Groq has announced $650 million in new growth capital to expand its AI inference cloud. The round was led by Disruptive and Infinitum, with existing investors also choosing to reinvest. The money will go toward fitting out Groq’s current data centers with its latest inference technology, including NVIDIA’s new LPX system, with a target of 200 MW of capacity by the end of 2027.
The company already runs 13 data centers across North America, Europe, the Middle East, and Asia-Pacific. It serves more than five million developers and thousands of AI-native companies, processing trillions of tokens each week. That scale puts Groq in a strong position as demand for AI inference keeps climbing.
This raise comes at a moment when the AI industry is shifting from model training to model deployment. Training gets most of the headlines, but running AI models in production, which is what inference means, is expected to require 15 to 20 times more compute over time. Most AI clouds were built for training, and no single company has claimed clear leadership in inference. That gap is exactly what Groq is trying to fill.
The company’s current direction took shape in December 2025, when Groq signed a non-exclusive licensing agreement with NVIDIA. At NVIDIA’s GTC conference earlier this year, NVIDIA announced its next-generation LPX platform incorporating Groq’s inference technology. Those milestones pushed Groq’s board and lead investors to focus the company entirely on one goal: becoming the go-to infrastructure layer for AI inference at scale.
Groq chairman Alex Davis, who is also founder and CEO of Disruptive, said the company has spent years building the technology and operational expertise needed for this moment. John Yetimoglu, Groq board member and founder of Infinitum, described inference as “the largest infrastructure market in technology” and pointed to Groq’s combination of technology, operating know-how, and global reach as the reason for backing the company.
The leadership team has also been updated. Groq is chaired by Alex Davis and led by CEO Adam Winter and CFO Matt Eng, both of whom have been with the company through its growth. They are joined by:
- Alan Rice as Chief Operating Officer, previously at xAI (now SpaceXAI) and Meta Datacenters, with an earlier career in U.S. Navy nuclear submarine operations
- Sinclair Schuller as Chief Technology Officer, starting in July, who founded enterprise cloud platform Apprenda, sold it to Atos, and later co-founded Nuvalence
- Rakesh Malhotra as Chief Product Officer, also starting in July, who co-founded Nuvalence with Schuller and previously spent about a decade at Microsoft leading cloud, data center management, and enterprise storage products
Nuvalence, the software engineering firm Schuller and Malhotra built together, was acquired by EY in 2024. Their combined background in enterprise software and infrastructure buildouts is the kind of experience Groq needs as it moves from a developer-focused platform toward broader enterprise adoption.
The core argument Groq is making to the market is straightforward. Most companies building AI products care about three things when running models: speed, reliability, and cost. Groq says its approach, built around its own Language Processing Units rather than traditional GPUs, delivers better performance per dollar for demanding applications. With NVIDIA now incorporating Groq’s technology into its LPX platform, Groq has also shown it can work within the broader AI infrastructure ecosystem rather than against it.
The inference market is still early, but the direction is clear. As more companies move AI from pilots into production systems, the infrastructure that runs those systems at scale will matter enormously. Groq is betting that its head start in inference-specific hardware, combined with an operational team that has hands-on experience running these systems globally, gives it an edge that will be hard to replicate quickly.
