Tensordyne

Tensordyne’s inference system uses logarithmic math to deliver high capacity with reduced energy, space, and cost.

Purpose-built logarithmic math, 
hard-wired in 3 nm silicon, Tensordyne is ready for the largest models and real-time 4K video.

Generative AI models have grown from 340 million parameters in 2018 to nearly 1.8 trillion in 2022, enabled by GPU-based systems scaling to over 10,000 GPUs.

Juniper Networks invested in Recogni and collaborates on Generative AI inference technology, participating in its $102 million Series C funding round.

Recogni appoints Kevin Johnson, former CEO of Starbucks and Juniper Networks, as an independent board member and investor.

Recogni Inc., the Generative AI Inference systems company, and DataVolt, the sustainable AI-cloud innovator, are teaming up to build the most energy-efficient AI cloud on the planet.

Tensordyne hires talent in computer science, deep learning, silicon engineering, systems engineering, networking, software, and business, with offices in Sunnyvale and Munich.

Tensordyne develops AI math, chips, hardware, and software using a logarithmic number system to reduce energy and cost.

Tensordyne replaces its former name, Recogni, to reflect its focus on tensors and AI inference efficiency.

Tensordyne offers technology information, demos, beta programs, and job applications.

Tensordyne presents next-generation AI inference systems, technology, products, and research at Tensordyne.ai.

This Cookie Policy was last updated on August 20, 2024 and applies to citizens and legal permanent residents of the European Economic Area and Switzerland.

Website Operator & Main Headquarters: Tensordyne, Inc. 1195 Bordeaux Dr Sunnyvale CA 94089 United States ‍