Groq announced it is raising up to $650 million from existing investors to launch a new company called Groq2, which will focus on AI neocloud services centered on inference processing rather than chip manufacturing [1, 2, 3].
The company plans to shift away from its original chip business toward building AI inference cloud platforms, a segment currently experiencing greater demand than model training [1, 2, 3]. Interim CEO Adam Winter and CFO Matt Eng are leading the new strategic direction [2, 3].
Groq’s new funding round benefits from strong backing. Key investors Disruptive and Infinitium have committed to cover the $650 million target if other participants do not invest pro-rata [2, 3].
This fundraising effort follows a major deal in December 2025, when Groq signed a $20 billion licensing and non-acquisition agreement with Nvidia. The deal includes licensing Groq’s hardware technology to Nvidia and the transfer of senior Groq employees to Nvidia [1, 2, 3].
Separately, the Chinese Industrial and Information Technology Ministry has been supporting research into space-based computing — the deployment of computing power on satellites. This emerging field could grow into a market valued at over $1 trillion by 2030 [1].
Groq’s plan to launch Groq2 aims to capitalize on AI inference cloud services, seeking to leverage its technology and investor support while moving beyond chip design. The company began raising the $650 million round in May 2026 [1, 2, 3].
