Image source: Groq

Groq is a startup developing chips that generate AI models that run faster than traditional processors. Neuberger Berman, Type One Ventures, Cisco, KDDI, and Samsung also participated in the round.

Groq reportedly initially hoped to raise $300 million at a slightly lower valuation ($2.5 billion), and the funding brings Groq's total funding to more than $1 billion and a valuation of $2.8 billion, which is a major win for Groq. That's more than double Groq's pre-valuation of about $1 billion in April 2021, when the company raised $300 million in a funding round led by Tiger Global Management and D1 Capital Partners.

Groq also announced today that Yann LeCun, chief AI scientist at Meta, will serve as a technical advisor to Groq, and Stuart Pann, former head of Intel's foundry business and former CIO of Hewlett-Packard, will join the startup as COO. LeCun's appointment comes as a bit of a surprise given Meta's investment in its own AI chips, but it certainly provides a strong ally for Groq in this highly competitive space.

Rising from stealth in 2016, Groq is creating a so-called LPU (Language Processing Unit) inference engine. The company claims that its LPUs can run existing generative AI models with an architecture similar to OpenAI's ChatGPT and GPT-4o, 10 times faster and only one-tenth the energy consumption.

Groq CEO Jonathan Ross is best known for helping invent the Tensor Processing Unit (TPU), Google's custom AI-accelerated chip for training and running models. Ross and Douglas Wightman, entrepreneurs and former engineers at Google's parent company, Alphabet X Lunar Labs, co-founded Groq nearly a decade ago.

Groq offers an LPU-powered developer platform called GroqCloud, which offers "open source" models such as Meta's Llama 3.1 series, Google's Gemma, OpenAI's Whisper, and Mistral's Mixtral, as well as APIs that allow customers to use their chips in cloud instances. (Groq also offers a platform for AI-powered chatbots, GroqChat, which launched late last year.) As of July, GroqCloud had more than 356,000 developers, and Groq said that part of the proceeds from the round will be used to expand production capacity and add new models and features.

Stuart Pann, Groq's chief operating officer, told TechCrunch, "Many of these developers are employees of large enterprises. We estimate that more than 75% of the Fortune 100 companies are customers. "

As the generative AI boom continues to heat up, Groq faces increasing competition from rival AI chip up-and-comers and powerful hegemons in AI hardware.

It is estimated that Nvidia controls between 70 and 95 percent of the market for AI chips used to train and deploy generative AI models, and the company is taking aggressive steps to maintain its dominance.

Nvidia has pledged to release a new AI chip architecture every year, instead of every other year as historically. Nvidia is reportedly building a new business unit focused on designing custom chips, including AI hardware, for cloud computing companies and others.

In addition to Nvidia, Groq competes with Amazon, Google, and Microsoft, all of which offer or will soon offer custom chips for AI workloads in the cloud. Amazon offers Trainium, Inferentia, and Graviton processors through AWS; Google Cloud customers can use the above TPUs and use Google's Axion chips for a certain period of time; Microsoft recently launched a preview of Azure instances for its Cobalt 100 CPUs, and Maia 100 AI Accelerator instances will also be available in the coming months.

Some analysts believe that in the AI chip market, which could reach $400 billion in annual sales over the next five years, Groq could also look to Arm, Intel, AMD and a growing number of startups as competitors. Arm and AMD's AI chip businesses are particularly thriving, thanks to soaring capital expenditures by cloud computing vendors to meet capacity demands for generative AI.

D-Matrix raised $110 million late last year to commercialize what it calls a first-of-its-kind inference computing platform. In June, Etched emerged from stealth and received $120 million in funding to customize a processor to accelerate Transformer, the mainstream generative AI model architecture today. SoftBank's Masayoshi Son is reportedly looking to raise $100 billion for a chip company to compete with Nvidia. In addition, OpenAI is said to be in talks with investment companies to launch an AI chip manufacturing plan.

In order to develop its own market, Groq has invested heavily in corporate and government promotions.

In March, Groq acquired Definitive Intelligence, a Palo Alto-based company that offers a range of enterprise-oriented AI solutions, to form a new business unit called Groq Systems. Groq Systems' role is to serve organizations that want to add Groq chips to existing data centers or build new data centers using Groq processors, including United States government agencies and sovereign nations.

Most recently, Groq partnered with government IT contractor Carahsoft to sell its solutions to public sector customers through Carahsoft's reseller partners, and the startup also signed a letter of intent with European company Earth Wind & Power's Norway data center to install tens of thousands of LPUs.

Groq is also working with Saudi Arabia consulting firm Aramco Digital to install LPUs in future data centers in the Middle East.

While building customer relationships, Groq, based in Mountain View, California, is also moving toward next-generation chips. In August last year, the company announced that it would sign a contract with Samsung's foundry business to produce 4nm LPUs, which are expected to improve performance and efficiency compared to Groq's first-generation 13nm chips.

Groq says it plans to deploy more than 108,000 LPUs by the end of the first quarter of 2025. 

This article is from Xinzhi self-media and does not represent the views and positions of Business Xinzhi.If there is any suspicion of infringement, please contact the administrator of the Business News Platform.Contact: system@shangyexinzhi.com