9 most powerful supercomputers in the world today

Supercomputers work on the same principles as conventional computers, but have superior performance and are more massive in size, reminiscent of the classical computers of old.

Unlike desktops or laptops, they process huge amounts of data and perform calculations at incredible speeds. According to IBM, supercomputers are the fastest machines in the world, requiring complex infrastructure to operate, including advanced cooling systems.

Architecturally, supercomputers are equipped with more powerful components than regular PCs. While a laptop has only one central processing unit (CPU) and one graphics processing unit (GPU), a supercomputer can have thousands of CPUs and GPUs, each with performance far exceeding that of conventional commercial components.

Their performance is measured in floating point operations per second (FLOPS). While PCs or laptops can only reach a few hundred gigaFLOPS (trillions of FLOPS), the world's most powerful supercomputers today have surpassed 1 exaFLOP, which is one billion billion (10¹⁸) FLOPS. Machines that reach this level are called exascale supercomputers.

Thanks to their ability to process vast amounts of data at lightning speed, supercomputers have become indispensable tools in many fields of science. They aid in drug development, discover new materials, predict the weather, and can even “learn” to play chess, like Deep Blue, the legendary IBM supercomputer that defeated Russian chess master Garry Kasparov in 1997.

1. El Capitan Supercomputer

- Location: Lawrence Livermore National Laboratory, California, USA

- Performance: 1,742 petaFLOPS (equivalent to 1,742 exaFLOPS)

- Components: 4th generation 24-core AMD EPYC CPU integrated in AMD Instinct MI300A APU

- Start of operation: November 2024

El Capitan is currently the world's most powerful supercomputer and the third exascale machine ever deployed, officially coming online in late 2024.

Housed at Lawrence Livermore National Laboratory, the supercomputer plays a vital role in protecting and securing America's nuclear arsenal without underground testing.

In addition, El Capitan also supports top-secret missions related to national security, including applying artificial intelligence (AI) and machine learning to advanced defense systems.

Not only limited to the military field, El Capitan is also used to solve complex scientific problems, especially in materials science and physics.

This supercomputer was built by the US Department of Energy as part of the CORAL-2 program, to replace Sierra, a supercomputer deployed in 2018. With its peak performance, El Capitan marks a major step forward in computing power, opening a new era of exascale supercomputers.

2. Frontier Supercomputer

- Location: Oak Ridge National Laboratory, Tennessee, USA

- Performance: 1.353 petaFLOPS (equivalent to 1.4 exaFLOPS)

- Components: 64-core AMD EPYC CPU and AMD Instinct MI250X GPU

- Start of operation: August 2022

Ranked second in the list of the world's most powerful supercomputers, Frontier, built by HPE Cray, made history by becoming the first exascale supercomputer in 2022.

According to the world's leading technology magazine IEEE Spectrum, Frontier was originally designed to support advanced research in many important fields such as cancer, drug discovery, nuclear fusion, new materials, super-efficient engine design and simulation of stellar explosions in the universe.

In the future, Frontier will continue to play an important role in scientific research, from developing next-generation transportation technology to breakthroughs in medicine. In addition, this supercomputer will be used for large-scale simulations of the universe.

Evan Schneider, Associate Professor of Astrophysics at the University of Pittsburgh (USA), shared with MIT Tech Review that he plans to use Frontier to study the evolution of the Milky Way over time.

With its outstanding performance, Frontier is not only a milestone in high-performance computing, but also opens a new era for science and technology.

3. Aurora Supercomputer

- Location: Argonne National Laboratory, Illinois, USA

- Performance: 1,012 petaFLOPS (equivalent to 1.0 exaFLOPS)

- Components: Intel Xeon Max Series CPU and Intel Data Center Max Series GPU

- Start of operation: June 2023

One of the newest and most promising supercomputers on the list, Aurora could become the most powerful machine on the planet. Located at the Argonne Leadership Computing Facility (ALCF), Aurora is the third exascale supercomputer ever built and, according to ALCF representatives, the system can reach 2 exaFLOPS, which is double the performance of Frontier.

Developed jointly by Intel and HPE, Aurora not only supports scientific tools and analytics, but also performs complex simulations and runs advanced AI algorithms. With its massive computing power, Aurora can create accurate models in areas such as climate forecasting, materials science, energy storage, and nuclear fusion.

Nuclear fusion is one of Aurora's key research focuses, according to high-performance computing news site HPCWire.

Scientists hope that with the power of this supercomputer, the mystery of large-scale fusion energy will soon be solved, opening a new era for clean and sustainable energy sources.

4. Eagle Supercomputer

- Location: Microsoft Azure

- Performance: 561 petaFLOPS (equivalent to 0.56 exaFLOPS)

- Components: 48-core Intel Xeon Platinum 8480C CPU and Nvidia H100 GPU

- Start of operation: August 2023

Unlike traditional supercomputers housed in research labs, Microsoft's Eagle exists entirely in the cloud. Deployed through the Microsoft Azure platform, Eagle is not only one of the fastest supercomputers in the world, but it can also be accessed by anyone willing to pay for its use.

As a distributed network of powerful systems, Eagle has enough power to rank 4th on the list of the world's fastest supercomputers. This marks a major step forward in cloud computing, where supercomputers are no longer confined to physical laboratories, but can be exploited remotely, serving research, AI, data science and simulation work on a global scale.

5. HPC6 Supercomputer

- Location: Eni Green Data Center, Ferrera Erbognone, Italy

- Performance: 477.9 petaFLOPS (up to 606 petaFLOPS)

- Components: 3rd generation 64-core AMD EPYC CPU in AMD Instinct MI250X GPU

- Start of operation: November 2024

In November 2024, Italian energy giant Eni launched HPC6, Europe's most powerful supercomputer and the world's first industrial supercomputer. It is the first non-US supercomputer to make it into the top five fastest supercomputers in the world.

