Groq, the AI inference startup that once competed directly with Nvidia, has secured $350 million in Series A funding and plans to expand its computing capacity from 54 megawatts to over 200 megawatts by 2027. This transformation marks a dramatic shift for the company, which pivoted from building its own specialized chips to operating data centers powered by Nvidia hardware after a $20 billion technology licensing deal in December 2025.

What Happened to Groq's Original Mission?

Groq was founded in 2016 by Jonathan Ross, a former Google engineer who helped develop Google's TPU (Tensor Processing Unit) chips. The company built its reputation on Language Processing Units, or LPUs, specialized processors designed to handle AI inference tasks with lower latency and higher throughput than Nvidia GPUs. For nearly a decade, Groq positioned itself as a direct competitor to Nvidia in the AI inference market.

That competitive relationship changed dramatically in December 2025. Nvidia reached a non-exclusive technology licensing agreement with Groq valued at approximately $20 billion, and brought Jonathan Ross, then-President Sunny Madra, and multiple core team members into Nvidia. The deal was structured as a licensing agreement rather than an acquisition, allowing Groq to continue operating independently while fundamentally transforming its business model.

"Groq plans to expand its computing power scale from 54 megawatts to over 200 megawatts next year," the company announced in its latest funding disclosure.

Groq Official Statement

How Is Groq Operating Now?

After the Nvidia technology licensing transaction, Groq shifted from developing self-designed chips to operating an AI inference cloud service. The company now functions as a Nvidia cloud partner, designing, deploying, and operating accelerated computing clusters based on Nvidia's reference architecture and operational standards. This represents a complete business model transformation for the startup.

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 companies. The platform processes trillions of tokens every week, positioning Groq as a significant player in the AI infrastructure space despite its pivot away from chip manufacturing.

Groq's new financing round, led by Disruptive and with Nvidia planning to participate, will fund the expansion of its Nvidia chip-based computing clusters. The company's valuation has dropped approximately 49% from its September 2025 peak of $6.9 billion to $3.5 billion post-money, though Groq disputes characterizing this as a "down round," instead framing it as a revaluation following the Nvidia technology licensing transaction.

Why Does This Matter for AI Infrastructure?

Groq's transformation reflects a broader trend in AI infrastructure: the shift from chip competition to computing capacity competition. As demand for AI training and inference computing power has exploded, no single company can meet global demand alone. This has forced former rivals to become essential partners.

The AI boom has created unprecedented demand for specialized hardware and computing resources. Nvidia CEO Jensen Huang demonstrated the severity of component shortages in June 2026 when he wrote "Please make more" on an SK Hynix memory wafer during an annual computer exhibition in Taipei, highlighting how critical supply chain partnerships have become.

Groq's expansion plans are ambitious. The company intends to increase its total computing power scale by at least 270% over the next year, from its current 54 megawatts to more than 200 megawatts in 2027. This expansion will require significant capital investment and operational expertise, both of which the new funding round is designed to support.

How to Understand Groq's New Business Model

• Infrastructure Provider: Groq now operates as a data center company offering AI training and inference computing power to external customers, rather than selling proprietary chips directly to end users.
• Nvidia Partnership: The company designs and deploys computing clusters according to Nvidia's specifications and standards, making it a cloud partner rather than a competitor in the chip market.
• Hybrid Technology Approach: While Groq continues operating some LPU-related businesses, the bulk of its new computing capacity is built on Nvidia GPUs, creating a mixed technology environment that hasn't been fully detailed publicly.

Groq's previous funding history demonstrates significant investor confidence despite the recent valuation adjustment. Before the Nvidia technology licensing transaction, the company had raised more than $1.75 billion through seed rounds, Series A through D, and additional financing from investors including Social Capital, Tiger Global, D1 Capital, BlackRock, Cisco, and Samsung. Including the $650 million bridge funding received in June 2026 and the current $350 million Series A round, Groq has secured over $1 billion in financing within just two months.

What Challenges Lie Ahead?

Groq's transformation from chip maker to data center operator introduces new business challenges. The company must now prove it can achieve strong utilization rates across its data centers, acquire and retain customers, and manage operational costs effectively. These are fundamentally different competencies from chip design and manufacturing.

Additionally, Groq has not disclosed critical operational metrics including current revenue, customer payment scale, data center utilization rates, or profitability. The company also hasn't specified the respective proportions of Nvidia GPUs versus self-developed LPUs in its new computing infrastructure, leaving questions about its long-term technology strategy unanswered.

The partnership between Groq and Nvidia is neither purely competitive nor purely cooperative. Nvidia currently only "plans to participate" in the funding round, with no official announcement of its investment amount. This ambiguity reflects the complex nature of their relationship, where Nvidia has acquired Groq's technology and key talent while allowing the company to operate independently as a customer-facing infrastructure provider.

As AI infrastructure becomes increasingly critical to global technology development, companies like Groq that can rapidly scale computing capacity will play essential roles in supporting the next generation of AI model development and deployment. Whether Groq can successfully execute its transformation from specialized chip maker to large-scale data center operator will significantly influence the competitive landscape of AI infrastructure over the coming years.