Groq has secured $650 million in new growth capital to accelerate the expansion of its AI inference cloud.
The round was led by Disruptive and Infinitum, with participation from investors who elected to reinvest in the company. Six months after the company handed over its founder, president, and core engineering team to Nvidia in a deal worth roughly $20 billion, it is back under new leadership and betting everything on the one market it was originally built for.
What Led to This
In December 2025, Nvidia agreed to a non-exclusive license to Groq’s technology. Groq’s founder Jonathan Ross, who helped Google start its AI chip program, as well as Groq President Sunny Madra and other members of its engineering team, joined Nvidia. The deal was structured as a licensing and talent arrangement rather than a straight acquisition, which allowed it to close unusually fast. And the official arrangement saw Nvidia acquire Groq’s assets and license its inference technology, while also absorbing the startup’s top leadership, with Nvidia agreeing to buy most of Groq’s AI chip assets for $20 billion in cash as part of the key terms.
This deal was good news for existing investors, who got paid out in cash in what would have been Nvidia’s largest purchase if the arrangement had been structured as a full acquisition. Those same investors have now been asked to back the company’s next phase.
The New Groq
Following the December milestones, Groq’s board and lead investors, Disruptive and Infinitum, worked alongside management to sharpen the company’s strategic focus around a single opportunity, which was to build the world’s leading AI inference cloud.
The company is now led by Chief Executive Officer Adam Winter and Chief Financial Officer Matt Eng, both long-standing leaders who spent years building and scaling the company’s technology, infrastructure footprint, and commercial operations. The leadership team has also expanded with new additions.
Chief Operating Officer Alan Rice, who was previously at xAI and Meta Datacenters, joins the company after an earlier career in the U.S. Navy. Starting in July, the company is also appointing Chief Technology Officer Sinclair Schuller and Chief Product Officer Rakesh Malhotra, longtime partners who worked together at Apprenda, the enterprise cloud platform Schuller founded and later sold to Atos. The pair then co-founded Nuvalence, a software engineering and digital transformation firm acquired by EY in 2024.
What Groq Is Building
The company’s core technology is the Language Processing Unit (LPU), a chip it designed specifically for AI inference. LPUs run large language models and other leading models at substantially faster speeds and, on an architectural level, up to 10x more efficiently from an energy perspective compared to GPUs. Unlike general-purpose GPUs, which were built for training workloads and adapted for inference, the LPU was designed from the ground up to meet the unique needs of AI.
The new capital will support the deployment of Groq’s latest inference technology across its existing data center network, including Nvidia’s new LPX system. Nvidia’s integration of Groq’s architecture goes further than just the licensing agreement. At GTC 2026, Jensen Huang highlighted the Groq 3 LPU, which has been integrated into the next-generation Vera Rubin platform, working alongside GPUs to form an inference acceleration ecosystem. Huang pointed out that with the growth of AI agent applications, inference workloads are becoming increasingly diversified, and some tasks prioritize interactivity and ultra-low latency where traditional GPUs may exhibit performance redundancy.
Today, Groq operates 13 data centers across North America, Europe, the Middle East and Asia-Pacific. The company serves more than five million developers and thousands of AI-native companies that consume trillions of tokens each week. Groq expects to scale toward 200 megawatts by the end of 2027.
Why the Inference Market Matters
The category Groq is betting on is one of the fastest-growing areas of AI infrastructure. Running AI models, known as inference, is a far larger opportunity than training them. Over time, inference will demand an estimated 15 to 20 times more compute. Most AI clouds today are built for training and no company yet leads the inference category.
The neocloud market already has CoreWeave and Lambda Labs scaling aggressively, making LPU speed the only defensible differentiator Groq has. And Speed is where the company’s pitch is strongest.
Groq’s board chair Alex Davis, who is also founder and CEO of lead investor Disruptive, said the company has spent years building the technology, infrastructure, and operational expertise required for its next phase. John Yetimoglu, Groq board member and founder and chief investment officer of Infinitum, said inference will become the largest infrastructure market in technology and that as AI moves from experimentation to production, demand for reliable, cost-efficient inference will continue to grow exponentially.
The company that was effectively hollowed out by Nvidia six months ago now has fresh capital, a new executive team, and a clear market thesis. Whether rebuilding from that point is viable against better-resourced competitors is the question the next few years will answer.
