Groq is making big moves in the world of AI chips. The startup, known for developing chips that can run GenAI models faster than traditional hardware, is setting its sights on the enterprise and public sector. In a recent announcement, Groq revealed the formation of a new division called Groq Systems, aimed at expanding its customer and developer ecosystem. This division will specifically cater to organizations, including government agencies, looking to incorporate Groq’s powerful chips into their data centers.
In addition to this exciting news, Groq announced that it has acquired Definitive Intelligence, a Palo Alto-based company specializing in AI solutions for businesses. Definitive Intelligence CEO Sunny Madra will now lead GroqCloud, Groq’s cloud platform offering hardware documentation, code samples, and self-serve API access for their cloud-hosted accelerators.
“At Groq, we’re dedicated to creating an AI economy that is accessible and affordable for anyone with a brilliant idea,” stated Jonathan Ross, co-founder and CEO of Groq. “We are thrilled to welcome Sunny and his team from Definitive Intelligence to help us achieve this mission. The Definitive team brings a wealth of expertise in AI solutions and go-to-market strategies, as well as a strong commitment to sharing knowledge with the community.”
Madra, who had previously co-founded the cloud-based platform Autonomic, brings a deep understanding of AI solutions and the market. Madra and his co-founder Gavin Sherry were also responsible for launching Autonomic, which was later acquired by Ford in 2018.
Definitive Intelligence’s products include OpenAssistants, a set of open-source libraries for creating AI chatbots, as well as Advisor, a visualization generator that connects to both public and enterprise databases. One of their flagship tools is Pioneer, an “autonomous data science agent” designed for predictive modeling and other data analytics tasks.
Prior to its acquisition, Definitive Intelligence had raised an impressive $25.5 million in venture capital funding.
“The world is just now realizing how crucial high-speed inference is for generative AI,” said Madra in a statement. “At Groq, we are providing developers with the speed, low latency, and efficiency necessary to deliver on the promise of generative AI. I have been a fan of Groq since I first met Jonathan in 2016, and I am excited to join him and the Groq team in their quest to bring the fastest inference engine to the world.”
Groq, which came out of stealth mode in 2016, is developing an LPU (Language Processing Unit) inference engine. The company boasts that its LPU can run large language models like OpenAI’s ChatGPT and GPT-4 at 10 times the speed of traditional hardware.
Ross, who played a significant role in inventing Google’s tensor processing unit (TPU) chip for AI model training and execution, is a key figure at Groq.
The acquisition of Definitive Intelligence marks Groq’s second purchase after acquiring Maxeler Technologies, a firm specializing in high-performance computing and AI infrastructure solutions, in 2022. And it may not be the last. The competition for custom AI chips is fierce, and this latest acquisition suggests that Groq is determined to establish a foothold before its rivals can catch up.
