Emerging firm Groq has secured $350 million, furthering its strategic transition from an AI chip manufacturer to a neocloud entity offering robust GPUs and AI infrastructure solutions.
This fresh investment, spearheaded by Disruptive and slated to include Nvidia, positions Groq’s valuation at $3.5 billion. This marks a decrease from the $6.9 billion valuation Groq attained last September. This prior valuation occurred only a few months before Nvidia recruited the startup’s founder and CEO, Jonathan Ross, along with other key personnel, as part of a $20 billion licensing agreement distributed to investors.
According to a company representative who spoke with TechCrunch, the firm views this adjusted valuation not as a “down round” but as the establishment of a new benchmark for “the iteration of Groq following the Nvidia licensing agreement.”
Initially, Groq concentrated on developing its proprietary chips, known as LPUs (language processing units), aiming to rival Nvidia in the inference domain — the computational power required for real-time AI workload execution. Subsequent to the departure of its highly skilled team, Groq transitioned from being solely an AI chip producer to a provider of cloud and data center services, utilizing Nvidia’s infrastructure.
Last June, Groq secured $650 million in funding to initiate its strategic realignment. The company aims to expand its capacity from 54 megawatts to over 200 megawatts by 2027.
Presently, Groq manages 13 data centers spanning North America, Europe, the Middle East, and Asia Pacific, catering to over 6 million developers, businesses, and AI-focused organizations. Groq indicated that the newly acquired capital will facilitate “users requiring access to medium and larger clusters of Nvidia’s accelerated computing for both training and inference tasks.”
“Our objective is to establish Groq as the premier AI inference cloud globally,” stated Alex Davis, who serves as Groq’s chairman and Disruptive’s CEO. He added, “Inference will undeniably evolve into the most substantial and essential component of AI infrastructure.”
Although demand for inference is robust as companies expand their AI operations, the long-term profitability of neoclouds and their capacity to yield returns on significant investments remains uncertain. CoreWeave demonstrated impressive revenue growth in its second quarter and recently secured substantial agreements with entities like Meta and Anthropic. Nevertheless, financiers expressed worries regarding the company’s substantial capital outlays, significant debt dependency, susceptibility to rapidly devaluing hardware, and its proficiency in converting expansion into free cash flow.
Currently, Groq’s financial details are not publicly disclosed, yet its strategic redirection positions the firm squarely within Nvidia’s AI infrastructure framework. This arrangement is not uncommon among contemporary neocloud providers. Nvidia furnishes the GPUs that power cloud services from CoreWeave, Lambda, and Nebius, concurrently injecting billions into several of these companies as they strive to augment their computational capabilities.
TechCrunch has sought additional details from Groq.