Groq lines up $650M internal raise to fund ‘Groq 2.0’ AI inference push

AI chip company Groq is reportedly raising up to $650 million in new capital from its existing investors to finance what insiders are calling its second phase, centered on an AI inference "neocloud" strategy. Axios and Reuters report that current shareholders are being invited to participate on a pro rata basis, with long-time backers Disruptive and Infinitum prepared to backstop the full target amount if other investors do not take their full allocations. StartupHub.ai also characterizes the transaction as an internal funding round and notes that the capital is intended to support Groq’s shift from a pure hardware model toward optimizing AI model response times. The raise follows Groq’s earlier $20 billion licensing agreement with Nvidia, which provided substantial returns and cash distributions to shareholders and was accompanied by significant executive turnover.

Public filings and secondary market data indicate that Groq had previously raised roughly $1.8 billion, including a $750 million Series D-3 round in September 2025 at a post-money valuation of about $6.9 billion and multiple Series D-related tranches in 2024 at a $2.8 billion valuation. The new $650 million raise has been reported as an internal financing for a "Groq 2.0" structure rather than as a labeled Series E or later equity round, and none of the sources to date disclose a specific valuation for the transaction. For Groq, the deal is intended to fund expansion of its inference cloud offering built on its own AI chips and systems, giving existing investors a way to re-up into the business after securing liquidity from the Nvidia licensing deal. At an industry level, the round underscores ongoing investor appetite for specialized AI infrastructure and inference platforms, and highlights how non-acquisition licensing arrangements with hyperscalers can recycle capital into follow-on private financings.