The collapse of [Crusoe Energy Systems’](https://crusoeenergy.com) $1.25 billion partnership with [Boom Supersonic](https://boomsupersonic.com) sends shockwaves through the AI infrastructure market, where power consumption has become the defining constraint for data center expansion. The deal would have represented one of the largest cross-industry collaborations between aviation and AI infrastructure, leveraging Boom’s supersonic jet engine technology for stationary power generation.
[Boom Supersonic](https://boomsupersonic.com) CEO Blake Scholl’s confirmation that the stationary power plants were “no longer in Crusoe’s near-term plans” suggests the partnership hit technical or economic roadblocks that couldn’t be resolved. The aviation company had been developing adapted versions of its supersonic aircraft engines for ground-based power generation, promising higher efficiency and lower emissions compared to traditional data center power solutions.
The abandoned partnership highlights the desperate search for sustainable power solutions in the AI sector. [Crusoe](https://crusoeenergy.com) has built its business model around using stranded energy sources – particularly flared natural gas from oil fields – to power cryptocurrency mining and AI compute operations. The company’s pivot away from Boom’s technology suggests either technical challenges with the turbine adaptation or economics that didn’t pencil out at scale.
Industry analysts had viewed the Crusoe-Boom partnership as a potential game-changer for AI infrastructure. Traditional data centers already consume roughly 1% of global electricity, and AI workloads are pushing that figure dramatically higher. [Microsoft](https://microsoft.com) recently signed a deal to restart Three Mile Island’s nuclear reactor specifically to power AI operations, while [Google](https://google.com) and [Amazon](https://amazon.com) have invested billions in renewable energy projects to meet their AI power demands.
The timing of Crusoe’s decision is particularly significant as the AI infrastructure market faces a power crunch. [Nvidia’s](https://nvidia.com) latest H200 and upcoming B200 AI chips consume substantially more power than previous generations, forcing data center operators to completely rethink their power and cooling strategies. Some facilities are hitting local grid capacity limits, creating bottlenecks for AI deployment.
For [Boom Supersonic](https://boomsupersonic.com), the setback represents a major blow to its diversification strategy beyond passenger aviation. The company has faced its own challenges bringing its Overture supersonic airliner to market, with first flights pushed back multiple times. The stationary power business was seen as a way to commercialize Boom’s engine technology while the aviation program matured.
The collapse also raises questions about Crusoe’s funding and strategic direction. The company raised $350 million in Series C funding in 2022, but the AI boom has dramatically increased both opportunities and capital requirements in the space. Without the Boom partnership, Crusoe may need to pursue more conventional power solutions or seek additional funding to scale its operations.
Market observers are now watching whether other AI infrastructure companies will pursue similar aviation-to-power partnerships. Several startups are exploring everything from small modular nuclear reactors to advanced geothermal systems to solve the AI power equation. The failure of the Crusoe-Boom deal may cool investor enthusiasm for unconventional power solutions, pushing the industry back toward proven but less innovative approaches.
Crusoe’s decision to abandon its $1.25 billion Boom Supersonic partnership reflects the harsh realities of scaling AI infrastructure in a power-constrained world. While innovative solutions like adapted jet engines offer theoretical advantages, the practical challenges of deployment, economics, and reliability often favor more traditional approaches. For the AI industry, this setback underscores that solving the power equation will require not just innovation, but proven, scalable solutions that can deliver reliable energy at the massive scale AI demands. The race for sustainable AI power continues, but it may take a more conventional path than many hoped.
