Valuation Drops from $6.9 Billion to $3.5 Billion
Raises $350 Million in New Funding
Pivots to Inference and Cloud Infrastructure
Chairman Amidi Sees Nvidia Deal as "Insurance"

Groq, an AI semiconductor startup, has secured $350 million (about 490 billion won) in new funding. However, the company's valuation has dropped to $3.5 billion, only half of last year's peak. Since Nvidia recruited Groq's founder and key talents, the company has been rapidly shifting its focus from being an AI chip competitor to becoming an operator of inference-focused data centers.


Chip Developed by Groq

Chip Developed by Groq

View original image

According to Bloomberg News on the 17th (local time), Groq raised $350 million at a valuation of $3.5 billion. This investment was led by Disruptive, an investment firm headquartered in Dallas, Texas, with Nvidia also expected to participate as an investor.


Groq's corporate valuation has declined significantly in just one year. In September of last year, Groq attracted $750 million in investment at a valuation of $6.9 billion. Its value, which was $2.8 billion in August 2024, more than doubled in just over a year, but in this recent round, the valuation has been reduced to about half of last year's peak, settling at $3.5 billion.


The turning point came in December last year with a deal involving Nvidia. Nvidia signed a contract to license Groq's AI inference technology and, at the same time, recruited key personnel including Jonathan Ross, the founder and then CEO, as well as President Sunny Madra. Legally, Nvidia did not acquire Groq and allowed it to remain an independent entity, but it was considered an unusual deal to secure both key talent and rights to use the technology.


Groq was founded in 2016 by Ross, who had participated in the development of Google's AI semiconductor, the Tensor Processing Unit (TPU). It garnered attention as a potential competitor to Nvidia in the AI inference market, developing a Language Processing Unit (LPU) with a structure different from Nvidia's Graphics Processing Unit (GPU).

Groq Halves Its Valuation...Shifts Focus from Competing with Nvidia to Data Centers View original image

The rapidly growing inference market, which runs pre-trained models in real services as opposed to training new ones, has attracted significant interest from investors.


In an exclusive interview with The Asia Business Daily in February, Saeed Amidi, chairman of Plug and Play, a leading Silicon Valley venture capital firm, commented on the deal between Nvidia and Groq, describing Nvidia CEO Jensen Huang's move as "a kind of insurance policy" against future competition.


Plug and Play invested in Groq when its valuation was around $2 billion. At that time, Chairman Amidi said, "I thought Groq could become a big tech company even with just a 3–5% market share," citing the company's different type of computation compared to Nvidia as a key investment rationale.


Regarding the deal between Nvidia and Groq, he also explained, "I think Nvidia decided it would be good to have Groq's solutions in its portfolio in case they became more significant."


As a result, Nvidia managed to secure key personnel and technology without acquiring Groq, while Groq pivoted from being a standalone AI chip competitor to focusing on data center and cloud infrastructure operations.


Currently, Groq is concentrating on expanding its inference data center business to capture surging demand for AI computation. In June, the company raised an additional $650 million to facilitate the business transition, and its valuation was also adjusted at that time, though specific numbers were not publicly disclosed. With the announcement of the $3.5 billion valuation in this latest fundraising, this is the first concrete update on Groq's post-Nvidia deal valuation drop.


Groq has also been strengthening its partnership with Nvidia. The company has decided to introduce Nvidia systems into its own AI cloud service, GroqCloud, data centers. As a result, customers will be able to use not only Groq's LPUs but also Nvidia's technology within the same cloud service.



Groq plans to use a portion of the new funding to expand its data center footprint, aiming to increase its total facility capacity to over 200 megawatts (MW) by next year. Alex Davis, chairman of Groq's board, stated, "Inference will unquestionably become the largest and most important tier in AI infrastructure," emphasizing the company's commitment to supporting major AI model developers.


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing