June 15 (Reuters) – Tensordyne said on Monday it expects more than $200 million in orders for its new inference system, as the AI chip startup positions itself as a direct challenger to Nvidia in a fast-growing but increasingly power-hungry market.
Its Tensordyne Napier chip, developed in collaboration with Broadcom and HPE-owned Juniper Networks, is being manufactured by TSMC, the world’s leading contract chipmaker.
The Sunnyvale, California-based company aims to ease AI infrastructure constraints by improving inference speed, power efficiency and rack density amid surging generative AI demand.
CEO Marc Bolitho told Reuters that the product is set for an official launch in the coming months and the startup is seeing “a great amount of interest.”
“We have over a dozen letters of intent for companies to evaluate our beta systems and we have over $200 million in forecasted demand going forward,” he said.
The company said AI infrastructure providers Cirrascale and BlueSky Compute had shown interest in the system, along with large technology companies and AI cloud service providers.
Founded in 2017 as Recogni, it rebranded as Tensordyne last year. It has raised about $176 million from investors including Celesta Capital, GreatPoint Ventures and Juniper Networks, and is preparing for a Series D funding round later this year.
(Reporting by Jaspreet Singh in Bengaluru; Editing by Tasim Zahid)
