Groq Raises $350 Million to Pivot Toward AI Cloud Computing

In a notable shift within the AI infrastructure sector, the startup «Groq» announced it has raised new funding worth $350 million, confirming its gradual transition from a company specialized in designing AI chips to a provider of cloud services and infrastructure built on powerful graphics processing units.

The investment round was led by «Disruptive», with expected participation from chip giant «Nvidia», bringing Groq's market valuation to $3.5 billion. Although this figure appears high, it represents a significant decline compared to a valuation of $6.9 billion the company received last September.

Why Did the Valuation Drop?

The reason for this decline goes back to a licensing deal Nvidia struck, through which it recruited Groq's founder and CEO Jonathan Ross, along with a number of the company's top talents. With the loss of this distinguished team, Groq's features changed radically.

Nevertheless, a company spokesperson clarified that it does not consider this funding a «down round», but rather sees it as a logical reevaluation of what he described as the «post-Nvidia licensing deal version of Groq»—that is, a new phase with a completely different business model.

From Chip Manufacturing to Cloud Computing

Groq previously focused on developing its own chips known as «Language Processing Units» (LPUs), with the aim of competing with Nvidia in the field of «inference»—the type of computing required to run AI models in real time.

But after the departure of its founding team, the company shifted its compass to become a provider of data center and cloud computing services that rely on Nvidia's systems. The irony here is that Groq transformed from a potential competitor to Nvidia into one of its customers.

Ambitious Expansion Plans

Groq began its new journey last June by raising funding worth $650 million. The company plans to significantly expand its operational capacity over the coming years, as follows:

• Moving from a current capacity of 54 megawatts to more than 200 megawatts by 2027.
• Operating 13 data centers distributed across North America, Europe, the Middle East, and the Asia-Pacific region.
• Serving more than 6 million developers, institutions, and companies that rely on artificial intelligence.

The company confirms that the new funding will be directed toward supporting customers seeking to use medium and large clusters of enhanced Nvidia computing, whether for training or running models.

Betting on the Future of Inference

Alex Davis, Chairman of Groq and CEO of Disruptive, said: «We are building Groq to be the world's leading AI inference cloud». He added that inference will undoubtedly become the largest and most important layer in AI infrastructure.

Challenges of the «Neocloud» Model

Despite the growing demand for inference services as companies expand their adoption of AI, the question remains about the extent to which «Neocloud» companies can achieve sufficient profits to justify their massive investments over the long term.

«CoreWeave» stands out as an example of this; despite its strong revenue growth and signing major contracts with «Meta» and «Anthropic», investors remain concerned about its rising capital expenditures and heavy reliance on debt, as well as the risks of rapid hardware obsolescence.

Groq's financial data remains private for now, but its transformation places it at the heart of Nvidia's ecosystem. This relationship is not unique; Nvidia supplies companies like CoreWeave, Lambda, and Nebius with graphics processing units, while simultaneously injecting billions of dollars into some of them as part of a frantic race to build more computing power.

✦ بقلم فريق دروب أيديا