Groq has raised $350 million as it moves from building its own artificial intelligence chips to running cloud and data center services. The funding, announced on August 17, 2026, gives the company new capital to expand its capacity and provide access to powerful Nvidia systems for AI training and inference.
Disruptive led the round, with planned participation from Nvidia. Groq’s valuation now stands at $3.5 billion, down from $6.9 billion last September. The change reflects a company reshaping its business around AI infrastructure instead of competing with Nvidia as a chipmaker.
From custom chips to Nvidia systems
Groq was founded around its own AI chips, called LPUs, which were designed to compete with Nvidia on inference. That strategy has now changed. After losing its star team, Groq shifted from being a pure AI chipmaker into a cloud and data center provider that operates Nvidia systems.
The new funding will support the use of medium and larger sized clusters of Nvidia accelerated computing for both training and inference. In simple terms, Groq plans to give customers access to the computing power needed to build AI systems and run them after they are trained.
That puts Groq inside Nvidia’s AI infrastructure ecosystem rather than outside it. Nvidia already supplies the GPUs powering clouds from CoreWeave, Lambda, and Nebius, and Nvidia is investing billions into some of those companies. Groq’s planned participation from Nvidia adds another connection between the chip company and the cloud providers building services around its hardware.
Groq’s chairman is Alex Davis, who is also the CEO of Disruptive. Davis described the market opportunity in direct terms: “Inference will without a doubt become the largest and most critical layer of AI infrastructure.”
A bigger data center footprint
Groq raised a $650 million round in June 2026 to support its pivot. The company now plans to scale from 54 megawatts of capacity to more than 200 megawatts in 2027, a major expansion of the power available to its AI infrastructure business.
Groq operates 13 data centers across North America, Europe, the Middle East, and Asia Pacific. Its services reach more than 6 million developers, enterprises, and AI-native companies, giving the company an existing customer base as it builds out its new role as a cloud provider.
The capacity goal matters because AI services need computing power not only when models are trained, but also each time those models respond to a request. Groq’s shift places that ongoing work at the center of its business. Its new clusters are intended to support both sides of the workload: training AI systems and handling inference after those systems are ready for use.
What the new valuation says
Groq’s $3.5 billion valuation is half of the $6.9 billion valuation recorded last September. The company’s financials are still private, so the funding and valuation figures offer the clearest public view of how investors value its new direction.
The lower valuation does not change the scale of Groq’s infrastructure plans. The company has raised $350 million in its latest round, raised $650 million in June 2026, and aims to grow from 54 megawatts to more than 200 megawatts in 2027. Those figures point to a business that is trading its original chip strategy for a larger infrastructure buildout.
Groq is not abandoning AI hardware as a customer need. Instead, it is using Nvidia systems to provide the computing capacity that developers, enterprises, and AI-native companies require. That move also places the company alongside cloud providers such as CoreWeave, Lambda, and Nebius, all of which use Nvidia GPUs.
The result is a very different Groq from the company that set out to challenge Nvidia with LPUs. Its future now depends on operating data centers, serving more than 6 million customers across its user base, and expanding the infrastructure available for AI training and inference. With Nvidia involved in the funding plans, Groq’s pivot has become part of the ecosystem it once sought to challenge.
Based on
