Groq Seeks $650 Million in New Funding Round to Power AI Inference Cloud Expansion

Groq, the Silicon Valley‑based artificial intelligence chip startup, is targeting a $650 million fundraise from its existing investor base as it doubles down on becoming a leading AI inference cloud provider. The move comes months after the company completed one of the most significant technology licensing agreements in recent memory, a deal with Nvidia valued at approximately $20 billion.

The fundraise, which was confirmed by sources familiar with the matter, marks a pivotal chapter for Groq as it formally transitions away from chip hardware sales toward operating as an AI inference neocloud. Existing backers Disruptive and Infinitum have agreed to backstop the full $650 million should other existing shareholders choose not to participate pro‑rata, meaning the raise is effectively underwritten.

What the Funding Will Be Used For

Groq's new direction centres on inference computing, the processing layer that activates every time a user submits a prompt to an AI model. As AI adoption accelerates across enterprise and consumer applications, the demand for fast, cost‑efficient inference infrastructure has grown considerably. Unlike model training, inference happens continuously at scale, making it one of the most capital‑intensive segments of the AI stack.

The company's inference cloud platform, which allows developers and enterprises to host inference‑heavy applications, has already attracted substantial developer interest. Groq's Language Processing Unit (LPU) chips are positioned as a faster, more cost‑effective alternative to conventional GPUs for inference tasks.

Capital raised from this round is expected to be directed toward:

• Scaling capacity on Groq's inference cloud platform
• Expanding enterprise customer deployments
• Strengthening infrastructure for what the company internally refers to as "Groq 2.0"

The Nvidia Deal and Its Aftermath

The $650 million raise follows a December 2025 licensing agreement with Nvidia, which involved a non‑exclusive arrangement granting Nvidia access to Groq's hardware technology. The deal, valued at approximately $17 to $20 billion depending on the reporting, resulted in the departure of several senior Groq executives who moved to the chip giant.

Groq's investors received cash distributions tied to the Nvidia agreement, with a final payout still expected. Those same investors are now being asked to reinvest in the company's next phase. The structure signals strong investor confidence in Groq's ability to build a standalone, recurring revenue business in the inference cloud market.

The round is being led under the watch of Groq's current management team, with Adam Winter serving as interim CEO and Matt Eng as CFO. Together, they are steering the company's pivot toward a services‑first model.

Groq 2.0: A New Identity in AI Infrastructure

Groq's reorientation reflects a broader shift playing out across the AI infrastructure sector. While GPU manufacturers continue to race for dominance in model training, the inference layer has emerged as the next major battleground. Enterprises deploying AI at scale need dependable, low‑latency inference capacity, and neocloud providers are increasingly stepping in to fill that gap.

Groq had previously raised $640 million in a Series D round at a $2.8 billion valuation, a round that included participation from BlackRock, Neuberger Berman, Cisco, Samsung Catalyst Fund, and others. The company also counts Bell Canada among its major infrastructure partners following an exclusive partnership for AI compute services.

With the Nvidia licensing deal providing both validation and liquidity, Groq now enters its next chapter better capitalised than most inference‑focused startups. The competitive field includes major hyperscalers as well as emerging neocloud platforms, but Groq's proprietary LPU architecture and track record give it differentiated footing.

For more information on Groq's cloud infrastructure and developer offerings, visit the official platform at groq.com.

The fundraise is not yet closed, and terms could still evolve, but the backstop commitments from Disruptive and Infinitum make full completion a near‑certainty. As AI inference demand continues to climb globally, Groq's $650 million push positions it as one of the more serious contenders in the infrastructure race.