors online is the highest-priority near-term step because shutting them down forces replacement with more fossil generation. SMRs and Gen 4 reactors may help, but only if utilities, government, and large capital providers are willing to underwrite the first-wave risk. AI and data-center demand make nuclear power a national competitiveness issue, not just an energy or climate issue. Open-source standards and shared learnings, similar to naval nuclear operations or Facebook’s Open Compute Project, could lower barriers and reduce vendor lock-in. The U.S. should not be defensive about foreign nuclear progress; global adoption of low-carbon power is beneficial even if other countries execute better.

Data Points: Diablo Canyon reactors: 2 reactors - Current operating units at the site Diablo Canyon planned capacity: 6 reactors - Site was originally provisioned for six reactors Current output: 2.5 gigs - Hosts describe the plant as producing about 2.5 gigawatts of clean power Construction time: 10 years - Maureen says Diablo Canyon took 10 years to build starting in 1970 Build cost: $16 billion - Cost to build both reactors at Diablo Canyon Online year: 1985 - Diablo Canyon went online in 1985 California power share: 8% to 9% - Diablo Canyon’s share of total California energy California renewable share contribution: about 20% - Hosts say Diablo provides roughly a fifth of the state’s renewable energy Diablo power cost: about 6 cents per kWh - Maureen cites the plant’s approximate generation cost Capacity factor: upwards of 90%; 93% to 94% cited - Used to argue nuclear provides highly reliable always-on power NRC license term: 40 years original + 20 years renewal - Standard U.S. nuclear operating license framework Potential lifetime: 60 to 80 years - They discuss successive renewals allowing reactors to operate up to 80 years Diablo current license expirations: November 2, 2024 and August 2025 - Unit 1 and Unit 2 current expiration dates referenced State-approved expiration dates: 2029 and 2030 - California approvals extended the projected shutdown dates Possible further extension: 15 years beyond 2030 - Governorship could extend operating life if federal renewal is approved U.S. operating reactors: 94 - Hosts cite the number of U.S. nuclear plants online U.S. share of world nuclear production: 31% - U.S. share of global nuclear energy production U.S. electricity share: 20% - Nuclear’s share of total U.S. power production U.S. clean energy share: nearly 50% - Nuclear’s share of U.S. clean energy production China build rate: 6 to 8 new reactors per year - Cited as China’s current pace of reactor construction China future additions: 25 to 30 reactors over the next 5 to 6 years - Projected reactor additions in China China/Korea delivered cost: about $2.5 billion per gigawatt - Hosts compare AP1000-class build costs in China Vogtle 3 and 4 cost: about $30 billion for 2.2 GW - Used as the U.S. benchmark for recent Gen 3 construction U.S. cost per gigawatt at Vogtle: over $13 billion per GW - Derived from Vogtle’s total cost and capacity South Korea delivery timeline: about 7 to 8 years - Hosts describe how long AP1000 builds take in South Korea U.S. delivery timeline: about 12 to 15 years - Estimated build time for comparable U.S. projects NRC startup fees: about $60 million to first approval - Startup costs mentioned for new nuclear entrants

Pivotal Quotes: "If someone showed up with the value proposition of fission today, it would blow all of our minds." — Host: Framing nuclear as an underappreciated technology that would look transformative if introduced now "It would be utter madness and scientifically wrong." — Maureen Zawalnik: Her view that shutting Diablo Canyon would be irrational given its reliability and grid value "The plant is upgradable and maintainable, and it's just a matter of keeping it in tip-top shape, meeting all the safety standards, and then reapplying for these licenses." — Host: Summary of the hosts’ new understanding of nuclear plant longevity

Implications: The episode argues that U.S. energy policy should treat nuclear as strategic infrastructure for climate, grid stability, and AI competitiveness. Keeping plants online, streamlining approvals, and scaling new builds could materially improve power availability and industrial competitiveness.