For many companies building large infrastructure today, the hardest question is no longer the price of electricity, but whether dependable power will be available at all.
Data centers, electrification, and advanced manufacturing facilities now require energy at a scale the grid wasn’t designed to supply. The ambitions shaping the next decade of the economy increasingly depend on power systems that can keep pace.
The Electric Power Research Institute’s analysis projects that data centers could account for 9–17% of U.S. electricity generation by 2030, more than double their current share. A single hyperscale campus can draw as much electricity as a mid-sized city, and developers are often planning the next expansion before the first buildings are finished.
The IEA’s Energy and AI report projects that data centers will account for nearly half of U.S. electricity demand growth through 2030, driven primarily by AI infrastructure. Much of that load is arriving in concentrated clusters, placing new pressure on regional grids that were designed for slower, more distributed growth.
For infrastructure developers, energy strategy is beginning to look less like procurement and more like capacity planning.
For many of these projects, the pressure ultimately converges on one requirement: baseload power—electricity that remains available regardless of time of day or weather. Data centers, semiconductor fabrication plants, and advanced manufacturing facilities operate continuously, and interruptions quickly translate into operational and financial risk, making dependable baseload power a foundational requirement.
NERC’s Long-Term Reliability Assessment warns that a growing number of North American grid regions already face elevated resource adequacy risk as load growth begins to outpace planned capacity additions. For companies planning large infrastructure projects, power availability increasingly shapes where and whether projects can move forward.
Speed compounds the constraint. Capital deployment timelines and competitive pressure mean developers rarely have the luxury of waiting years for new generation to come online. Power that arrives too late to keep a project viable is, for practical purposes, no solution at all.
And for a growing share of these organizations, the power also must be clean. Corporate decarbonization commitments, investor scrutiny, and evolving carbon accounting standards mean that the emissions profile of a facility’s power supply now affects financing, reporting, and long-term operating risk.
Recent Goldman Sachs analysis of hyperscaler energy procurement puts it plainly: timing constraints are pushing developers toward behind-the-meter natural gas as a near-term bridge, even as their long-term preference remains grid-connected power. There’s a clear gap between what developers need and what clean firm generation can deliver on current timelines, and it’s driving procurement decisions that most organizations would prefer to avoid.
Firm, clean baseload power that can be deployed quickly remains difficult to secure.
Solar and wind have reshaped electricity markets, and installed capacity has grown faster than most forecasts predicted. Their output, however, follows weather and time of day. Storage can help smooth those fluctuations, and batteries are becoming an important part of grid flexibility. But for facilities that run continuously, storage alone rarely closes the gap.
Other sources of clean baseload power face different constraints. Geothermal can provide dependable baseload power, but is geographically limited. Nuclear offers reliable generation at scale, but development timelines often stretch a decade or more before first power.
Companies needing firm capacity on timelines that match project development frequently land on natural gas. Gas turbines offer dispatchable, scalable baseload power, but for organizations with decarbonization commitments, that choice solves the near-term reliability problem while creating a longer-term emissions problem.
Carbon capture is often presented as a way to reconcile natural gas generation with climate goals, but in practice, conventional capture systems introduce their own constraints.
When fuel combusts in air, the exhaust stream is heavily diluted, dominated by nitrogen with only a small fraction of CO₂. Separating that CO₂ requires chemical solvents and significant additional processing. Typical systems capture roughly 80–90% of emissions, still produce NOx and other criteria pollutants, and require large capture facilities alongside the plant. It’s an improvement over unabated gas, but it doesn’t deliver zero operating emissions, and it doesn’t address deployment speed.
There’s also a supply issue. Demand for gas turbines has surged as new data center and industrial projects come online, pushing manufacturing lead times years into the future.
At Arbor, we’re approaching the problem at the level of the thermodynamic cycle itself. The result is a compact system designed to deliver clean baseload power without the constraints that limit conventional gas turbines.
Instead of combusting fuel in air and bolting on carbon capture afterward, the system separates nitrogen from the air before combustion occurs. Fuel then reacts with pure oxygen, producing an exhaust composed only of CO₂ and water. Cooling condenses the water, leaving a high-purity CO₂ stream ready for permanent geologic storage while the remaining CO₂ recirculates as the working fluid in a supercritical CO₂ power cycle. Because supercritical CO₂ is far denser than steam, the turbomachinery required to move it can be dramatically smaller.
That change in scale opens the door to a different manufacturing approach. HALCYON components are produced using additive manufacturing techniques common in aerospace and rocket propulsion, like 3D printing. That allows for faster design iteration, tighter tolerances, and more compact system layouts than traditional casting processes allow. Aerospace-grade simulation tools are used throughout the design process, allowing engineers to test the system across operating conditions before hardware is manufactured.
Each HALCYON moduarborle delivers 25 MW of power with greater than 98% carbon capture and no NOx, SOx, or particulate emissions. Arbor’s domestic sourcing shares no components with conventional gas turbine supply chains and requires no critical minerals.
For decades, baseload power was largely a grid planning concept; today it is becoming a core input to economic development. Companies building the next generation of data centers, manufacturing facilities, and digital infrastructure increasingly find that power availability determines where projects can move forward. In many regions, the question is no longer how much electricity costs but whether dependable baseload power can be secured at all.
Natural gas generation still fills that role for many projects, but it introduces emissions liabilities that grow harder to manage as climate commitments tighten and carbon accounting evolves.
Meeting that need requires power systems designed around different assumptions from the start. Systems where carbon capture is inherent to the cycle itself. Systems compact enough to be pre-fabricated and deployed faster. Systems designed to deliver clean baseload power at the pace modern industry now demands.
For companies planning the next generation of infrastructure, access to that kind of power will increasingly determine what gets built. The systems that succeed will be the ones able to keep pace with the ambitions driving this new wave of infrastructure.
—
If baseload power availability is shaping how your next project moves forward, get in touch with our team.