Akamai Deploys Thousands of NVIDIA Blackwell GPUs Across 4,400+ Locations to Cut AI Latency 2.5x
Akamai is deploying thousands of NVIDIA Blackwell GPUs across its 4,400+ edge locations to create a globally distributed AI inference platform that reduces latency up to 2.5x and cuts inference costs by up to 86% vs hyperscale clouds. The infrastructure supports localized fine-tuning, post-training optimization and AI decisions at the edge.
1. NVIDIA Blackwell GPU Deployment
Akamai is integrating thousands of NVIDIA Blackwell GPUs into its global network of over 4,400 edge locations, creating one of the world’s most widely distributed AI compute fabrics designed specifically for inference workloads.
2. Performance and Cost Advantages
The new architecture routes AI inference tasks to optimized GPU clusters, delivering up to 2.5x lower latency and up to 86% cost savings on inference compared to traditional hyperscale clouds, while enabling localized fine-tuning and post-training model optimization.
3. Strategic Impact and Applications
By extending AI processing to the edge, Akamai positions itself to support real-time decision-making across industries—ranging from autonomous delivery and smart grids to surgical robotics and fraud prevention—strengthening its competitive edge against centralized cloud providers.