Constellation CEO stated that long lead-time equipment (large pumps and motors) has already been ordered, which is critical because manufacturing capacity for these specialized nuclear-grade components has largely left North America and must be rebuilt or sourced from overseas.

Fukushima was a watershed event that forced the nuclear industry to move beyond design basis standards to 'beyond design basis' planning, accounting for worst-case scenarios that exceed historical precedent; this has made modern reactors far safer than they were before.

Nuclear power plants must shut down for refueling (typically every 18 months to annually), meaning Microsoft will need backup power or a second reactor to avoid data center downtime, which significantly increases the operational complexity and cost of the arrangement.

China is currently building approximately 20 nuclear reactors simultaneously, demonstrating industrial capacity that the US lost over the past 30 years of minimal nuclear construction; the US must somehow rebuild this capability in a few years to compete.

When all existing uranium mines are fully contracted out (all their long-term production sold), spot price will spike because utilities and new reactors will have no choice but to buy on the spot market at whatever price is required; this is the key trigger event for uranium price appreciation.

China has built strategic uranium stockpiles and storage facilities on the Kazakhstan border to ensure long-term energy independence; this represents sophisticated energy security thinking unavailable to the West.

OpenAI has requested 35 GW of electrical capacity from the US government, equivalent to 5-7 data centers of 5 GW each, which represents 25-35% of current total US nuclear electricity production and demonstrates a demand scale that dwarfs existing supply capacity.

TMI-1 currently has a 60-year operating license, is already 50 years old with only 10 years of license life remaining, so Constellation must apply for a license extension to 80 years as part of the $1.6 billion refurbishment to support the 20-year Microsoft contract.

Data centers cannot operate with solar or wind power because they require 24/7 high-quality, stable electricity with no voltage reductions or frequency variations, which only nuclear can reliably provide at scale.

Producing 35 GW of capacity from wind requires 9,375 square miles of land; from solar requires 1,350 square miles; from nuclear requires 54 square miles—a ~170x difference in land efficiency between nuclear and wind.

If Three Mile Island replaces 50% of its core with new fuel and increases enrichment to extend the fueling interval from 12 months to 18 months, uranium demand would be roughly 3x the annual requirement of a steady-state reactor, creating a significant spike in near-term uranium purchases.

Uranium mining in remote locations (Northern Saskatchewan, Utah desert) faces persistent labor supply challenges because workers do not want to relocate or work on fly-in/fly-out schedules; this structural problem has existed for 50 years and will cause wage inflation, making mines less profitable unless uranium prices rise.

Uranium spot price and long-term contract price have historically diverged significantly; spot price is volatile due to trading and speculation, while long-term contract price reflects actual utility purchasing patterns and is a better indicator of true supply-demand fundamentals.

Germany's energy transition away from nuclear toward renewables has resulted in higher energy prices and industrial competitiveness problems; companies are relocating to other countries with cheaper, more reliable power.

The uranium market is illiquid and small compared to oil and gold markets; large institutional investors (pension funds, mutual funds) will not invest until the market has greater liquidity and scale, which will only happen as major capital flows into the nuclear supply chain.

SMRs (small modular reactors) face a fundamental economics challenge: building many smaller reactors costs more per unit of capacity than building one large reactor, and mass production at the scale needed (1,200+ SMRs for 35 GW) has never been demonstrated.

Rolls-Royce in the UK is positioning to mass-produce SMRs with contracts from Poland, Czech Republic, and UK National Grid, but no SMR has been built and operated at scale anywhere, creating technology and production risk.

There are only four SMRs under construction in North America (Darlington in Canada), none of which have been completed or operated, representing zero proven operating track record for SMR technology.

Some SMR designs use exotic fuel types like TRISO fuel, which have been tested in research reactors but never operated in commercial power reactors, creating uncertainty about real-world fuel performance.

Most countries recognize that large-scale nuclear deployment is in their national interest; the only major exception is Germany, but even Germany will likely reverse course as energy security concerns mount.