Groq raises $350 million at $3.5 billion valuation as it builds an Nvidia-powered AI cloud

Groq raised $350 million at a $3.5 billion valuation in a Series A round led by Disruptive, with planned participation from Nvidia, as the company pivots to operating an Nvidia-powered AI cloud. The company has raised $1 billion in new capital since June and plans to expand from 54 megawatts of capacity to more than 200 megawatts in 2027. Groq operates 13 data centers and serves more than six million developers.

Groq raises $350 million at $3.5 billion valuation as it builds an Nvidia-powered AI cloud - Groq raised $350 million at a $3.5 billion valuation, led by Disruptive, with planned participation from Nvidia. 1 https://groq.com/newsroom/groq-closes-usd350-million-series-a-building-the-world-s-leading-ai-inference-cloud - The company says it has raised $1 billion in new capital since June and plans to expand from 54 megawatts of capacity to more than 200 megawatts in 2027. 1 https://groq.com/newsroom/groq-closes-usd350-million-series-a-building-the-world-s-leading-ai-inference-cloud - Groq’s pivot follows Nvidia’s December 2025 licensing agreement for its inference technology and the departure of founder Jonathan Ross and other senior employees. 2 https://techcrunch.com/2026/06/22/ai-chipmaker-groq-confirms-650m-raise-re-staffs-after-nvidias-20b-not-acqui-hire-deal/ 3 https://www.forbes.com/sites/phoebeliu/2026/03/18/groq-cofounder-ross-explains-whirlwind-ai-chip-deal-with-nvidia/ - Groq says it operates 13 data centers and serves more than six million developers, Fortune 500 enterprises and thousands of AI-native companies. 1 https://groq.com/newsroom/groq-closes-usd350-million-series-a-building-the-world-s-leading-ai-inference-cloud Groq has raised $350 million at a $3.5 billion valuation as the former AI-chip challenger shifts toward operating a cloud infrastructure business built to offer Nvidia accelerated computing. Disruptive led the Series A round, while Nvidia is expected to participate subject to customary closing conditions, according to Groq. 1 https://groq.com/newsroom/groq-closes-usd350-million-series-a-building-the-world-s-leading-ai-inference-cloud The financing gives Groq $1 billion in new capital across two rounds announced since June. The company said it will use the money to expand its global inference footprint and support customers seeking medium- and large-sized Nvidia computing clusters for both training and inference. 1 https://groq.com/newsroom/groq-closes-usd350-million-series-a-building-the-world-s-leading-ai-inference-cloud A new company after the Nvidia deal Groq’s strategy changed after Nvidia entered a non-exclusive licensing agreement for Groq’s inference technology in December 2025. Nvidia also hired founder and chief executive Jonathan Ross, president Sunny Madra and other employees in a transaction widely reported at about $20 billion. Groq remained independent, while Nvidia gained access to the company’s language-processing-unit technology and talent. 2 https://techcrunch.com/2026/06/22/ai-chipmaker-groq-confirms-650m-raise-re-staffs-after-nvidias-20b-not-acqui-hire-deal/ 3 https://www.forbes.com/sites/phoebeliu/2026/03/18/groq-cofounder-ross-explains-whirlwind-ai-chip-deal-with-nvidia/ Nvidia later introduced the Groq 3 LPX platform, which combines Groq’s low-latency inference accelerator with Nvidia’s Vera Rubin platform. Nvidia describes LPX as an accelerator for inference workloads, while Groq’s own cloud business continues to operate separately. 4 https://www.nvidia.com/gtc/session-catalog/sessions/gtc26-s81595/ Before the agreement, Groq developed LPUs for real-time AI inference and offered them through GroqCloud and on-premises hardware clusters. After the senior departures, the company began repositioning itself as a cloud and data-center operator using Nvidia accelerated computing alongside its experience operating inference systems. 2 https://techcrunch.com/2026/06/22/ai-chipmaker-groq-confirms-650m-raise-re-staffs-after-nvidias-20b-not-acqui-hire-deal/ Capacity is the main use of capital Groq says it currently operates 13 data centers across North America, Europe, the Middle East and Asia-Pacific. Its announced capacity target is to grow from 54 megawatts to more than 200 megawatts in 2027. In June, the company said it was processing trillions of tokens each week and serving more than five million developers; its latest announcement raised that developer figure to more than six million. 