Groq raises $750M as inference demand surges, partners with DOE and Saudi Arabia for global expansion

AI chip startup Groq just closed a $750 million funding round, the company disclosed Wednesday, as the rush for faster inference shows no sign of cooling. The new money, coming on the heels of a $640 million raise last year and a prior $2.8 billion valuation, underscores just how fierce the competition among AI infrastructure providers has become.

Alongside the funding, Groq also revealed a non-exclusive licensing agreement with Nvidia, signaling a pragmatic shift in its hardware playbook. Under that deal, Groq will weave Nvidia's GPU technology into its own inference stack, letting customers run models on either architecture through a single API. Financial terms stayed under wraps.

Government and sovereign AI partnerships

On the public-sector front, Groq inked a partnership with the U.S. Department of Energy to push forward AI inference and next-gen computing infrastructure. The collaboration, which covers high-performance computing workloads, cements Groq's role in sovereign AI. The company recently landed on a shortlist of the top 10 national security tech firms in the United States.

Global expansion: Saudi Arabia, Asia-Pacific, and Europe

Saudi Arabia is putting $1.5 billion behind an AI-powered economy built with Groq. Aramco Digital, the digital arm of Saudi Aramco, is moving ahead on what it bills as the world's largest inferencing data center in the kingdom, following a memorandum of understanding signed at the LEAP conference. The facility is expected to house tens of thousands of Groq's Language Processing Units.

Down in Asia-Pacific, Groq opened a new data center in Sydney, Australia, dedicated to AI inference for the region. That site will serve local enterprises and government agencies. Meanwhile, the company launched a European data center in Helsinki, Finland, and struck a deal with Bell Canada to become the exclusive inference provider for Bell's sovereign AI network.

Paytm, McLaren, and enterprise deployments

Groq also partnered with Indian fintech heavyweight Paytm to deliver real-time AI for payments and platform intelligence. On the enterprise side, Groq powers HUMAIN One, a real-time AI operating system, and launched OpenAI's new open models on day zero. McLaren Racing named Groq an official partner of its Formula 1 team, a sign the company is pushing into sports and entertainment.

Leadership and financial outlook

Simon Edwards has been appointed Chief Financial Officer, a move that hints at preparations for a potential IPO or additional financing. The fresh capital will go toward scaling data center capacity, beefing up the engineering team, and speeding up product development.

Industry context

Groq goes head-to-head with Nvidia in the inference market, but the licensing agreement suggests a more nuanced dynamic, one where Groq may be hedging its bets on LPU adoption. Nvidia's GPUs still dominate training, but inference has become a battleground as models like Llama 4, Grok-3, and DeepSeek-R1 push latency limits. Groq claims its LPUs deliver lower latency for transformer-based models, a pitch that has caught the eye of hyperscalers.

The DOE partnership and Saudi expansion also underline how AI infrastructure is turning into a geopolitical chess game. As nations race to build sovereign AI capabilities, companies like Groq are positioning themselves as essential enablers.