You could say that supercomputers are the titans of the tech world. They’re the most advanced systems in the world, rapidly expanding the limits of science, engineering, and virtually every other field of human knowledge. In 2025, we’re entering a new era of exascale and quantum-powered computers that will take the tech to new heights. In this article, we’ll explore the world’s most powerful supercomputers, how they work, what they’re used for, and where the technology is headed. 

What Defines a Supercomputer?

Supercomputers are built to do one thing—crunch massive amounts of data at lightning speed. Their power is measured in FLOPS (floating-point operations per second). Today’s systems can handle quadrillions (petaFLOPS) or even quintillions (exaFLOPS) of these calculations every second. These machines are critical for simulating climate change, atomic behavior, advancing AI, and securing national defenses around the world.

The Rise of Exascale Computing

Exascale computing means reaching at least one exaFLOP—one quintillion floating-point operations every second. The exascale barrier was first broken by a supercomputer in 2022. Breaking into this new frontier marked a historic milestone and opened the door to a brand new level of simulation and data analysis capabilities. In three years, exascale computing has already reshaped how scientists tackle everything from cancer research to renewable energy.

El Capitan: The Current Champion

Right now, El Capitan at Lawrence Livermore National Laboratory in California holds the crown as the world’s fastest supercomputer. It holds steady at 1.742 exaFLOPS, but can reach 2.79 exaFLOPS at its peak. El Capitan uses AMD’s 4th-Gen EPYC CPUs and Instinct MI300A APUs. While it does lots of work on climate science, astrophysics, and pharmaceutical discoveries, its main job is national security tasks such as nuclear stockpile management.

Frontier: The First to Break the Exascale Barrier

Next up, the Frontier supercomputer is located at Oak Ridge National Laboratory in Tennessee. Frontier made history in 2022 as the first computer to break the exascale barrier, clocking in at 1.4 exaFLOPS. Like El Capitan, it is based on AMD hardware, though earlier generations—EPYC CPUs and Instinct MI250X GPUs. It has led to significant improvements in a wide variety of research, including cancer treatments, fusion energy, materials science, and astrophysics. Though it was the first to enter exascale territory, the Frontier computer continues to set the benchmark for scientific supercomputing.

Aurora: Intel’s Exascale Contender

Aurora, based at Argonne National Laboratory in Illinois, became the second exascale machine in the U.S. in 2023. It has a baseline of 1.0 exaFLOPS, but room to double that speed. Aurora was developed on Intel Xeon Max CPUs and Intel Data Center Max GPUs, as opposed to AMD hardware. Its focus is on modeling climate change, improving energy storage, and advancing research into nuclear fusion.

Cloud Supercomputers: Microsoft Eagle

Microsoft threw their hat in the ring with Eagle, a distributed supercomputer running on the Azure cloud. However, it hasn’t reached exascale speeds, with 561 petaFLOPS (about 0.56 exaFLOPS). Despite its lower speed, Eagle has a distinct advantage compared to some of the others. It is accessible through the cloud, which means businesses, researchers, and startups can take advantage of supercomputer-level performance without needing to set foot in a government lab.

International Leaders: Fugaku and Alps

Outside the U.S., Japan’s Fugaku, operational since 2020, was the first ARM-based supercomputer on the global leaderboard. It reached 442 petaFLOPS on the HPL benchmark, but stretched to a massive 2 exaFLOPS in mixed-precision AI tasks. Additionally, Switzerland’s Alps system has contributed to global supercomputing. It is powered by Nvidia Grace CPUs and Hopper GPUs, with a focus on researching energy efficiency and other scientific output. Both systems highlight the increasing global competition in the race for lightning-speed computation.

The Quantum Computing Revolution

2025 is shaping up to be another milestone year in lightning-fast computing. IBM is preparing to launch a quantum-centric supercomputer with over 4,000 qubits (the quantum version of a typical bit of data). This computer holds promise to reshape fields as varied as cryptography and materials science. NVIDIA’s own QuantumX AI Supercomputer, which launched earlier this year, combines AI and quantum-inspired designs. Its focus is massive simulations and ultra-fast AI training. With these advances, quantum machines are gearing up to tackle problems previously impossible to solve with classical computers.

Power, Scale, and Sustainability

Though their capabilities are remarkable, the size and power needs of today’s top supercomputers are unfortunately huge. El Capitan, for example, takes up 87 computer racks, weighs in at 1.3 million pounds, and draws 30 megawatts of power—the output of a small power plant—all on its own. As supercomputer energy consumption climbs, there is a growing emphasis on sustainability. New machines like JEDI and Alps seek to top the Green500 rankings by delivering peak performance with greater energy efficiency.

The Future of Supercomputing

This next era of supercomputing will be shaped by the combination and intermingling of both exascale and quantum technologies. Supercomputers will expand beyond the realm of traditional scientific research and national defense fields. Soon, you’ll find supercomputers behind the scenes in industries such as healthcare, finance, transportation, and advanced AI. Additionally, new challenges around cybersecurity and data privacy will emerge as quantum systems mature in their ability to decode and decrypt data. Innovation to meet these system design and encryption challenges will likely become additional tasks for supercomputers. One thing you can be sure of—tomorrow’s supercomputers will be exponentially faster, smarter, more sustainable (through energy-efficiency), and more widely accessible than ever before.

 

 

 

Thank you for reading! Have some feedback for us?
Contact the 24/7 Wall St. editorial team.