Oklo: Building the Nuclear Infrastructure Layer for the AI S-Curve

The paradigm shift is no longer theoretical. It is a seismic change in the fundamental demand for electricity, and it is accelerating along an exponential curve. For decades, U.S. commercial electricity demand was essentially flat. That era has ended. Through the first nine months of 2025, that demand is up 2 percent year-over-year, following a 3 percent increase in 2024. This isn't a blip; it's the early, undeniable signal of a new technological S-curve taking hold.

The catalyst is artificial intelligence. The consulting firm Grid Strategies projects that data centers alone will account for 55 percent of the growth in U.S. electricity demand over the next five years. This isn't just incremental growth. It is a structural reordering of the power sector's load profile, driven by the immense compute power required to train and run generative AI models. The numbers are staggering. U.S. power demand is expected to surpass 4,200 billion kWh in 2026, a clear inflection point from the 2024 record. Globally, the forecast is even more dramatic. Goldman Sachs Research projects that global data center power demand could more than double by 2030.

Meeting this surge requires building new infrastructure at a pace that matches the adoption curve. The scale is immense. Goldman Sachs estimates that to meet all of that data center power demand growth by 2030, the world would need 85-90 gigawatts of new nuclear capacity. That is a massive build-out, and it underscores a critical vulnerability: the current energy mix is not on track to deliver it. Well less than 10% of that required nuclear capacity will be available globally by 2030. This creates a window of opportunity for the companies that can construct the fundamental rails for this new paradigm. The investment thesis is clear: the infrastructure layer for the AI era is not yet in place, and the first movers will be those that can scale the necessary technologies-like advanced nuclear-fast enough to ride the exponential adoption curve.

Oklo's First-Principles Infrastructure Play

Oklo's strategy is a textbook example of building the infrastructure layer for a new paradigm. The company is not chasing incremental efficiency gains; it is deploying a technology designed from first principles to solve the core constraints of the AI power S-curve: safety, fuel sustainability, and rapid deployment. Its fast reactor design is the bedrock of this approach. This technology is not theoretical. It has been built and operated before, with over 400 years of cumulative operating experience worldwide. This proven track record, from the pioneering Experimental Breeder Reactor-I to the decades-long operation of EBR-II, provides a crucial foundation of reliability and scalability. In the context of an exponential demand curve, this isn't just a safety feature-it's a critical infrastructure advantage that reduces execution risk.

The company's flagship product, the Aurora powerhouse, embodies this infrastructure-first thinking. These are small, fast-spectrum reactors engineered for inherent safety. They are self-stabilizing, self-controlling, and cooled by natural forces, meaning they are walk-away safe. More importantly, they are uniquely able to recycle used nuclear fuel. This transforms a waste stream into a fuel source, directly addressing a key bottleneck in scaling nuclear power. For a company building the rails for the AI era, this closed-loop capability is a fundamental feature, not a bonus.

This technological alignment is now being backed by a major market commitment. In January, Oklo announced a landmark agreement with Meta to develop a 1.2 GW power campus in Pike County, Ohio. Meta's prepayment provides critical funding to secure nuclear fuel and advance Phase 1 of the project. The scale is significant: the development will occupy 206 acres of land in a strategic location within the PJM grid. This isn't just a power purchase agreement; it is a pre-commitment to a multi-year build-out, with the first phase targeted for 2030 and the full capacity expected by 2034. The deal provides the project certainty Oklo needs to move from demonstration to deployment.

The bottom line is that Oklo is positioning itself as the fundamental infrastructure layer for the next energy paradigm. It combines a proven, inherently safe reactor design with the unique ability to use nuclear waste as fuel. This combination, now backed by a major anchor tenant like Meta, aims to deliver clean, reliable power at the scale and speed required by the AI S-curve. The company is building the rails, not just the locomotive.

Execution Risk vs. Exponential Growth Potential

The exponential growth thesis for Oklo is built on a simple, powerful equation: the AI power S-curve is accelerating, and its infrastructure rails are not yet laid. The company's first commercial reactor, the Aurora-INL, is the critical first step in proving the build-out can happen. Its target timeline is late 2027 to early 2028. This is the foundational milestone. Success here de-risks the entire deployment model; failure would severely undermine the credibility of scaling to the multi-gigawatt levels required.

The scale challenge is immense and defines the execution risk. Goldman Sachs Research projects that 85-90 gigawatts of new nuclear capacity would be needed globally by 2030 to meet all data center power demand growth. Yet the firm forecasts that well less than 10% of that required capacity will be available globally by 2030. This creates a massive gap. Oklo's Pike County project is a direct attempt to fill a tiny sliver of that gap. The deal with Meta is a pre-commitment to develop a 1.2 GW power campus on 206 acres of land in southern Ohio. The first phase is targeted for 2030, with the full capacity expected by 2034. This is a multi-year, multi-billion dollar build-out that must proceed on schedule.

The critical path for this project is clear. Oklo must use Meta's prepayment to secure nuclear fuel and advance Phase 1 of development. This includes pre-construction and site characterization, slated to begin in 2026. The company's ability to manage this capital efficiently, navigate permitting, and execute the build-out on the stated timeline will determine if it can transition from a technology demonstrator to a scalable infrastructure builder. The risk is not just technical-it is the familiar, daunting challenge of scaling any complex industrial project in a capital-intensive, regulated sector. The exponential growth potential is vast, but it is entirely contingent on Oklo's ability to deliver its first reactor on time and then replicate that success at an unprecedented pace across the 85-90 GW gap.

Catalysts, Scenarios, and the Path to Singularity

The investment thesis for Oklo hinges on a series of near-term milestones that will validate its ability to scale. The primary catalyst is the tangible progress on the 1.2 GW Meta project in Ohio. The company has already secured the land and the prepayment, but the critical path now begins in 2026 with the start of pre-construction and site characterization. Success here will demonstrate Oklo's operational execution and its capacity to navigate the complex regulatory and logistical hurdles of a multi-year build-out. Any delay or setback in this phase would directly challenge the project's timeline, which targets the first phase for 2030.

Beyond this anchor deal, the market signal to watch is the emergence of additional corporate power purchase agreements (PPAs) from tech giants. Goldman Sachs Research notes that several big tech companies have already signed contracts for new nuclear capacity, signaling a growing market demand for low-carbon, round-the-clock power. The pattern of these follow-on deals will be a key indicator of broader market adoption. If Oklo can leverage its Meta partnership to secure similar commitments from other data center operators, it would confirm that its infrastructure model is being recognized as a viable solution to the AI power crunch.

The foundational catalyst, however, remains the successful operation of its first commercial reactor, the Aurora-INL. This reactor is the essential proof point for the entire deployment S-curve. Its target timeline of late 2027 to early 2028 is non-negotiable. A successful startup and sustained operation would de-risk the technology at scale, providing the credibility needed to accelerate the Pike County project and attract further investment. Conversely, any delay or technical issue would severely undermine the exponential growth narrative.

Viewed through the lens of technological singularity, the role of energy infrastructure becomes paramount. The AI S-curve is not just about smarter algorithms; it is about the physical capacity to run them. Oklo's path from a single reactor to a 1.2 GW campus is a microcosm of the broader challenge: building the fundamental rails for a new paradigm. The company's ability to execute on these near-term catalysts will determine whether it becomes a foundational layer in that future, or merely a promising concept left behind as the demand curve accelerates beyond its reach.