Boom Superpower and the Race to Power AI Data Centres

The Boom Superpower, a 42-megawatt turbine for data centre power generation from aviation company Boom Supersonic. (Photo: Boom Supersonic)

Boom Supersonic, best known for its plans to bring back commercial supersonic flight, has just stepped directly into the AI infrastructure story.

The company has announced “Superpower”, a 42-megawatt natural gas turbine designed to provide dedicated power for AI data centres. Alongside the launch, Boom revealed a backlog of more than $1.25 billion for the turbine, its first customer – energy-first AI infrastructure company Crusoe – and an additional $300 million funding round led by Darsana Capital Partners and other investors.

Superpower uses the same core technologies being developed for Boom’s Symphony supersonic jet engine. The turbine is pitched as a way to accelerate AI data-centre deployment while simultaneously generating real-world operating data for the supersonic engine programme – a neat example of dual-use turbomachinery.

What’s special about Superpower?

According to the announcement, Superpower’s key characteristics include:

• 42 MW of ISO-rated power in a shipping-container-scale package

• Ability to deliver full rated output even at ambient temperatures above 110°F (~43°C)

• Waterless operation, avoiding the need for evaporative cooling or water-intensive systems

• Operation on natural gas with diesel back-up capability

• A production ramp targeting over 4 GW of turbine capacity per year by 2030

Crusoe has reportedly ordered 29 units to power its advanced AI data centres, positioning Superpower as a distributed, high-density power source for AI clusters.

The positioning is very deliberate: traditional grid connections in some regions are struggling to keep up with the pace and scale of AI data-centre demand, and large-scale, on-site generation is being explored as a way to bridge that gap.

The bigger context: AI, data centres and power

Superpower sits at the intersection of several trends already reshaping the energy and digital-infrastructure landscape:

• Global electricity use by data centres is projected by the IEA to grow from roughly 460 TWh in 2024 to over 1,000 TWh in 2030, with AI as the main driver of that growth.

• Goldman Sachs Research forecasts that global data-centre power demand could increase by up to 165% by 2030 compared with 2023, driven largely by generative AI.

• Analyses from McKinsey and others suggest that the world will need hundreds of gigawatts of new data-centre capacity, much of it in locations chosen primarily for power availability rather than proximity to traditional tech hubs.

At the same time, a number of industry pieces have highlighted how, in the near term, natural gas is playing a major role in meeting the surge in AI power demand – especially in regions where renewable build-out and grid reinforcement cannot keep pace with the speed at which AI campuses are being commissioned. DataCenterDynamics+1

Boom’s Superpower turbine is a high-profile example of that trend: using aeroderivative gas-turbine technology, optimised for high ambient temperatures and water-constrained sites, as a way to deliver tens of megawatts of dedicated capacity per site.

Thermal and environmental angles

From an engineering perspective, several aspects of Superpower stand out:

• High-temperature performance: Turbines typically derate (lose output) at high ambient temperatures. Superpower is explicitly marketed as maintaining full capacity above 110°F: suggesting advanced materials, cooling and compressor/combustor design to push the thermal envelope.

• Waterless operation: Many large power plants and some data-centre cooling systems depend on evaporative cooling and therefore on water availability. By avoiding water altogether, Superpower targets hot and arid locations where water is a critical constraint, a theme increasingly visible in discussions of AI and data-centre sustainability.

• Integration with data-centre cooling: When gas turbines are coupled to data centres, they create opportunities (and challenges) around waste-heat utilisation, combined-cycle operation, and integration with liquid or two-phase cooling systems inside the data hall. All of this has to be carefully designed and optimised if the overall system is to be efficient, not just the prime mover.

From a climate perspective, this development also sharpens an uncomfortable question that the IEA and others are now raising: how much of the AI boom will be powered by low-carbon sources versus fossil fuels, and over what timeframe?

Superpower’s natural gas base – even with high efficiency and potential for lower-carbon fuels in the future – underlines the reality that short-term AI power solutions are often carbon-intensive, even as longer-term plans lean heavily on renewables, nuclear and other low-carbon options.

Where simulation and Mansim come into the picture

Although Boom’s work is in aerospace-derived turbomachinery, the underlying engineering challenges are very familiar in Mansim’s core domains:

• High-temperature turbomachinery and combustor flowsDesigning turbine components that maintain performance at high ambient temperatures relies on detailed CFD and conjugate heat-transfer modelling, including blade cooling, hot-streak mixing, and thermal-stress considerations.

• Gas-turbine + data-centre integrationWhen a data centre is tied directly to on-site generation, there is a coupled system of:

  • turbine and exhaust conditions,

  • heat-recovery and cooling plant,

  • rack-level liquid or air-plus-liquid cooling,

  • and potentially district-heat or process-heat integration.System-level simulation is essential to understand efficiency, resilience and failure modes.

• Hot-climate and water-constrained designWaterless turbines and high-density AI cooling in hot regions raise questions about airflow, plume behaviour, thermal recirculation and environmental impact, areas where high-fidelity CFD is often the only way to see the full picture before hardware is built.

Whether future AI campuses are powered by gas turbines like Superpower, nuclear SMRs, large-scale renewables + storage, or hybrids of all three, the underlying need will be the same: careful modelling of flows, heat and systems-level behaviour to ensure that the physical infrastructure keeps pace with the digital ambitions built on top of it. If you’re exploring how to tackle the energy and cooling challenges of AI or high-performance infrastructure, we’d be happy to discuss how advanced simulation can help, please get in touch via our contact page.