Claims

Gen 4 SMRs are inherently safe with walk-away shutdown capability, making meltdowns physically impossible based on fuel and reactor design, unlike older generation reactors.

Doomberg is less enthusiastic about fusion than at founding because fission solves all the problems anyway; fusion is a distraction from pursuing deployment-ready fission technology.

Small modular reactors (SMRs) operating at Gen 4 designs can provide steam at 500-1000 degrees Celsius, covering a significant portion of industrial base-load heat demand currently supplied by coal, natural gas, and oil; this addresses carbon-intense industrial steam requirements that most people don't realize exist.

Intermittent renewables with low energy density, questionable energy payback periods, and dependence on China for key supply chain components (polysilicon, cells, wafering, machines) represent the totality of the green energy agenda's practical alternatives to fossil fuels and nuclear.

Energy demand will increase due to AI and other technologies, and global electrical generation capacity will need to expand from current levels (estimated at 3,000-4,000 GW); fusion is likely to eventually dominate electrical generation globally as the cheapest, most abundant, reliable, and densest energy source, though solar plus storage is also a viable path.

Before the war, Ceraweek 2026 would have been dominated by discussions about powering AI data centers, with major tech company executives presenting, but the crisis shifted focus back to oil and gas alongside electricity generation questions.

Managing an electricity grid with much more intermittent renewable energy is technically solvable but requires intentional planning; the transition from fossil-dominated stable baseload generation to renewable intermittency is a major infrastructure challenge.

Jet fuel cannot be easily replaced by renewable energy investments because mobility and refining require upstream hydrocarbon sources and refining capacity that take years to develop.

Neither Canada nor the US will run out of oil on any timeframe that matters to anyone listening, because of technological advancement; peak oil is a myth that reflects a sucker's bet against human ingenuity.

Just three to four years before the shale revolution, serious analysts claimed the US would never economically produce viable natural gas from the Marcellus Shale; now it is producing at roughly twice Qatar's rate from a single deposit, demonstrating the folly of betting against technological progress.

Natural hydrogen is a potential game changer on par with the shale revolution; if hydrogen can be drilled from the ground like oil, it would be a real GameChanger because it can be burned cleanly without producing carbon dioxide.

Microsoft's Stargate 5 GW reactor will draw power equivalent to five New York boroughs' consumption 24/7, supporting a single 5 GW data center, illustrating the massive power demands of AI infrastructure and why nuclear is the only baseload solution.

The uranium fuel cycle consists of seven stages: mining/extraction, milling (crushing to yellow cake), conversion (to UF6 gas), enrichment (cascading centrifuges raising U-235 from 0.71% to 3-5%), fabrication (into pellets and fuel rods), and use in reactors; each stage is dominated by a small number of countries with high regulatory barriers.

Modern nuclear plants would be 'light years more sophisticated' than those built in the 1970s that are still operating, offering significantly better safety and efficiency, making the decades-long moratorium on new nuclear builds particularly wasteful.

Battery costs have declined by 90% since 2010, and the levelized cost of electricity (LCOE) of solar plus batteries is expected to be cheaper than gas power in the US by 2023.

Solar and wind made up more than 75% of all new global capacity additions in 2022 and now represent more than 10% of global power generation.

A major constraint on AI data center natural gas adoption is corporate carbon emission concerns; however, if Microsoft and Google decline to use natural gas for environmental reasons, other companies and countries will do so, suggesting carbon consciousness will not prevent natural gas utilization.

Natural gas offers significant advantages for powering AI data centers: it is very cheap at approximately $2.75 per million BTU (equivalent to $15-16 per barrel oil), burns cleanly, and can be brought online quickly, enabling co-located closed-loop systems where natural gas input and data output eliminate grid connection requirements.