Highlights
- Big Tech has invested over $10 billion in small modular reactors (SMRs) to power AI data centers with 24/7 carbon-free energy.
- No commercial SMR operates in the U.S. or Europe, with construction timelines of 7-10 years instead of the advertised 2-4 years.
- First-reactor costs have exploded to $14,600/kW—five times 2020 projections—and SMRs currently cost 50% more per kilowatt than traditional reactors, undermining Big Tech's energy economics.
- SMRs require large quantities of critical minerals like zirconium, hafnium, and rare earth alloys.
- The U.S. and Europe lack domestic processing of these minerals, while China dominates supply chains, creating choke points similar to those affecting battery metals and permanent magnets.
How about a billion-dollar promise outrunning physics? Big Tech has unleashed more than billion into small modular reactors (SMRs), betting that compact nuclear units will deliver 24/7 carbon-free power for AI data centers. The narrative—factory-built reactors deployed by 2030—is elegant. But as Sydney Westrick, an energy analyst at Vienna Capital Partners, notes in his viral LinkedIn analysis, no commercial SMR operates in the U.S. or Europe, and the gap between projection and reality has widened, not narrowed. Construction timelines are not the advertised 2–4 years—they are 7–10 years.
Table of Contents
First-reactor costs are not falling—they’ve exploded to $14,600/kW, more than five times 2020 projections. And despite decades of nuclear standardization, SMRs currently cost 50% more per kilowatt than traditional reactors, raising fundamental questions about Big Tech’s energy math.
Where The Numbers Crack—and the Critical Mineral Angle Emerges
For Rare Earth Exchanges readers, the overlooked angle is the materials footprint. SMRs require large quantities of high-spec steel, zirconium alloys, hafnium, specialty nuclear-grade tubing, and rare earths (particularly for control systems, sensors, and coolant pumps).
Westrick’s critique is accurate: the economics don’t add up—but he misses the underlying supply chain bottleneck. The U.S. and Europe lack domestic processing for many nuclear-adjacent critical minerals, and China remains dominant in zirconium, hafnium, and rare earth magnet alloys. If SMRs were to scale rapidly—unlikely, but theoretically possible—the world would run straight into the same choke points that already plague battery metals and permanent magnets. That is the hidden structural issue: nuclear modularity still depends on mineral immobility.
Hype, Hope, and the Rare Earth Reality Check
Westrick’s skepticism is well-founded. Argentina’s SMR is 600% over budget; China’s and Russia’s designs are years behind schedule; and the U.S. NuScale project collapsed under its own economics. His analysis leans pessimistic, but not incorrectly so. Where bias creeps in is the implication that SMRs are destined to fail entirely—historically, early nuclear models always blew deadlines before later generations stabilized.
The misinformation risk: suggesting SMRs are a solved technology simply awaiting Big Tech enthusiasm. They are not. But neither are they fantasy. The honest framing is this: SMRs may transform energy—but not on AI’s timeline. And in the meantime, Big Tech is powering data centers with current energy sources—natural gas (over 40%) followed by renewables (about 24%), nuclear (around 20%), and coal (about 15%). Natural gas is a major contributor because it is reliable, scalable, and can be quickly dispatched to meet the high energy demands of data centers.
© 2025 Rare Earth Exchanges™ – Accelerating Transparency, Accuracy, and Insight Across the Rare Earth & Critical Minerals Supply Chain.
0 Comments