Groq, which develops technology for running trained AI models (a process called inference), raised $350 million at a $3.5 billion valuation—a 49% drop from its $6.9 billion valuation nine months earlier.
The decline follows Groq's shift from selling hardware to operating a cloud-based inference service and Nvidia's $20 billion licensing deal for Groq's technology in December 2025, which also brought founder Jonathan Ross, President Sunny Madra, and other Groq staff to Nvidia.
For Nvidia, the investment and talent acquisition strengthen its position in inference, where customers increasingly prioritize cost and speed as deployed AI models handle massive numbers of requests.
What happened
Groq, an AI inference startup, raised $350 million at a $3.5 billion valuation in a round led by Disruptive and including Nvidia. This represents a roughly 49% decline from the company's $6.9 billion valuation after raising $750 million last September.
Why it matters
The valuation drop reflects Groq's dramatic business pivot away from selling proprietary hardware toward operating a cloud-based inference service, following Nvidia's December 2025 licensing agreement for Groq's inference technology (reportedly worth $20 billion) and the departure of founder Jonathan Ross, President Sunny Madra, and other employees to Nvidia. For Nvidia, the relationship is strategically important as inference—running already-trained AI models to handle billions of user requests—becomes increasingly critical alongside training.
What to watch
Nvidia's upcoming data-center results should show whether inference demand is becoming a larger growth driver. The valuation also signals that massive AI infrastructure spending does not guarantee rising valuations for startups whose business models shift fundamentally.
Ask the AI about this article →
Groq's valuation collapse from $6.9 billion to $3.5 billion in under a year is striking, but the shift reflects a fundamental reordering of the company's business and strategic role. The catalyst was Nvidia's December 2025 licensing agreement for Groq's inference technology—a non-exclusive deal reportedly valued at $20 billion that also pulled Groq's founder, president, and other key staff into Nvidia. What remains is no longer a hardware competitor but a cloud-based inference service operator, a business model that investors are valuing at half its prior price.
For Nvidia, the move is strategically coherent. While training has driven enormous GPU demand, the company recognizes that inference—the repeated execution of already-trained models serving billions of user requests—is where the next wave of AI computing growth and margin pressure will emerge. By licensing Groq's inference technology, absorbing its talent, and now investing in the company's remaining cloud business, Nvidia is attempting to lock in influence over a critical phase of the AI stack. The combination of technology licensing, personnel acquisition, and equity investment gives Nvidia multiple levers to defend its position as the center of gravity shifts from model training toward inference at massive scale.
For example, today's edition would include:
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. Q&As are published on this page for other readers too.
Visko raised $10 million in pre-seed funding from Llama Ventures and opened public access to its first foundat…
U.S. markets ended August higher, with the S&P 500 up 2.6% and the Nasdaq up 3.9%
Neurovia AI, an Abu Dhabi-based company, is pitching Saudi security agencies software that it says can compres…
AI company Runway has unveiled Solaris, the first model in a new category it calls "Interface World Models." I…
John Deere introduced JD, a conversational AI tool that lets farmers ask open-ended questions about their hist…
Nvidia CEO Jensen Huang said on Fox Business that AI is creating 'hundreds of thousands' of jobs, including in…