1 https://groq.com/newsroom/groq-closes-usd350-million-series-a-building-the-world-s-leading-ai-inference-cloud 5 https://groq.com/newsroom/groq-raises-usd650m-to-scale-its-ai-inference-cloud-business The company is an Nvidia Cloud Partner, a designation that Groq says certifies it to design, deploy and operate Nvidia accelerated-computing infrastructure according to Nvidia’s reference architecture and operating standards. That arrangement gives Groq access to Nvidia’s hardware ecosystem while putting it in the same broad category as other Nvidia-dependent neoclouds. 1 https://groq.com/newsroom/groq-closes-usd350-million-series-a-building-the-world-s-leading-ai-inference-cloud Groq has announced deployments with infrastructure partner Equinix, including a 4.5-megawatt Sydney facility opened to support low-latency inference in Australia. The company has said Canva is among the Australian customers using its inference services. 6 https://groq.com/newsroom/groq-expands-to-asia-pacific-with-sydney-data-center-to-power-the-next-generation-of-ai-inference In Saudi Arabia, Groq and Aramco Digital have worked on an AI inference center. Saudi Aramco’s own reporting says the companies agreed to establish a large inference center and that Groq LPUs had arrived at an Aramco Digital data center. 7 The company’s 2025 announcement of a $1.5 billion Saudi commitment was described by Groq as an expansion of infrastructure and services, not as a disclosed equity investment in Groq. 8 https://groq.com/blog/saudi-arabia-announces-1-5-billion-expansion-to-fuel-ai-powered-economy-with-groq Revenue remains difficult to measure Groq does not publish financial statements or current revenue figures. Forbes reported that the company generated about $3 million in revenue and lost $88 million in 2023. Sources later told the publication that annual revenue was closer to $100 million around the time of the Nvidia agreement, well below earlier projections of $2 billion for 2025 and a later projection of $500 million. Groq’s chief executive at the time declined to comment on 2025 revenue. 3 https://www.forbes.com/sites/phoebeliu/2026/03/18/groq-cofounder-ross-explains-whirlwind-ai-chip-deal-with-nvidia/ The company’s customer metrics show broader usage than those earlier revenue figures, but they do not establish how much of that activity converts into paid, recurring cloud revenue. Groq says six million developers, Fortune 500 enterprises and thousands of AI-native companies use or are served by its platform, while its latest financing announcement does not disclose bookings, revenue, gross margins or cash-flow targets. 1 https://groq.com/newsroom/groq-closes-usd350-million-series-a-building-the-world-s-leading-ai-inference-cloud The financing also leaves Groq exposed to the economics of the neocloud model: large upfront spending on data-center capacity and hardware, followed by the challenge of keeping those systems utilized as equipment becomes outdated. TechCrunch cited CoreWeave’s capital spending, debt burden and hardware depreciation as investor concerns that apply broadly to the sector. 9 https://techcrunch.com/2026/08/17/groq-raises-350m-to-fuel-its-pivot-from-ai-chips-to-neocloud/ Groq’s immediate operating plan is clear: build more Nvidia-based capacity, sell access to inference and training clusters, and use its experience operating specialized inference systems as a differentiator. Whether that produces a durable standalone business will depend on customer utilization and revenue growth that the private company has not yet quantified. Companies mentioned Further sources 1 Groq’s August 17, 2026 announcement details the $350 million Series A, $3.5 bil… ↗ https://groq.com/newsroom/groq-closes-usd350-million-series-a-building-the-world-s-leading-ai-inference-cloud 2 TechCrunch reported Groq’s pivot after Nvidia licensed its technology, hired se… ↗ https://techcrunch.com/2026/06/22/ai-chipmaker-groq-confirms-650m-raise-re-staffs-after-nvidias-20b-not-acqui-hire-deal/ 3 Forbes reported Groq’s historical revenue, losses, estimated revenue around the… ↗ https://www.forbes.com/sites/phoebeliu/2026/03/18/groq-cofounder-ross-explains-whirlwind-ai-chip-deal-with-nvidia/ 4 Nvidia’s official GTC material describes the Groq 3 LPX rack as an inference ac… ↗ https://www.nvidia.com/gtc/session-catalog/sessions/gtc26-s81595/ 5 Groq’s June 22, 2026 announcement describes the 13-data-center footprint, token… ↗ https://groq.com/newsroom/groq-raises-usd650m-to-scale-its-ai-inference-cloud-business 6 Groq announced its Sydney deployment with Equinix, including the 4.5 MW facilit… ↗ https://groq.com/newsroom/groq-expands-to-asia-pacific-with-sydney-data-center-to-power-the-next-generation-of-ai-inference +3 more The stories that matter, in one email. Free — unsubscribe anytime.