Welcome to the era of “BYO Power.” Crusoe has been building it for years.

The White House’s recent initiative urging tech companies to “build, bring, or buy” their own power is a smart, and necessary, step to avoid straining local grids and impacting ratepayers. The “BYO power” pledge between the White House and major tech companies is ushering in a new era of energy-first AI infrastructure.

At Crusoe, this has always been how we build.

Since our founding, we have operated on a core thesis: the two greatest bottlenecks to AI progress are energy and compute — and the only way to solve them is to tackle both at once. We started by harnessing wasted, stranded energy, like flare gas, to power co-located compute. Today, we’re scaling that vision into massive, gigawatt-scale AI factories powered by a differentiated portfolio of energy solutions ranging from natural gas to second-life battery storage. We are, at our core, an energy-first company. We find pockets of abundant, underutilized power and turn them into the most valuable resource on earth: intelligence.

The principle that guides us is straightforward: move the compute to the energy, not the other way around. Traditional data center development follows the grid – waiting interconnection approvals, straining existing infrastructure, and too often driving up electricity costs for the communities it enters. We take the opposite approach. We locate AI workloads near sources of abundant, often underutilized energy – tapping into places traditional data centers can't or won't reach. This is what it means to build energy-first.

As the industry’s first vertically integrated AI factory company, we manage the entire value chain of AI infrastructure – from power generation and land, to the design and construction of data centers, to the data center components we manufacture domestically at Crusoe Industries, to our Crusoe Cloud platform. Traditional AI infrastructure is fragmented across a power company, a real estate developer, a data center operator, and a cloud platform provider. This is slow, expensive, and creates misaligned incentives. Vertical integration allows us to compress development timelines, and build gigawatt-scale campuses, and provide an AI cloud platform faster and more cost-effectively than anyone else.  

In other words, instead of relying on the utilities to get us power, we are helping to build the power ourselves.

This “BYO power” approach is already delivering unprecedented results. In 2025, our power pipeline quadrupled to 45 gigawatts – enough energy to power New York City eight to ten times over. We energized the first data center buildings on our flagship 1.2 gigawatt Stargate campus in Abilene, Texas just one year after breaking ground. Our partnership with Redwood Materials has produced the world's largest second-life battery deployment, pairing 63MWhs of repurposed EV batteries with a solar microgrid to power our modular data centers in Nevada, using the grid only for backup. By co-locating our compute with clean energy and microgrid capabilities right at the point of generation, we’re proving that data centers can operate without waiting for – or straining – the grid.

But the energy-first approach isn't only about building faster. It's about building smarter, and in ways that are good for the grid and good for the communities we enter.

AI represents the world's first truly dynamic industrial load. Unlike steel mills or refineries that must run continuously to stay viable, AI workloads are inherently flexible. We envision a world where we can locate them near abundant energy sources, pause them during periods of peak grid stress, or provide excess capacity back to the grid when it's needed most. By shifting from the traditional "perfect uptime" model to a more flexible, grid-aware approach, we can help transform data centers from passive consumers into active participants in grid stability — with the potential to operate as Virtual Power Plants.

The demand for AI compute is growing faster than the grid can keep pace. Meeting this demand requires a proactive energy-first approach – not joining the growing interconnection queue and hoping for the best. We’re excited to see the broader industry embrace the necessity of self-sufficient power generation, and we look forward to delivering the AI infrastructure that will power America's leadership for decades to come.