The new scientific stack

The convergence of HPC and AI is an architectural change. AI does not reduce the importance of physics-based modelling; in many cases, it increases the value of simulation by making its outputs reusable across a larger workflow. A 2023 Nature review documented this shift across hypothesis generation, experimental design and simulation interpretation.¹

Building the adaptive scientific workflow

The scientific stack that supports this work has to behave less like a queue of isolated jobs and more like an adaptive system. Researchers need to move between simulation, data processing, model development, analysis and validation without rebuilding their environment at every step. This requires processor performance, memory bandwidth, low-latency networking, orchestration, governance and AI tooling to work together. Cloud-based HPC gives teams flexibility to align infrastructure with workloads and reduce friction between stages.

AWS and the infrastructure for scientific discovery

At AWS, this direction is reflected in an advanced computing portfolio built for scientific work. For HPC users, the foundation is Amazon Elastic Compute Cloud (Amazon EC2), which includes purpose-built HPC instances for workloads that depend on throughput, memory bandwidth and efficient communication across many nodes, most recently with Amazon EC2 Hpc8a instances powered by 5th Generation AMD EPYC processors.²

The significance is not any single capability, but the range of infrastructure that scientific discovery now requires. CPU-based HPC instances support tightly coupled simulation. GPU and accelerated compute infrastructure support AI training, inference and visualisation. Orchestration services connect these stages so simulation can generate data, AI can explore a wider design space, and researchers can return to higher-fidelity methods with better questions.

This pattern is visible in engineering workflows on AWS. In one automotive example, AWS Batch launched computational fluid dynamics simulations that created ground truth data, Amazon SageMaker-trained drag prediction models, and GPU-based Amazon EC2 infrastructure served models through an interactive design workflow.³

AWS Parallel Computing Service extends this workflow approach for organisations that use Slurm. In chemical formulation and discovery, Aionics migrated to AWS Parallel Computing Service in 2025 to support computational chemistry workloads, including density functional theory simulations and GPU-accelerated molecular dynamics with machine-learned interatomic potentials.⁴ 

This example shows the direction of the stack: not just more compute, but a managed HPC environment that helps researchers focus on scientific search rather than infrastructure maintenance. 

This extends beyond engineering. The Allen Institute has used HPC and generative AI on AWS for neuroscience research, including omics pipeline runs that moved from weeks on premises to as little as one day on AWS.5 The importance is not only an isolated performance gain. It is the ability to change the pace at which teams move from question to computation, and from computation to insight.

Quantum and the next stage of scientific computing

Quantum computing belongs in this same stack, not as a replacement for HPC or AI, but as a future accelerator for selected problems in chemistry, materials science, optimisation and quantum systems. Amazon Braket gives researchers access to quantum computers and simulators to explore quantum and hybrid algorithms alongside classical workflows. Recent work with Kvantify Qrunch on Amazon Braket demonstrates how quantum chemistry calculations can be developed using hybrid quantum algorithms on today’s quantum devices.⁶

AWS is also investing in the longer-term hardware path. In February 2025, Amazon announced Ocelot, a quantum computing chip developed by the AWS Center for Quantum Computing. Ocelot uses a hardware-efficient approach to quantum error correction based on cat qubits, aiming to reduce the overhead of fault-tolerant quantum computing.⁷

Much of today’s AI conversation is focused on productivity, automation and consumer applications. The deeper opportunity is to expand the rate at which researchers can explore, test and validate ideas. When HPC provides the physical and numerical foundation, AI helps navigate vast search spaces, and quantum computing begins to extend what can be modelled, the effect is not simply faster computing. It is a different cadence of discovery, one that will shape progress in energy, medicine, materials and fundamental science.

Charunethran Panchalam Govindarajan heads Product Marketing for Advanced Computing at AWS, spanning high-performance computing, AI for science and quantum computing. He holds a Master’s degree in Electrical Engineering from Stanford University.

References

1. Hanchen Wang et al., "Scientific discovery in the age of artificial intelligence," Nature, 2023. https://www.nature.com/articles/s41586-023-06221-2
2. "A technical deep dive into Amazon EC2 Hpc8a performance for engineering and scientific workloads," AWS HPC Blog, March 2026. https://aws.amazon.com/blogs/hpc/a-technical-deep-dive-into-amazon-ec2-hpc8a-performance-for-engineering-and-scientific-workloads/
3. "Using machine learning to drive faster automotive design cycles," AWS HPC Blog, May 2024. https://aws.amazon.com/blogs/hpc/using-machine-learning-to-drive-faster-automotive-design-cycles/
4. "How Aionics accelerates chemical formulation and discovery with AWS Parallel Computing Service," AWS HPC Blog, November 2025. https://aws.amazon.com/blogs/hpc/how-aionics-accelerates-chemical-formulation-and-discovery-with-aws-parallel-computing-service/
5. "Powering discovery: Allen Institute advances in-silico science using AWS," AWS Case Study. https://aws.amazon.com/solutions/case-studies/allen-institute-case-study/
6. "Building advanced quantum chemistry calculations with Kvantify Qrunch on Amazon Braket," AWS Quantum Technologies Blog, February 2026. https://aws.amazon.com/blogs/quantum-computing/building-advanced-quantum-chemistry-calculations-with-kvantify-qrunch-on-amazon-braket/
7. "Amazon announces Ocelot quantum chip," Amazon Science, February 2025. https://www.amazon.science/blog/amazon-announces-ocelot-quantum-chip