Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.
Back to Podcast Digest
A Nuclear Masterclass: The Past, Present, and Future In One Conversation
TL;DR
America didn’t stop getting nuclear power — it stopped building new plants — James Karenstein says U.S. reactor count peaked around 112 and is now about 94-95, but capacity factors rose from roughly 55% in 1975 to more than 90% today, effectively doubling output from the same fleet.
The real collapse started before Three Mile Island — Nixon’s 1973 dream of 1,000 reactors by 2000 ran into inflation, Paul Volcker-era interest rates, labor shortages, slower electricity-demand growth, and regulatory chaos; by the time Three Mile Island happened in 1979, new orders had already dried up.
China’s nuclear surge is built on old-school reactor discipline, not startup novelty — China is scaling pressurized water reactors derived from Westinghouse and French designs, with about 40 reactors under construction, and Karenstein says they’ll likely pass the U.S. within five or six years.
France won by abandoning ‘advanced’ reactors and just copying what worked — Karenstein’s favorite historical lesson is that France ditched its indigenous gas-cooled design, licensed proven Westinghouse reactors, standardized them through EDF and Framatome, and used that repetition to decarbonize its grid fast.
Karenstein thinks most SMR hype ignores basic physics — his blunt argument is that thermal power systems naturally favor larger plants, and once you add shielding and nuclear safety requirements, many small modular reactor pitches look like trying to power the grid with “thousands of Chevy Tahoe engines.”
His company’s bet is to squeeze 6-10 GW from existing U.S. nuclear plants before building new ones — Alva Energy wants to add a second turbine system and upgrade steam production at current reactors, claiming 200-300 MW per reactor at about one-seventh to one-tenth the cost of an SMR, with first deployment potentially within three years.
The Breakdown
Nixon’s 1,000-reactor dream and the reality check
Ashley Vance opens with the wild contrast: in the late 1960s and early 1970s, the U.S. imagined 1,000 reactors by 2000; instead it topped out a little above 100 and now sits around 94 or 95. He frames the episode as a badly needed reality check for a moment when SMRs, fusion startups, Trump-era nuclear enthusiasm, and AI energy demand are all colliding.
The U.S. didn’t build much — but it learned to run reactors far better
Karenstein immediately complicates the “nothing happened” story. The last truly new U.S. plant that started construction and actually finished before Vogtle was Sharon Harris Unit 1, begun in 1978, but the fleet got dramatically better at operating: average nuclear capacity factor went from about 55% in 1975 to over 90% today. His point is that the industry spent decades learning how to run plants cheaply, safely, and reliably, even while new construction stagnated.
China’s rise is really a story about disciplined copying
On China, Karenstein is almost annoyingly precise in the best way: their current buildout is dominated by proven pressurized water reactors, including the CAP1000/AP1000 line from Westinghouse and the Hualong One, which traces a lineage from Westinghouse to France to China. He notes that the first AP1000s ever completed were in China, not the U.S., after Westinghouse transferred major technology there in the mid-2000s — “maybe not America’s finest hour.”
Why France became the nuclear outlier
Karenstein lights up when he tells the French story. France had its own “advanced” gas-cooled reactor program, then made the cold-blooded decision to abandon it, license Westinghouse pressurized water reactors, and build the same thing over and over through EDF and Framatome. The lesson he keeps hammering: don’t get distracted by fancy tech when a proven design can be standardized, vertically integrated, and repeated across dozens of sites.
What actually killed the U.S. buildout
Three Mile Island mattered, but Karenstein says it was “the cherry on top” of a breakdown already underway. Nuclear plants got hammered by inflation, rising commodity and labor costs, high interest rates, a fragmented utility system, and slower-than-expected electricity demand growth after 1973; utilities were suddenly betting the company on giant capital projects they no longer clearly needed. By his telling, it wasn’t one disaster but a whole stack of economic and institutional failures.
Nuclear power, climate politics, and the environmentalist contradiction
The conversation gets spicy here. Karenstein argues that much of the anti-nuclear movement fused civilian power with anti-weapons politics and later wrapped climate rhetoric around positions that still opposed the one technology that had already decarbonized whole countries, pointing to France and Germany as the contrast. His sharpest line is that nuclear is a “cheat code” for climate: you can use huge amounts of energy without changing your lifestyle, because the grid just gets cleaner underneath you.
Vogtle: the comeback project that proved how hard this still is
Vogtle Units 3 and 4 are the centerpiece of the modern U.S. revival — and also a warning. Karenstein says the AP1000 design wasn’t finished when construction began, the supply chain wasn’t ready, and the workforce had lost much of its new-build muscle after three decades; the result was a more-than-$30 billion project that came online years late, in 2023 and 2024. Southern Company gets real credit in his view for gutting it out when other utilities, like VC Summer’s backers in South Carolina, collapsed or worse.
Why he’s skeptical of SMRs — and what Alva Energy wants instead
This is the episode’s most contrarian turn. Karenstein says the industry already tried small reactors in the 1950s, then rapidly scaled up because thermal systems favor size; his analogy is that nobody powers the grid by stacking thousands of car engines together, even though cars are mass-produced brilliantly. Alva Energy’s alternative is less flashy and more surgical: use upgrades and a proprietary second-turbine setup to pull 200-300 more megawatts from existing reactors, potentially unlocking 6-10 GW across the U.S. fleet while building the execution muscle to eventually construct new large reactors the right way.
Share
Keep Reading
Two Ways to Follow
The Weekly Echo. The inbox-shaped summary of what mattered.
New editorials announced here.
