Bringing cloud-native AI supercomputing to a data center near you
Technical note
Talia Gershon, Yoonho Park, Nicole Saulnier, Daniel Klyashtorny, Jamey Roe, and Edward Barth
Hybrid Cloud Infrastructure
We’re driving infrastructure innovations spanning physical infrastructure, operating systems, and software-defined compute, network, and storage. Our goal is to create technologies that improve performance, reliability, security, and ease of operation across hybrid and multi-cloud environments.
Our work
Reinventing data preparation on cloud to scale enterprise AI
NewsMike MurphyA new chip architecture points to faster, more energy-efficient AI
NewsMike MurphyComposable systems with OpenShift
Technical noteChristian Pinto, Michele Gazzetti, and Michael JohnstonWhy we built an AI supercomputer in the cloud
ResearchTalia Gershon, Seetharami Seelam, Jay Jubran, Eran Gampel, and Drew ThorstensenCars “see around corners” in EU’s CLASS project
ReleaseErez Hadad8 minute read- See more of our work on Hybrid Cloud Infrastructure
Publications
Towards Efficient Key-Value Cache Management for Prefix Prefilling in LLM Inference
- Yue Zhu
- Hao Yu
- et al.
- 2025
- CLOUD 2025
Routing Strategies for RoCE Networks in AI Clouds
- Alim Alim
- Ali Sydney
- et al.
- 2025
- CLOUD 2025
Mind the Memory Gap: Unveiling GPU Bottlenecks in Large-Batch LLM Inference
- Pol G. Recasens
- Ferran Agullo
- et al.
- 2025
- CLOUD 2025
Rex: Closing the language-verifier gap with safe and usable kernel extensions
- Jinghao Jia
- Ruowen Qin
- et al.
- 2025
- USENIX ATC 2025
Scalable and Efficient LLM Serving with the vLLM Production Stack
- Junchen Jiang
- Yue Zhu
- 2025
- OSSNA 2025
How Low Can LoRA Go: System-Level Throughput, Energy, and Model Quality Tradeoffs when Fine-Tuning Adapters
- Connor Espenshade
- Umesh Deshpande
- et al.
- 2025
- ISCA 2025