When a major competitor pays a substantial licensing fee for a startup's intellectual property and walks away with its top executives, the path forward for what remains of that company is rarely straightforward. For AI chip startup Groq, that path has involved raising fresh capital, rebuilding its leadership bench, and refocusing the business around a different product line entirely.
Groq announced on Monday that it has closed a new funding round totaling $650 million, confirming what had been reported in the weeks prior. The round was co-led by Disruptive, a Dallas-based late-stage investment firm whose founder, Alex Davis, also serves as chairman of Groq's board, and Infinitum, a hedge fund headquartered in Fort Lauderdale. The company did not announce a new valuation as part of the raise. Its most recent publicly known valuation stood at $6.9 billion, established after a $750 million funding round completed in September of last year.
The new capital arrives approximately six months after Nvidia entered into a non-exclusive licensing agreement to use Groq's chip technology, a deal that was accompanied by the departure of several of the company's most prominent figures. Among those who left for Nvidia were Jonathan Ross, who co-founded Groq and had been serving as its CEO, and Sunny Madra, who had been the company's president. Other employees followed as well.
Origins of the Company and Its Chip Technology
Ross built his reputation in the AI hardware space while working at Google, where he played a significant role in developing the company's Tensor Processing Unit, widely known as the TPU - a specialized chip designed to accelerate machine learning workloads. He later partnered with fellow Google engineer Doug Wightman to start Groq roughly ten years ago. When the Nvidia deal was finalized, Wightman chose to remain with the company and stepped into the CEO role.
Groq's core hardware product was a chip the company branded as a language processing unit, or LPU, which was purpose-built for AI inference tasks. The company offered access to this technology either through a cloud-based service or as part of on-premises hardware clusters that customers could deploy directly. The LPU was positioned as a faster and more efficient alternative to traditional GPU-based inference setups.
Following the licensing deal, Nvidia incorporated the acquired IP into its own product line. At its GTC conference in March, Nvidia unveiled the Nvidia Groq 3 LPX inference hardware system, a cluster-based solution that drew directly from the technology it had licensed from Groq.
Pivoting to the Neocloud Business
With its core chip IP now shared with one of the largest and most powerful companies in the semiconductor industry, Groq has repositioned itself around its neocloud infrastructure business. This segment of the company had previously been overseen by Madra following Groq's 2024 acquisition of Definitive Intelligence, an AI data analytics firm he had founded.
According to the company, the neocloud operation has expanded to encompass 13 data centers spread across North America, Europe, the Middle East, and the Asia-Pacific region. Groq says it is now serving more than five million developers along with thousands of AI companies, and that its infrastructure is processing trillions of tokens on a weekly basis. The inference cloud business, in other words, has scaled considerably even as the company underwent significant internal disruption.
New Leadership Joining the Ranks
Alongside the funding announcement, Groq has been actively filling the executive gaps left by the departures tied to the Nvidia deal. The company has brought in Alan Rice as its new Chief Operating Officer. Rice comes to the role following stints at xAI, Elon Musk's AI venture, and at Meta, and has a background that includes a career in the United States Navy.
Two additional hires round out the new leadership team. Sinclair Schuller has joined as Chief Technology Officer, and Rakesh Malhotra has come aboard as Chief Product Officer. The two have a long history of working together. Schuller previously founded Apprenda, an enterprise cloud software company, where Malhotra also worked. The pair later co-founded Nuvalence, a software engineering consultancy that was acquired by professional services firm EY in 2024. Before his work with Schuller, Malhotra spent close to a decade contributing to cloud product development at Microsoft.
What Lies Ahead for Groq
The central question facing Groq now is whether a company that came close to being absorbed by a competitor can carve out a durable and competitive position in the AI infrastructure market. The conditions are not entirely unfavorable. Demand for AI inference technology has surged in recent years, drawing significant venture capital investment and prompting a wave of new entrants and product development across the industry.
At the same time, the competitive dynamics are intensifying. Groq no longer holds exclusive control over the hardware technology that originally differentiated it, and it now operates in a space where hyperscalers, chip manufacturers, and well-funded startups are all racing to capture inference workloads.
There is some precedent for companies recovering from similar situations. Scale AI CEO Jason Droege recently told Forbes that his company has bounced back following a $14.3 billion deal with Meta that similarly saw key talent depart, and that Scale AI is now on track to generate $1 billion in annual revenue. The parallel is not exact, but it offers a data point suggesting that companies can rebuild and grow even after absorbing major structural shocks.
For Groq, the combination of fresh funding, a growing data center footprint, millions of active developers on its platform, and a newly assembled executive team gives it a foundation to work from. Whether that foundation proves sufficient in an increasingly crowded and fast-moving market remains to be seen, but the company is clearly not standing still.
