Zettascale

Description

Builds XPU chips for AI inference and training, with an FPGA prototype, planned serving API, and hosted XPU clusters designed to act as one chip.

Related

• Together AIBuilds a full-stack AI platform offering inference APIs, batch inference, GPU clusters, custom infrastructure, training, fine-tuning, and…
• LambdaBuilds modular AI factories and provides managed GPU clusters, supercomputers, and instances for labs and teams running AI training,…
• fal.aiProvides generative media model APIs, galleries, serverless GPU inference, on-demand clusters, and dedicated compute for developers…
• krtklAccelerated intelligence — FPGA + multiprocessor compute platforms for embedded vision, signal processing, and edge AI.
• SatlytProvides above-cloud service provider offerings, including hosted computing and data access services, to help customers manage information…
• Nodus ComputeBuilds an AI compute platform that places and runs training, inference, and agent workloads, with checkpoint recovery and an…
• AnyscaleBuilds Ray-based software for AI teams to scale distributed training, data curation, batch inference, and serving workloads for foundation…
• Standard MachinesBuilds environments and evaluations for advanced chip design, including reinforcement learning environments for long-horizon design tasks…
• FluidstackBuilds AI compute and power infrastructure, including solar generation, large data centers, and XPU operations for organizations needing…
• MatXBuilds high-throughput chips for large language models, supporting frontier labs with hardware for training, reinforcement learning,…
• Lattice Semiconductor ($LSCC)Designs programmable chips, low-power FPGAs, software tools, and secure management platforms for industrial, automotive, communications,…
• StarcloudDevelops space-based data centers for AI training clusters and publishes related white papers, while engaging suppliers and customers for…