Powering the AI Age with Atoms
The numbers driving the nuclear renaissance
AI is the most power-hungry technology humanity has ever built
AI datacenters consumed an estimated 260 TWh of electricity in 2026 -- roughly equal to the entire annual consumption of Spain. By 2030, that figure is projected to reach 945 TWh, a 3.6x increase in just four years. Every ChatGPT query uses ~10x the electricity of a Google search. Every AI training run demands megawatts for weeks.
Solar and wind cannot serve as baseload power. They are intermittent by nature -- the sun sets, the wind stops. Battery storage at datacenter scale remains economically prohibitive. Natural gas is the bridge fuel, but it comes with carbon emissions and price volatility.
Nuclear is the only proven zero-carbon baseload power source that can operate 24/7 at scale. A single 1GW reactor produces ~8 TWh per year with a 90%+ capacity factor. To meet the projected 685 TWh of incremental datacenter demand by 2030, the world needs the equivalent of ~85 new gigawatt-scale reactors. The race is on.
The hyperscalers are betting billions on atomic power
$80B+ capex in 2025 alone. Signed a 20-year PPA with Constellation Energy to restart Three Mile Island Unit 1 -- a 1.2GW deal, the largest corporate nuclear PPA ever. Nuclear is central to Microsoft's carbon-negative-by-2030 pledge.
Signed a PPA with Kairos Power for small modular reactor (SMR) deployment. First reactor planned for 2030. Google is also exploring advanced geothermal and next-gen nuclear to power its AI infrastructure at scale.
Acquired a $500M nuclear-powered datacenter campus from Talen Energy adjacent to the Susquehanna nuclear plant. Also investing in X-energy SMR development and exploring additional nuclear site acquisitions.
Issued an RFP for 1-4 GW of new nuclear generation capacity. Meta's AI training clusters require sustained multi-gigawatt power -- intermittent renewables alone cannot meet the load profile.
Demand is accelerating while supply contracts
Uranium spot prices have risen from $20/lb in 2020 to $77-80/lb in early 2026 -- a nearly 4x increase. Long-term contract prices are approaching $86/lb. The driver is simple: structural supply deficit.
Annual global uranium demand is approximately 180 million pounds. Current mine supply is roughly 140 million pounds. The deficit has been filled by secondary supplies -- utility inventories, government stockpiles, and underfeeding enrichment -- but these sources are depleting rapidly.
Kazatomprom, the world's largest uranium producer (~45% of global output), cut production guidance by 17% due to sulfuric acid shortages and construction delays. Meanwhile, 440 reactors operate globally with 65 more under construction. Every new reactor adds ~500,000 lbs of annual uranium demand. The supply gap is widening.
~40M lb annual shortfall growing. Kazatomprom cuts. Mine development takes 10-15 years. Secondary supply sources depleting. New reactor demand accelerating. Long-term contracts locking in higher floors.
Kazatomprom could reverse production cuts. New mines in Canada and Africa in development. Enrichment underfeeding can stretch supply. Uranium recycling and reprocessing expanding. Spot price already 4x from 2020 lows.
Bipartisan nuclear support from Washington to the world stage
Streamlined NRC licensing for advanced reactors. Reduced review timelines from 42+ months to target 25 months. Fee reductions for small companies. The most significant nuclear regulatory reform in decades.
$900M+ allocated for High-Assay Low-Enriched Uranium (HALEU) production. HALEU is the fuel required by virtually every advanced reactor and SMR design. Without domestic HALEU, the US SMR industry cannot scale.
$30/MWh production tax credit for nuclear power under the Inflation Reduction Act. This makes existing nuclear plants highly profitable and improves the economics of new builds. The credit applies to all zero-emission electricity.
22 nations signed a pledge to triple nuclear energy capacity by 2050 at COP28. Signatories include the US, UK, France, Japan, South Korea, and Canada. This is the strongest multilateral nuclear commitment in a generation.
The companies powering the atomic renaissance
From uranium mine to datacenter -- the nuclear ecosystem
Fission wins 2025-2035. Fusion is the 2035+ option.
While fission is the proven, deployable technology for the next decade, fusion is attracting unprecedented private capital. 43 private fusion companies are now operating globally, with over $7B in total private investment.
SPARC tokamak targeting net energy gain. Backed by $2B+ in funding. HTS magnets represent a breakthrough in compact fusion. ARC commercial reactor design targets the 2030s.
Raised $5B+ -- the most-funded private fusion company. Signed a PPA with Microsoft for fusion electricity by 2028. Uses a field-reversed configuration (FRC) approach that directly converts fusion energy to electricity.
$1.2B raised with a hydrogen-boron (p-B11) aneutronic fusion approach -- no radioactive waste. If successful, this is the cleanest possible energy source. Timeline extends into the 2030s.
The investment thesis is clear: fission is the bridge, fusion is the destination. The fission companies (CCJ, LEU, BWXT, CEG) generate revenue and cash flow today. Fusion remains pre-revenue and speculative, but the potential market is unlimited -- literally the energy source of the sun. Smart capital allocation means owning fission now and watching fusion carefully.
Positioning all tickers by risk and reward potential
OKLO and NuScale are years from meaningful revenue. OKLO's NRC application was rejected once. NuScale lost its UAMPS project. Execution risk is severe -- these are capital-intensive, decade-long development cycles with regulatory uncertainty.
The Nuclear Regulatory Commission reviews only a handful of applications at a time. Even with the ADVANCE Act, permitting new reactor designs takes years. The NRC is understaffed for the volume of SMR applications expected.
If Kazatomprom reverses its production cuts or new mines come online faster than expected, uranium prices could correct significantly. The current spot price already reflects considerable optimism about future demand.
OKLO trades at $15B+ with zero revenue -- a pure narrative stock. If NRC permitting delays extend or the Meta deal structure disappoints, the valuation has no earnings floor. Public sentiment on nuclear remains mixed despite improving trends.
Key nuclear inflection points from 2026 to 2035+