Cloud Computing Solutions

Perplexity uses the NVIDIA accelerated computing platform on AWS to power AI training and inference. 

The company reduces model training time by up to 40% with Amazon SageMaker HyperPod, accelerated by NVIDIA GPUs. During spike periods, it delivers near-real-time inference for 10,000 concurrent users and 100,000 queries per hour using Amazon EC2 P5 Instances, accelerated by NVIDIA Hopper™ GPUs and NVIDIA GPU-optimized software.

BMW transformed its electric vehicle production system by leveraging NVIDIA AI and Azure Machine Learning, enabling real-time, AI-powered automated inspections that dramatically improve quality control and operational efficiency in electric drive system manufacturing.

Writer, a full-stack generative AI platform for enterprises, leverages NVIDIA H100 and L4 Tensor Core GPUs on GKE with the NVIDIA NeMo™ framework and TensorRT-LLM to train and deploy over 17 large language models that scale up to 70 billion parameters.

Beamr uses NVIDIA L40S GPUs on OCI for accelerated video processing and achieves 30% more efficient video encoding.

A-Alpha Bio uses NVIDIA BioNeMo™ and NVIDIA Hopper GPUs on AWS to accelerate antibody drug discovery.

The company achieved 12X faster inference and a 10X increase in predictions, leading to higher-quality drug candidates.

Encina, a plastic recycling innovator, is tackling climate change by leveraging Microsoft Azure, CPFD Software, and NVIDIA accelerated computing to run simulations 506X faster than CPU-based methods, significantly reducing costs and accelerating the design of next-generation recycling facilities. This computational leap enables Encina to revolutionize sustainable plastic recycling and contribute to global efforts to reduce plastic waste and carbon emissions.

LiveX AI leverages the power of NVIDIA NIM microservices on Google Kubernetes Engine with NVIDIA GPUs to achieve a 6.1X increase in average token speed. This enhancement lets LiveX AI deliver personalized experiences to customers in real time, including seamless customer support, instant product recommendations, and reduced returns.

Modal Labs uses a wide range of bare-metal machines accelerated by NVIDIA GPUs on OCI to quickly scale resources when launching their customers' demanding generative AI workloads.