What the round covers

Groq has raised $350 million as it continues to pivot from an AI chipmaker to a neocloud company providing powerful GPUs and AI infrastructure services. The round was led by the investment firm Disruptive, with planned participation from Nvidia.

The new capital values the company at $3.5 billion. That is a little over half the $6.9 billion Groq was valued at last September. A spokesperson told TechCrunch that despite the difference in valuation, the company does not see it as a down round, but rather as establishing a new valuation for the "post-Nvidia-licensing-deal version of Groq".

Why the pivot happened

Groq was originally focused on building its own chips, dubbed LPUs, or language processing units. The goal was to compete with Nvidia on inference: the type of compute needed to run AI workloads in real time.

What changed the picture was losing the key team. Nvidia hired Groq's founder and CEO Jonathan Ross, along with other top talent, as part of a $20 billion licensing deal that the company paid out to investors. After it lost its star team, Groq shifted from being a pure AI chipmaker into a cloud and data center provider that operates Nvidia systems.

Current scale

The company's present position is described by these figures:

  • It operates 13 data centers across North America, Europe, the Middle East and Asia Pacific.
  • It says it serves more than 6 million developers, enterprises and AI-native companies.
  • In June it raised a $650 million round to kick off the pivot.
  • It intends to scale from 54 megawatts to more than 200 megawatts in 2027.

Groq says the fresh funds will support "those seeking usage of medium and larger sized clusters of Nvidia accelerated computing for training and inference". Alex Davis, Groq's chairman and CEO of Disruptive, framed the goal as "building Groq into the world's leading AI inference cloud".

How to read it

What this story really shows is how hard it is to build an alternative to Nvidia in inference hardware. Groq was attempting a technically different architecture and carried a valuation to match; after the team changed hands through a licensing deal, the company turned into a customer operating the systems of the firm it had set out to rival.

That the company frames the halved valuation as a redefinition rather than a down round is telling for the same reason: what was sold was the technology and the team, and what remains is data center operation. That can be a profitable business, but it is a different one; competing with Nvidia on your own chip and renting out Nvidia hardware do not provide the same moat. It should also be noted that the figures rest on company and investor statements, with no independent confirmation.