The U.S. Department of Energy just handed Oklo and four other nuclear startups a deal that sounds too good to be true: take our surplus weapons-grade plutonium and turn it into fuel for your advanced reactors. If you're wondering why that matters for your investment portfolio, the answer is simple: AI data centers are running out of power.

Oklo announced it's been selected for advanced negotiations under the DOE's Surplus Plutonium Utilization Program. The program aims to convert designated surplus plutonium material into fuel for advanced nuclear reactors. In partnership with European developer newcleo, Oklo would lead the utilization while newcleo brings fuel experience and up to $2 billion in potential project capital.

Let's translate that from government-speak. The U.S. has been sitting on plutonium that was set aside for disposal, which is expensive and politically complicated. Instead of paying to bury it, the DOE is letting companies use it as bridge fuel for next-generation reactors. Material that was a liability becomes a revenue-generating asset.

Here's why this matters beyond the nuclear industry. Every major tech company is racing to build AI infrastructure, and they're all hitting the same wall: power. Training large language models and running inference at scale requires staggering amounts of electricity. Traditional grid power can't scale fast enough, and renewables remain intermittent.

Nuclear has always been the obvious answer for baseload AI data center power, but fuel supply has been a constraint on deployment speed. Programs like this one create alternative pathways to source fuel while domestic enrichment infrastructure scales up. That potentially accelerates the timeline for advanced reactor deployment.

Oklo's CEO Jacob DeWitte stated it plainly: "Fuel supply constraints are a key throttle to advanced reactor development. This program creates a pathway to use existing surplus material as bridge fuel for advanced reactors to bring more reactors online sooner."

The investment angle here is infrastructure, not environmentalism. Companies like Microsoft, Google, and Amazon are already signing power purchase agreements with nuclear developers. They need reliable, carbon-free electricity at data center scale, and they need it within this decade, not the next one.

Oklo went public via SPAC in 2024 and has been volatile, like most early-stage nuclear plays. The company is pre-revenue and faces significant regulatory hurdles. But programs like the plutonium utilization pathway de-risk one of the major bottlenecks: fuel supply.

The newcleo partnership adds another dimension. newcleo initiated pre-application engagement with the Nuclear Regulatory Commission in February 2026 for both an advanced fuel fabrication facility and a lead-cooled fast reactor design. That's a transatlantic supply chain being built in real time.

Should you buy Oklo stock tomorrow morning? Hold on. This is a long-term infrastructure play with significant execution risk. Regulatory approval for advanced reactors is measured in years, not quarters. The $2 billion investment commitment is subject to definitive agreements and customary approvals, which is code for "this might not happen."

But the broader thesis is compelling. AI power demand is real, growing, and not going away. Nuclear is the only proven technology that can deliver clean baseload power at the scale these data centers require. And fuel supply has been the chokepoint. Programs like this one remove obstacles.

If you're looking for exposure to the AI infrastructure build-out beyond chips and cloud services, nuclear fuel and power generation is an under-followed corner of the market with actual fundamental drivers. Just don't confuse a DOE program selection with guaranteed success.