Beyond the GPU: What Will Power the Next Wave of AI? | VSORA, SiPearl, ZML, SpiNNcloud & Scaleway

As inference expands and power constraints tighten, AI infrastructure faces demands that raw compute alone cannot solve. Which workloads call for specialized chips, and where do established architectures still win? Bringing together VSORA, SiPearl, ZML, and SpiNNCloud, this panel examines competing approaches to AI compute — and the software needed to make them usable. From performance per watt to developer adoption, the discussion explores what it takes for an alternative architecture to earn its place in production.