New major project operation from Nvidia. Popular startup Groq announced the signing of a non-exclusive licensing agreement with the chip giant for its inference technology, indicating that the alliance reflects a shared approach to expanding access to high-performance, low-cost inference.
As part of the agreement, Jonathan Ross, founder of Groq, Sunny Madra, president of the start-up, and other team members will join Nvidia to pilot and evolve the licensed technology.
According to CNBC and other US media outlets, the deal includes Nvidia’s purchase of Groq’s assets, valued at $20 billion, in what would be the largest corporate deal for Nvidia, the world’s largest company by market capitalization.
Groq has announced that it will continue to operate as an independent company, with Simon Edwards taking on the role of delegated advisor. Additionally, it clarified that its cloud business, GroqCloud, will continue to operate without interruption.
Last September, AI inference pioneer Groq announced a new funding round of $750 million, with a post-money valuation of $6.900 million. The round was led by Disruptive with significant investment from Blackrock, Neuberger Berman, DTCP and a major US West Coast investment fund manager. The review also includes continued support from Samsung, Cisco, D1, Altimeter, 1789 Capital and Infinitum.
The company then said it is providing more than two million developers and Fortune 500 companies with fast, affordable computing, in addition to expanding its global presence, leveraging existing data centers in North America, Europe and the Middle East.
With Groq on its side, Nvidia is poised to become more dominant in making chips for artificial intelligence (AI).
Nvidia’s GPU has become the industry standard. However, Groq is working on a different type of chip called an LPU (language processing unit), which the company says can run LLM at 10 times the speed and consuming a tenth of the power. Indeed, the aforementioned Ross is an expert in this technology.
Industry sources explain that large acquisitions in the United States have recently been made through the purchase of assets or acquisitions of less than 50% to avoid interference from the regulator, which takes longer than expected, such as for example the exploitation of Meta and ScaleAI, where Facebook headquarters bought 49% but is not able to do it with the whole team. Groq has its main businesses: hardware (chips) and the cloud. In this case, Nvidia does not want to launch into cloud services because that would mean directly competing with its main customers like AWS, Azure or Google Cloud, but it is interested in TPU technology and equipment to develop developments.
In this way, buy all the assets, take the technological equipment you are interested in and the cloud service, Groq Cloud within the company, and do not buy it.
So he works with the assets and talent he’s interested in and avoids buying the company, thereby avoiding jurisdictional and regulatory issues that would delay the transaction.
About The Author
