The Hidden Truth About the US Energy Grid: Subsidies, Demand & Rising Prices - Chris Wright
Summary
Wright argued hydrocarbons will remain the dominant global energy system because they supplied 85% of global energy in both 1973 and 2024. Natural gas grew at 3% annually over that period, versus 2% for coal and 1% for oil; nuclear now supplies 4%, while wind, solar, and batteries together contribute less than 3%.
Nuclear’s extraordinary fuel economics are being overwhelmed by permitting, overengineering, and public fear rather than technical limitations. Chamath framed the prize as roughly $12,000 of uranium fuel producing $3 million-$4 million of electricity daily in a 1 GW plant; Wright called nuclear the safest energy source and said government had “killed nuclear over the last 4 decades.” China’s advantage, he argued, is a more pragmatic safety regime and faster construction.
Wright attributes rising US electricity prices to paying for intermittent generation while retaining the traditional grid needed for peak demand. Natural gas supplies 43% of US electricity, nuclear about 20%, coal 15%-16%, and those sources plus hydro roughly 83%; on a cold winter evening, Wright said wind, solar, and batteries may provide only 2%-3%. A 4-cent-per-kWh federal wind subsidy can exceed the 2-cent avoided gas cost in Texas, leading to his deliberately combative conclusion: “If you’re not dispatchable… you’re just a parasite.”
The sharpest disagreement was whether solar is being unfairly dismissed or merely assigned an exaggerated role. Chamath pointed to China installing two to three times more solar per capita, while Wright offered a 50-year bet that solar “never gets to 10% of global energy” and stressed that electricity itself supplies only 20% of total energy. He still allowed a role for solar, citing a firmed 1-gigawatt UAE facility using 8 gigawatts of panels and extensive batteries. Chamath split the difference: remove distorted subsidies, but expect America’s largest utility eventually to be a distributed network of homeowners using solar and batteries.
Wright accepts climate change as real but ranks energy poverty far above it, a hierarchy with major policy implications for development finance. He said atmospheric CO2 has risen 50%, yet characterized projected end-of-century income damage as perhaps months to a year of foregone growth over 75 years, compared with 2 million-3 million preventable annual deaths from indoor wood burning. Chamath asked whether methane, roughly 80 times more heat-trapping than CO2, and even 1% leakage could erase natural gas’s advantage; he also noted that without leakage gas is probably better, perhaps with half the footprint. Wright did not directly answer the leakage question.
Natural gas is Wright’s near-term answer to AI data-center demand, while regulatory changes could unlock gigawatts from assets already installed. He cited retail electricity increases of 25% and wholesale increases above 50% during the Biden administration, acknowledged turbine supply constraints, and proposed allowing peaker plants and backup gas or diesel generators to run during the few annual peak hours. “We have slack capacity 98% of the time.”
DOE plans to pair federal land and accelerated permitting with preserved national-lab research capacity. Wright expects a next-generation small modular reactor to reach criticality on federal land by July 4 next year—though it would not yet sell power into the grid—and said an RFP for lab-land data centers drew 300 responses. At the same time, after cutting more than $500 billion of energy subsidies, he said he wanted to cut the full $1 trillion but vowed to defend the roughly $10 billion annual budget for 17 national labs: “I will succeed at that.”
Deep dive
1. Hydrocarbons still supply the energy system policymakers expected to replace
Wright’s historical anchor was 1973, when the oil shock prompted calls to get off oil and gas: oil, gas, and coal supplied 85% of global energy then—and still supplied 85% in 2024. Natural gas compounded at 3%, coal at 2%, and oil at 1%.
Wright distinguished the fuels by use case: oil is the most expensive and flexible, coal is cheapest without infrastructure, and natural gas can be cheapest with infrastructure—helping explain why gas is growing fastest.
Nuclear fell from 6% of global energy in 2000 to 4% today; wind, solar, and batteries together remain below 3%. Traditional biomass is twice their combined contribution.
Chamath emphasized the human consequence: 2 billion people still cook or heat with indoor wood fires, causing 2 million-3 million preventable deaths annually. His preferred intervention is not an abstract transition but “just a simple propane stove.” Wright later used the same clean-cooking problem in his “humans first” comparison.
