Groq has raised $350 million in a Series A round as the AI infrastructure company moves to expand its inference cloud across global markets. Disruptive led the financing, with planned participation from NVIDIA, giving Groq a valuation of $3.5 billion. The new capital follows $650 million raised in June 2026, bringing the company’s recent funding haul to $1 billion. For Groq, the financing arrives as AI infrastructure spending increasingly shifts from building models toward running them at scale.

The company’s strategy centers on making inference a major infrastructure layer rather than treating it as an afterthought to model training. Groq already operates 13 data centers spanning North America, Europe, the Middle East, and Asia Pacific, giving it a footprint across several of the world’s most important AI markets. Its customer base includes more than six million developers, Fortune 500 enterprises, and thousands of AI-native companies. That combination gives Groq a substantial installed audience as demand for production AI workloads continues to grow.

Groq Targets 200+ Megawatts For 2027

The funding will help Groq support customers looking for access to medium and larger clusters of NVIDIA accelerated computing for both training and inference. The company currently operates at 54 megawatts and expects to push beyond 200 megawatts in 2027. That planned expansion represents a significant change in infrastructure scale, particularly as AI workloads require larger pools of compute capacity closer to users and applications. The move also positions power availability as a central component of Groq’s next phase rather than simply a supporting requirement.

Scaling from 54 megawatts to more than 200 megawatts means Groq will need to coordinate compute, facilities, networking, cooling, and deployment capacity across multiple regions. The company’s existing international footprint gives it a foundation for that expansion, while the planned access to NVIDIA accelerated computing adds another layer to its infrastructure strategy. Groq’s approach therefore goes beyond adding servers and focuses on building an inference platform capable of handling workloads at commercial scale. That distinction matters as enterprises increasingly evaluate AI infrastructure around latency, availability, throughput, and operational consistency.

NVIDIA Partnership Strengthens Groq’s Infrastructure Position

Groq operates as an NVIDIA Cloud Partner, with certification to design, deploy, and operate NVIDIA accelerated computing according to NVIDIA’s reference architecture and operational standards. The relationship gives Groq a formal position within NVIDIA’s cloud ecosystem while allowing the company to expand access to NVIDIA-based infrastructure for customers. Planned participation from NVIDIA in the Series A also deepens the strategic relationship between the two companies. Together, the financing and partnership create a clearer path for Groq to build infrastructure around NVIDIA accelerated computing at greater scale.

The significance extends beyond the capital itself. AI cloud providers increasingly need to combine specialized infrastructure expertise with access to high-performance accelerator capacity, particularly as enterprises demand larger and more reliable clusters. Groq can use its operational experience across 13 data centers while expanding its NVIDIA-based offering to customers pursuing both training and inference. Meanwhile, its infrastructure expansion could give customers another route to deploy demanding AI workloads without building the underlying capacity themselves.

Groq Bets Inference Will Become AI’s Critical Infrastructure Layer

Alex Davis, Groq Executive Chairman and CEO of Disruptive, described the company’s ambition and the broader infrastructure shift directly: “We are building Groq into the world’s leading AI inference cloud. We look forward to continuing our partnership with NVIDIA at such an important juncture for the ecosystem. Inference will without a doubt become the largest and most critical layer of AI infrastructure. Our team has unmatched experience operating LPUs at scale and delivering the performance, efficiency and reliability that the next generation of AI demands. We will be focused on supporting the most important model makers to further contribute to the global growth of innovation and American ingenuity.”

That statement frames Groq’s strategy around the economics of running AI after models leave the training environment. Inference workloads can generate sustained infrastructure demand as applications serve users continuously, creating a different capacity profile from periodic training campaigns. Groq’s planned increase beyond 200 megawatts suggests the company expects that demand to translate into significant physical infrastructure requirements. The real test will come from how efficiently Groq can turn that additional capacity into dependable, geographically distributed inference services.

Groq’s Next Challenge Is Infrastructure Execution

Groq enters this expansion with a sizeable developer and enterprise audience, but scaling an AI cloud introduces operational complexity that funding alone cannot solve. The company will need to expand power capacity while maintaining performance and reliability across regions with different infrastructure conditions. Its NVIDIA Cloud Partner certification provides an established framework for deploying accelerated computing, but execution at several times the current power footprint will remain critical. In practical terms, Groq’s 2027 target will test whether its infrastructure model can scale as quickly as its financing.

The company’s $1 billion in recent funding gives it considerable financial capacity to pursue that expansion. Its 13-data-center footprint also provides a starting network from which to grow, rather than building a global presence from scratch. Still, the competitive landscape will depend on how Groq differentiates its inference cloud as hyperscalers, specialist AI clouds, and accelerator providers compete for the same enterprise workloads. The company’s emphasis on inference positions it around one of the fastest-changing segments of AI infrastructure.

Groq’s $3.5B Valuation Reflects The Inference Opportunity

The $3.5 billion valuation attached to the Series A signals investor confidence in the potential scale of AI inference infrastructure. Disruptive’s leadership of the round, alongside planned NVIDIA participation, gives the financing both financial and strategic weight. Groq now has a larger capital base to expand its facilities, increase accelerator availability, and support customers requiring larger AI clusters. The company’s stated trajectory toward more than 200 megawatts in 2027 makes that ambition measurable.

Groq’s next phase will ultimately be defined by infrastructure deployment rather than the size of its funding headline. If the company can convert its existing developer and enterprise reach into sustained inference demand, its planned capacity expansion could become a major component of its growth strategy. The combination of global facilities, NVIDIA accelerated computing, and inference-focused operations gives Groq a distinctive position in an increasingly crowded AI cloud market. For Compute Forecast, the more important signal is clear: AI inference is rapidly becoming a power, compute, and infrastructure planning problem of its own.