Tenstorrent, Samsung Partner for Next-Gen AI Chipsets

The Canadian startup Tenstorrent, which, among other things, develops artificial intelligence (AI) processors, announced a new partnership with Samsung’s semiconductor manufacturing division.

The startup announced its partnership with Samsung on October 2, stating that it will use the partnership to deliver the “next generation of AI chipsets” to market. Data center processors and intellectual property (IP) are manufactured by Tenstorrent.

The agreement stipulates the use of one of Samsung’s sophisticated manufacturing processes, the SF4X process and 4nm architecture, for the production of its next-generation chips.

According to the announcement, the product to be developed in collaboration with Samsung will be a chipset that will be packaged with other chipsets.

The objective, according to Jim Keller, CEO of Tenstorrent, is to develop “high-performance compute” and deliver these solutions to “customers around the world.” Tenstorrent completed a $100 million funding round led by Samsung and Hyundai in August.

Tenstorrent stated at the time that the funds would be used to accelerate product development and design, develop AI chipsets, and create a roadmap for machine learning software.

This development occurs amid a fierce competition to develop the most powerful AI processors for the creation of advanced AI systems. Nvidia is currently the market leader in terms of chip development and the quality of the chips they have produced.

During the AI fervor, the company’s valuation briefly surpassed $1 trillion earlier this year. While Nvidia remains at the head of the field for chip development, they recently came under scrutiny by authorities in France in an antitrust investigation.

After French antitrust authorities conducted a broad investigation into the cloud computing industry, the local police invaded Nvidia’s offices.

According to a Reuters report, the European Commission stated on October 2, following the events in France, that EU antitrust regulators have not launched a formal investigation into AI processors.

In October 2022, the United States will prohibit the export of high-level Nvidia technology to China as a result of regulators’ efforts to exert control over the industry.

In recent months, additional measures have been implemented. As an American corporation, Nvidia is subject to U.S. regulations and restrictions.

The U.S. also recently concluded a multibillion-dollar deal with Vietnam, a traditional trading partner with China and erstwhile adversary of the U.S., regarding AI chips and AI technology.