New US supercomputer turbo charges US hypersonic weapons research

China has been using supercomputers and Artificial Intelligence (AI) combined with the world’s fastest hypersonic wind tunnels now for years to accelerate their hypersonic weapons development programs. And now the US Air Force Research Laboratory (AFRL) has unveiled a new $20 million supercomputer designed to tackle some of the military’s most computationally demanding challenges, from hypersonic weapons research to AI development and next-generation aircraft design.

Known as Flyer, the machine was recently commissioned at Wright-Patterson Air Force Base in Ohio. According to AFRL officials, the system can solve in a single day problems that would take an average laptop roughly 500 years to complete.

Flyer has been part of a multi-year modernisation effort under the US Department of Defense High Performance Computing Modernization Program (HPCMP). The system was originally announced alongside its classified counterpart, Raven, as part of AFRL’s TI-23 supercomputing initiative.

The launch was first reported by local outlets like Springfield News-Sun and WHIO-TV. Flyer is designed to support large-scale simulations that would be impractical, prohibitively expensive, or outright impossible to conduct in the real world.

The 22 Biggest Societal & Technological Changes by 2040, a keynote by futurist Matthew Griffin

The system contains approximately 186,000 processing cores, 800 terabytes of RAM, and 18 petabytes of storage. AFRL officials noted that matching its memory capacity alone would require roughly two million laptops.

While those specifications are impressive, the real value lies in what the machine enables. Military researchers increasingly rely on digital engineering, Computational Fluid Dynamics (CFD), AI training, and virtual testing environments to evaluate new technologies before building physical prototypes. Instead of developing multiple aircraft designs or conducting thousands of expensive flight tests, engineers can simulate those scenarios digitally and rapidly iterate on designs.

AFRL has previously highlighted hypersonic vehicle development as one area where supercomputers provide a major advantage. Vehicles travelling at several times the speed of sound are difficult and expensive to test repeatedly in real-world conditions, making advanced simulation essential.

The growing importance of supercomputers is directly related to how militaries develop new technology. Modern defence programs generate enormous amounts of data from sensors, weapons systems, aircraft, satellites, and autonomous platforms. Processing that information requires computing infrastructure capable of performing trillions and even quadrillions of calculations every second.

AFRL officials have repeatedly described modelling and simulation as one of the primary uses of their supercomputing resources, allowing researchers to conduct higher-fidelity analyses while dramatically reducing development timelines. Previous AFRL systems have reduced simulation projects from months to weeks while enabling more detailed studies of complex engineering problems.

The laboratory has also emphasised the role of high-performance computing in the development of AI, where greater computing resources enable researchers to train more sophisticated models and run larger-scale experiments.

Flyer joins a growing fleet of supercomputers operated by AFRL and the Department of Defense. The laboratory previously announced that Flyer and Raven would deliver approximately 14 petaflops of computational capability. A petaflop represents one quadrillion calculations per second, placing the system among the most powerful military research computers in the United States.

As defence development becomes increasingly digital, systems like Flyer are emerging as critical infrastructure. The machine may never leave Wright-Patterson Air Force Base, but the technologies it helps design, from future aircraft and autonomous systems to advanced weapons and AI models, could influence military capabilities for decades to come.