$NVDA - Valor Atomics powered the first AI chip (NVIDIA Blackwell) directly from a nuclear reactor, demonstrating nuclear's capability to meet AI compute power demands. The company is hosting their website directly on the NVIDIA chip powered by their reactor.
$NUCLEARSMR - Small modular reactors (SMRs) represent a shift from civil infrastructure to advanced manufacturing approach in nuclear, which aligns with US capabilities. Valor demonstrates this is achievable through hardware iteration rather than just modeling.
$ENERGY - Energy demand is fundamentally driven by price - making energy cheaper creates its own demand. Nuclear has potential to make energy 10x cheaper, which would unlock massive new demand and applications.
$POWER - Power infrastructure is a critical bottleneck for AI compute growth. Building gigasites with abundant cheap power will attract data center load, creating a 'build it and they will come' dynamic.
$ROBOTICS - AI and robotics will convert human labor inputs to energy inputs in manufacturing, making energy the fundamental cost of all physical goods. As energy gets cheaper, manufactured goods approach free.
Bearish:
$TRADNUCLEAR - Traditional nuclear companies are modeling and simulation companies, not hardware execution companies. They focus on complex designs rather than simple, scalable reactors, and charge inflated prices (e.g., $5M for systems that can be built for $400k).