Groq, the AI inference startup that lost its founder and core team to Jensen Huang's Nvidia, has secured $350 million in Series A funding and plans to expand its computing capacity by more than 270 percent within the next year. The move marks a dramatic pivot: Groq shifted from competing directly with Nvidia's chips to operating as a Nvidia-powered cloud partner, a transformation that raises questions about whether the company can succeed in its new role.

What Happened When Nvidia Poached Groq's Leadership?

In December 2025, Nvidia reached a technology licensing deal with Groq worth approximately $20 billion. As part of that agreement, Nvidia brought in Groq's founder and CEO Jonathan Ross, then-President Sunny Madra, and multiple core employees to work at Nvidia. The deal was not a traditional acquisition; Groq remained independent but fundamentally changed its business model.

Before the Nvidia deal, Groq had built its reputation on developing Language Processing Units, or LPUs, specialized chips designed to handle AI inference tasks with lower latency and higher throughput than Nvidia's GPUs (Graphics Processing Units). The company had raised more than $1.75 billion from investors including Social Capital, Tiger Global, and BlackRock. But the loss of its founder and leadership team created an existential challenge: how could Groq compete without the visionary who built it ?

The answer was to stop competing and start collaborating. After the Nvidia deal closed, Groq pivoted to operating AI data centers powered by Nvidia's chips rather than its own LPUs. The company now positions itself as a Nvidia cloud partner, designing and deploying accelerated computing clusters according to Nvidia's reference architecture and standards.

How Is Groq Expanding Its Computing Power?

Groq's latest $350 million funding round, led by Disruptive and with Nvidia planning to participate, will fuel an aggressive expansion of its data center footprint. The company currently operates 13 data centers across North America, Europe, the Middle East, and Asia-Pacific, serving more than 6 million developers, Fortune 500 companies, and thousands of AI-native startups. The platform processes trillions of tokens every week.

• Current Computing Scale: Groq operates at 54 megawatts of computing power today, a baseline for training and running AI models at scale.
• Planned Expansion: By 2027, Groq plans to increase total computing power to more than 200 megawatts, representing a 270 percent increase in capacity.
• Funding Timeline: In June 2026, Groq secured $650 million in bridge funding to support its strategic transformation; combined with this $350 million round, the company has raised $1 billion in financing within two months.

The expansion is designed to serve major AI model developers who need massive amounts of inference computing power. At Nvidia's GTC conference in March 2026, the company unveiled the Groq 3 LPU and Groq 3 LPX inference accelerator racks built on Groq's technology. When paired with Nvidia's Rubin GPU, these systems can increase inference throughput per megawatt by 35 times when running trillion-parameter models, meaning they can process vastly more AI requests simultaneously.

Why Does This Deal Matter for the AI Infrastructure Race?

The Groq-Nvidia partnership illustrates a broader shift in the AI industry: inference, the process of running trained AI models to generate responses, is becoming as critical as training. While training large language models requires enormous upfront investment, inference happens billions of times per day as users interact with AI chatbots, search engines, and other applications. The computing infrastructure needed to handle inference at scale is becoming one of the largest and most expensive components of AI systems.

"Groq will focus on serving major model developers, and believes that AI inference will become one of the largest and most critical AI infrastructure links," stated Alex Davis, Executive Chairman of Groq and CEO of Disruptive.

Alex Davis, Executive Chairman of Groq and CEO of Disruptive

However, Groq's transformation comes with significant risks. The company's valuation dropped approximately 49 percent from its $6.9 billion peak in September 2025 to $3.5 billion post-money valuation in this round. While Groq's leadership characterized this as a "revaluation" rather than a "down round" following the Nvidia technology licensing transaction, the decline reflects market skepticism about the company's ability to execute in its new role.

Groq has not disclosed critical metrics including current revenue, customer payment scale, data center utilization rates, or profitability. The company also has not specified the respective proportions of Nvidia GPUs versus self-developed LPUs in its new computing infrastructure. These gaps make it difficult for investors and analysts to assess whether Groq can generate stable revenue and compete effectively in the crowded data center market.

What Challenges Does Groq Face in Its New Business Model?

Groq's shift from chip maker to cloud operator requires the company to master an entirely different set of skills. Operating profitable data centers demands expertise in power management, cooling systems, customer acquisition, pricing strategy, and operational efficiency. The company must now compete not just on technology but on cost per computation, customer service, and reliability.

Additionally, Groq still operates LPU-related businesses alongside its Nvidia-powered infrastructure. This creates a complex relationship with Nvidia that is neither purely competitive nor purely cooperative. Groq must prove that its LPU technology remains valuable while simultaneously betting its future on Nvidia's hardware and software ecosystem. If Nvidia's GPUs become sufficiently powerful and cost-effective, Groq's LPUs may become redundant.

The company's success will ultimately depend on whether it can acquire and retain major customers, maintain high data center utilization rates, and control costs in a capital-intensive business. The $1 billion in financing secured in recent months provides runway, but the path to profitability remains uncertain. Market observers will be watching closely to see whether Groq can transform from a chip startup into a sustainable cloud infrastructure provider.