Groq raised $350 million in a Series A round announced August 17, 2026, at a $3.5 billion valuation, led by Dallas-based Disruptive with participation from NVIDIA. The funding represents roughly half the $6.9 billion valuation Groq commanded in September 2025, before NVIDIA hired founder Jonathan Ross and ~90% of its engineering team under a reported $20 billion technology licensing deal.
The company has pivoted from building its own AI chips (Language Processing Units) to operating a neocloud data center business focused on NVIDIA-accelerated inference. Groq now operates 13 data centers globally across North America, Europe, Middle East, and Asia Pacific, serving 6+ million developers and enterprises. Combined with a $650 million raise in June 2026, Groq has raised $1 billion in two months.
Groq plans to scale from 54 megawatts current capacity to 200+ megawatts by 2027. The company reports handling trillions of tokens weekly and positions itself as an NVIDIA Cloud Partner certified to deploy NVIDIA's reference architecture, including the new LPX systems that incorporate licensed Groq inference technology.
For infrastructure teams evaluating inference capacity, Groq's pivot signals that post-licensing incumbent AI chipmakers are consolidating neocloud capability through technology transfer rather than competition. The company's developer base and 13-DC footprint provide credible inventory against CoreWeave and Lambda, but without founder/tech differentiation, execution on 200MW expansion and profitability is now the only moat.