Introducing the first sub-1 nanometer node chip — the smallest, most powerful chip technology in the worldNewsMike Murphy25 Jun 2026Semiconductors
Building PyTorch-native support for the IBM Spyre AcceleratorTechnical noteRaghu Ganti, Mudhakar Srivatsa, Elpida Tzortzatos, Viji Srinivasan, Vaidyanathan Srinivasan, and Tabari Alexander06 Mar 2026AI Hardware
Expanding AI model training and inference for the open-source communityNewsPeter Hess21 Oct 2025AIAI HardwareGenerative AIHybrid CloudOpen Source
How low-precision computing boosts efficiency — without hurting accuracyExplainerPeter Hess17 Jan 2025AIAI Hardware
Serving AI models faster with speculative decoding NewsKim Martineau24 Jun 2024AIFoundation ModelsGranite
IBM’s CodeFlare significantly cuts the time to automate transfer learning tasks for foundation models ResearchCarlos Costa, Raghu Ganti, Nick Fuller, Bishwaranjan Bhattacharjee, and Mudhakar Srivatsa16 Dec 20214 minute readData ManagementFoundation ModelsHybrid Cloud PlatformMachine Learning