2. Nuclear’s constraint is institutional fear, not fuel economics
Chamath’s framing captured nuclear’s energy density: a 1 GW reactor can turn roughly $12,000 of uranium fuel into $3 million-$4 million of daily electricity. Why, then, does construction take years and cost billions?
Wright’s answer: invisible radiation is easy to weaponize politically, producing slow permits, restricted enrichment, and excessive engineering. Although nuclear is “the safest form of energy production we’ve ever seen,” government has “strangl[ed]” it for four decades.
China, he argued, designs around human safety rather than satisfying layers of political opposition, so it builds faster and cheaper. Wright cited more than 20 reactors under construction; Chamath countered with 33 underway and 200 planned while pressing why America keeps foregrounding coal.
3. Intermittency, not generation cost alone, drives Wright’s grid critique
US electricity now comes 43% from natural gas, about 20% from nuclear, and 15%-16% from coal; with hydro, those traditional sources total roughly 83%. Wright contrasted today’s increases with a century of declining inflation-adjusted electricity prices.
His cost mechanism is additive: intermittent sources require new transmission while traditional sources must remain available for peak demand. Peak demand is a cold winter evening, when an outage can be fatal; Wright cited more than 200 deaths from a Texas electricity outage. Wind’s 4-cent-per-kWh federal subsidy can exceed Texas’s roughly 2-cent avoided gas cost, before utility payments or state mandates.
Chamath’s pushback—worth keeping—was that China installs two to three times as much solar per capita while America praises “clean, beautiful coal.” Wright replied that China also built 100 coal plants last year and offered a 50-year wager that solar never reaches 10% of global energy.
Wright still assigned solar a role, citing a firmed 1-gigawatt UAE facility using 8 gigawatts of panels and extensive batteries, but argued that solar cannot supply the other 80% of global energy demand, including processed heat and fuels for jets and ships.
Chamath reframed the divide around broken utility economics: transmission and distribution can overwhelm cheap generation, eventually pushing households to self-supply with solar and batteries. He applauded ending subsidies that supported “the shadiest of companies”: “Let the best companies compete.”
4. Climate risk is acknowledged, but subordinated to development and security
Chamath asked whether methane’s roughly 80-times greater heat retention and even 1% supply-chain leakage could make natural gas worse than coal or oil. He also said that without leakage, gas is probably better and may have half the footprint. Wright did not resolve that arithmetic; he pivoted to CO2, accepting its 50% atmospheric increase and contribution to warming.
Wright cited climate-economic projections ranging from roughly 0.2% to perhaps 3%-4% lower per-capita income by century-end—“a couple months” to one year of growth lost over 75 years. His hierarchy reverses the previous administration’s: humans first, energy as enabler, climate in “the appropriate role.”
Chamath’s synthesis was strategic rather than climatic: China invests in every energy source because it cannot risk foreign dependence. America should likewise treat energy as “a national security issue,” not cherry-pick whichever technology confirms an ideological bias.
5. DOE wants federal assets to accelerate AI power without gutting science
Wright defended all 17 national labs and their roughly $10 billion annual budget, citing MRI machines alongside open-ended work on dark matter and neutrinos. After cutting more than $500 billion in energy subsidies—and saying he wanted to cut the full $1 trillion—he said he would defend the labs and “succeed” in stopping cuts.
For AI data centers, natural gas is “by far” the cheapest and fastest scalable source, though turbine supply is constrained. Wright cited 25% retail and more than 50% wholesale electricity-price increases during the Biden administration as the inherited backdrop.
His quickest capacity unlock is regulatory: grids have slack 98% of the time but need additional electrons during a few peak hours. Allowing restricted peakers and backup diesel or gas generators to sell power then could add gigawatts without waiting for entirely new infrastructure.
DOE’s lab-land data-center RFP drew 300 responses, with fast federal permitting intended to prevent consumer-rate spillovers. A small modular reactor should reach criticality at Idaho National Laboratory by July 4 next year; it would demonstrate that it can sell electricity but would not yet be selling power into the grid. Former DOGE personnel—now effectively rebranded—remain embedded across DOE operations.