Thanks to groundbreaking improvements in the infrastructure at the Eni Green Data Center, the performance of HPC6 has increased almost tenfold, from 70 petaFLOPS to a maximum of 606 petaFLOPS. The system is designed to support research on decarbonization and energy efficiency optimization, contributing to sustainable development.

HPC6 will undertake many important tasks in advanced science and technology areas such as geological and fluid dynamics research for CO₂ storage, as well as high-performance battery development, biofuel supply chain optimization, materials development for biochemistry applications, and simulation of plasma behavior in magnetically confined fusion, supporting research on fusion energy.

The HPC6 is also equipped with a modern liquid cooling system, using the "direct" method to dissipate 96% of heat, improving performance and saving maximum energy.

With its enormous computing power and groundbreaking applications, HPC6 is not only the pride of Europe but also a big step forward in the field of supercomputing for industry and clean energy.

6. Fugaku Supercomputer

- Location: Riken Center for Computational Science, Kobe, Japan

- Performance: 442 petaFLOPS (equivalent to 0.44 exaFLOPS)

- Components: CPU A64FX

- Start of operation: June 2020

Having held the title of the world's most powerful supercomputer from June 2020 to June 2022, Fugaku remains one of the oldest systems on the list of top supercomputers today.

Fugaku, named after Mount Fuji, Japan's iconic volcano, is located about 100 kilometers from Tokyo. The system replaced Summit, which for the first time took the No. 1 spot in the TOP500 ranking of the world's most powerful supercomputers.

Over the years, Fugaku has been used for important scientific research. In particular, during the COVID-19 pandemic, scientists have taken advantage of Fugaku's computing power to simulate the spread of the virus.

According to Nikkei Asia, this supercomputer has helped prove that non-woven fabric masks are more effective in blocking droplets released into the air, contributing to the direction of health policy.

Today, Fugaku continues to assert its pioneering role by being used to train Japan's large language AI models (LLMs), similar to how ChatGPT was developed. According to the Japan News, this is part of Japan's AI development strategy, to build AI models optimized for Japanese language and culture.

Although no longer in the top position, Fugaku is still one of the most important supercomputers in the world, making great contributions to science, medicine and AI.

7. Alps Supercomputer

- Location: Swiss National Supercomputing Center, Lugano, Switzerland

- Performance: 270 petaFLOPS (equivalent to 0.27 exaFLOPS)

- Components: Nvidia GH200 Grace Hopper processor

- Start of operation: September 2024

Inaugurated in September 2024, the Alps supercomputer is one of the world's most powerful computing machines, designed to process the massive amounts of data needed for modern science.

It is also the replacement for Piz Daint, the former supercomputer of the Swiss National Supercomputing Centre (CSCS). In particular, Alps is open to global researchers, helping to accelerate cutting-edge scientific research.

Alps' infrastructure is designed with flexible resource allocation capabilities, allowing a portion of its computing power to be dedicated to different specialized purposes.

Although the main Alpine hub is at CSCS, the system does not operate from a single location. It is distributed across several major research facilities, including the Swiss Federal Institute of Technology in Lausanne, the Paul Scherrer Institute (PSI) in Villingen, and the European Centre for Medium-Range Weather Forecasts (ECMWF) in Bologna, Italy.

The Alps supercomputer is not only a powerful computing machine but also an open scientific research center, supporting fields from meteorology to AI. With its flexible architecture and distributed network, Alps is helping to shape the future of science and technology globally.

8. LUMI Supercomputer

- Location: CSC Data Center, Kajaani, Finland

- Performance: 380 petaFLOPS (equivalent to 0.38 exaFLOPS)

- Components: 3rd generation 64-core AMD EPYC CPU and AMD Instinct MI250X GPU

- Start of operation: June 2021

LUMI, based in Finland, was once Europe's most powerful supercomputer and is now the eighth fastest in the world. According to European Union (EU) officials, it runs on 100% renewable hydroelectric power and its waste heat is used to heat nearby buildings. It began testing in June 2021 and is expected to be fully operational in February 2023.

LUMI is not only a powerful supercomputer, but is also designed to serve researchers across Europe, promoting large-scale scientific collaboration projects.

The system is specifically optimized for AI-based workloads, accelerating advanced data analytics and simulations, officials previously said.

One of LUMI's breakthroughs is its role in combining the power of supercomputing with quantum computing. Specifically, LUMI acts as a "partner" to two Finland-based quantum computing systems, QAL 9000 and Helmi.

This collaboration helps scientists tap into the potential of quantum computing, although the technology is still limited compared to what it could achieve in the future.

The combination of quantum computing and supercomputers like LUMI creates a more powerful system that supports cutting-edge research in areas such as quantum chemistry, materials simulation, and AI optimization.

LUMI is not only an advanced AI supercomputer, but also a bridge between classical and quantum computing, providing researchers with today's most powerful tools to explore new horizons in science and technology.

9. Leonardo Supercomputer

- Location: CINECA Data Center, Bologna, Italy

- Performance: 239 petaFLOPS (equivalent to 0.23 exaFLOPS)

- Components: 32-core Intel Xeon Platinum 8358 CPU and Nvidia A100 GPU

- Start of operation: November 2022

Leonardo is one of Europe's most advanced supercomputers, part of the European Union's EuroHPC program. The system is built from three powerful modules, which together make up the second fastest supercomputer in Europe, providing enormous computing power for scientific and technological research.

The supercomputer is operated by CINECA, a leading Italian technology consortium comprising leading universities, public research centers and government agencies.

Leonardo is not only a powerful supercomputer, but also a symbol of scientific and technological cooperation in Europe, contributing to important breakthroughs in many fields.