Trend Trend 2: heterogeneous supercomputing platforms become the substrate for advanced models

Summary

Gartner's second trend concerns AI supercomputing platforms that provide the vast computing capacity that training and operating frontier AI models requires. They bring together high-performance computing, specialised chips and architectures that can scale to handle data-heavy workloads. Gartner's illustration places GPUs, CPUs, custom AI chips, quantum processors, neuromorphic chips and even biological components under a single hybrid orchestration layer and programming model. Example uses include optimisation and simulation, energy and climate modelling, new materials and drug discovery. The trend is driven by organisations building ever larger models that exceed the limits of traditional infrastructure.