Zettascale
ActiveEnergy efficient chips for AI
About
Zetta is building the next NVIDIA, energy-efficient chips ("XPUs") for AI training and inference, offering superior speed and energy efficiency compared to traditional GPUs and TPUs. Our XPUs are reconfigurable, capable of optimizing the dataflow of each model, making them faster and more energy-efficient than the current SOTA chips on the market. This saves data centers *billions* in cooling and energy costs.
Founders · 2
Co-founder & CEO of the Zettascale Computing Corp., autodidact since age 9, with a background in Computer Science & Computer (Architecture) Engineering (BSc + MSc, dropout), mathematics, and aerospace.
Co-founder & CTO of the Zettascale Computing Corp., with a background in Mechanical Engineering, Generative Modeling, and Electrical Engineering
Launch
Reconfigurable chips for AI that are 27.6x more energy efficient.
Exa Laboratories announces reconfigurable AI chips that adapt to each model architecture to improve efficiency and performance, with simulations showing up to 27.6x gains versus H100 GPUs. The product targets data centers and AI workloads, aiming to reduce energy use and associated costs while enabling diverse neural network designs.
Formerly “Exa Labs”, “Exa Laboratories”
Related startups
Run AI pipelines 10x faster
Efficient and compact power supply for AI hardware
