Groq secures $350 million to develop its neocloud offering

Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Groq announces $350 million raised, led by Disruptive, with Nvidia's participation anticipated, to continue its transition from in-house chips to neocloud backed by Nvidia GPUs. The company aims for over 200 MW of capacity by 2027 and already operates 13 data centers. However, the neocloud model remains under scrutiny regarding its profitability, while Groq's finances remain private.
Nvidia powers CoreWeave, Lambda, Nebius, and invests billions
Groq's refocus places it directly within Nvidia's infrastructure ecosystem. The GPU provider is already equipping the clouds of CoreWeave, Lambda, and Nebius, and is injecting billions of dollars into some of these platforms to expand their capabilities. Groq's accounts remain confidential, while the long-term viability of the neocloud model is still undecided, despite sustained demand for inference as companies increase their workloads. CoreWeave, which reported strong growth in the second quarter and signed contracts with Meta and Anthropic, still raises investor concerns about the scale of capital expenditures, debt reliance, exposure to rapidly depreciating hardware, and converting growth into free cash flow. This type of relationship with Nvidia is not isolated within the sector.
Groq aims for over 200 MW by 2027 and operates 13 data centers
Groq claims to operate 13 data centers across North America, Europe, the Middle East, and Asia-Pacific, serving over 6 million developers, businesses, and AI-native companies. The roadmap plans to ramp up from 54 megawatts to over 200 megawatts of capacity by 2027. According to the company, the new funds will support clients seeking medium to large-sized Nvidia compute clusters for training and inference. Groq positions itself in inference, with its leaders presenting the goal of building the "most powerful" inference cloud, while estimating that this layer will become the most significant part of AI infrastructure.
$350M raised, $3.5B valuation, and Nvidia participation expected
The $350 million raise is led by Disruptive, with Nvidia's participation announced as anticipated. It sets a valuation at $3.5 billion, down from $6.9 billion last September. A spokesperson for Groq does not characterize this round as a "down round" and sees it as establishing a post-licensing contract valuation with Nvidia. In June, Groq had already mobilized $650 million to initiate this shift. The company, which initially aimed for in-house chips called LPUs to compete with Nvidia in inference, lost its star team and transformed into a cloud and data center provider operating Nvidia systems. Nvidia has also hired Jonathan Ross and other top profiles as part of a $20 billion licensing rights deal paid to investors. Groq is now continuing this transformation towards neocloud.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.