Groq Raises $650M to Expand AI Inference Cloud Toward 200 MW by 2027

Groq announced $650 million in new growth capital on June 22, 2026, to expand its AI inference cloud. The round was led by Disruptive and Infinitum, with participation from investors who chose to reinvest. The company operates 13 data centers across North America, Europe, the Middle East, and APAC, serving more than five million developers and processing trillions of tokens each week. The new capital will accelerate fitting out its existing footprint with Groq's latest inference technology, including NVIDIA's LPX system, and Groq expects to scale toward 200 MW by the end of 2027.

Leadership and Strategy

Groq's board and lead investors worked with management to sharpen the company's focus on building what it calls the world's leading AI inference cloud. This follows a non-exclusive licensing agreement with NVIDIA in December 2025 and NVIDIA's announcement of its next-generation LPX platform at GTC this year. Groq says its near-term competitive advantage is operational expertise in AI inference, supported by what it believes is the only engineering team with hands-on experience operating LPUs at scale.

The company is chaired by Alex Davis of Disruptive, with Adam Winter as CEO and Matt Eng as CFO. New appointments include Alan Rice as COO, previously at xAI (now SpaceXAI) and Meta Datacenters, after a career in U.S. Navy nuclear submarine operations. Starting in July, Sinclair Schuller will join as CTO and Rakesh Malhotra as CPO. The pair previously worked together at Apprenda, the enterprise cloud platform Schuller founded and later sold to Atos, and then co-founded Nuvalence, acquired by EY in 2024. Malhotra earlier spent about a decade at Microsoft leading cloud, data-center-management, and enterprise-storage products.

The Inference Opportunity

Groq frames inference as a far larger opportunity than training, estimating that inference will demand 15 to 20 times more compute over time. The company notes that most AI clouds today are built for training and that no company yet leads the inference category. The challenge, it says, is delivering fast, reliable output at a competitive price, and Groq believes its approach will offer the best value-to-performance for demanding AI applications.

For inference engineers and model builders, this funding signals a push to scale dedicated inference infrastructure, not just add capacity. The move toward 200 MW by 2027 and the integration of NVIDIA's LPX system point to a focus on high-throughput, low-latency serving at scale. The leadership hires bring experience from hyperscale data centers and enterprise platforms, which could shape how Groq's cloud handles production workloads. The company's claim that inference will require 15 to 20 times more compute than training, if accurate, underscores the growing demand for efficient inference solutions as AI moves from experimentation to production.