Arbor Energy has clinched a roughly $1 billion purchase from GridMarket for up to 5 gigawatts of its Halcyon modular turbines, a significant vote of confidence in rocket-grade turbomachinery repurposed for the power grid. The deal targets fast-growing data centers and heavy industrial loads that need reliable, low-carbon electricity on tight timelines.
The order equates to about 200 machines at 25 megawatts apiece. Halcyon’s core is built around 3D-printed and precision-machined components derived from space propulsion, packaged for rapid deployment and sited close to major power users. For buyers that can’t wait through long interconnection queues, the appeal is obvious.
Inside the GridMarket deal and Arbor’s delivery plan
Arbor says its first grid-tied Halcyon unit will precede a production ramp aimed at triple‑digit annual output once its manufacturing lines stabilize, with an eventual goal of enabling 10 gigawatts of new capacity each year. The GridMarket agreement is fuel‑flexible by design, allowing projects to run on biomass‑derived syngas or natural gas while keeping the same carbon capture pathway.
The buyer, GridMarket, aggregates and develops power solutions for data centers and industrial campuses. Expect the initial turbines to be deployed in modular blocks near large loads—four to a dozen units per site—where on‑site generation can ease transmission constraints and cut curtailment risk.
Rocket turbomachinery meets the grid for firm power
Halcyon borrows from rocket engine turbopumps—hardware built to thrive at extreme pressures and temperatures—then adapts that DNA for terrestrial duty cycles. The resulting hot‑section can be additively manufactured, shortening lead times and reducing reliance on artisanal single‑crystal casting of blades and vanes that constrains traditional gas turbine supply chains.
That manufacturing angle matters. Incumbent heavy‑duty turbine makers have been cautious about re‑expanding capacity after past market swings, and buyers routinely face multi‑year waits. By leaning on 3D printing and advanced machining, Arbor is betting it can deliver power plants in the window data center operators actually need them.
Emissions math and fuel choices for Halcyon turbines
Arbor designed Halcyon around oxy‑combustion: fuel is burned with pure oxygen rather than air, producing an exhaust that’s mostly CO₂ and water. That simplifies capture to a dense CO₂ stream, avoiding the energy penalty of scrubbing dilute flue gas with amines. Department of Energy and academic literature show oxy‑combustion configurations can achieve capture rates above 95% when properly integrated with air separation and compression systems.
On biomass‑derived syngas, the system functions as BECCS (bioenergy with carbon capture and storage). Because the carbon originated in plants, capturing and storing it makes the net electricity carbon‑negative when feedstock sourcing and logistics are well managed. On natural gas, Halcyon is not carbon‑negative, but Arbor targets less than 10 grams of CO₂ per kilowatt‑hour after capture and storage—far below the roughly 400 g/kWh typical of standard gas plants without capture, according to the International Energy Agency.
The big caveat is methane leakage. Peer‑reviewed work supported by Environmental Defense Fund has measured U.S. oil and gas supply chain leakage around 1%–3%, which erodes lifecycle gains if not tightly controlled. Arbor says it is prioritizing low‑leak suppliers and permanent geologic storage. U.S. 45Q incentives can pay up to $85 per metric ton of CO₂ captured and stored from point sources, improving project economics, with additional value possible in regional low‑carbon fuel markets.
Data Center Demand Turbocharges Adoption
Data centers have become the unexpected kingmakers in power markets, driving a surge in firm capacity needs that intermittent renewables and long‑lead transmission can’t always satisfy. The IEA estimates global data center electricity use could roughly double this decade from the mid‑hundreds of terawatt‑hours annually and may top the terawatt‑hour mark each year if AI build‑outs continue apace.
Meanwhile, interconnection queues are swollen. Research from Lawrence Berkeley National Laboratory shows new projects routinely face waits approaching five years on average to secure grid access. That mismatch—fast IT deployment versus slow grid upgrades—is pushing operators to procure on‑site or near‑site generation that can scale in 25‑ to 50‑megawatt increments with clear schedules and known emissions profiles.
Manufacturing Bottlenecks And Arbor’s Bet
Traditional turbines depend on a concentrated ecosystem of specialty alloys, single‑crystal casting houses, and veteran craftspeople. Those chokepoints don’t expand quickly. Arbor’s use of additively manufactured hot sections and rocket‑style high‑speed rotating machinery aims to sidestep the bottlenecks and compress delivery times without sacrificing performance.
If the company can consistently print and qualify critical parts at scale—maintaining tolerances, surface finishes, and creep life—it will unlock not just faster deliveries but also faster design iteration. That feedback loop has already transformed aerospace propulsion and could do the same for power generation.
What to watch next as Arbor and GridMarket execute
Three milestones will determine whether this order becomes a broader market shift. First, successful grid‑tied demonstrations with stable capture performance and bankable operating data. Second, reliable access to oxygen and CO₂ transport and storage, including Class VI well permitting where applicable. Third, proof that unit economics pencil for both biomass and natural gas configurations once fuel, capture, and sequestration are included.
For now, Arbor’s billion‑dollar pact signals how quickly energy procurement is changing. If rocket‑inspired turbines can deliver firm, low‑carbon power on developer timelines, they’ll become a serious contender alongside combined‑cycle gas, small modular reactors, and long‑duration storage in the race to keep the digital economy energized.
