Nvidia has agreed to pay approximately $20 billion to acquire the core assets and technology of AI chip startup Groq, marking what would be the largest deal in the company’s history and significantly reshaping the competitive landscape of the artificial intelligence hardware market. The transaction, reported by financial outlets including CNBC, reflects Nvidia’s strategic push to strengthen its position in AI inference computing — the process of running trained models efficiently in real-time applications.
Groq, a U.S.-based company founded in 2016 that specialises in AI accelerator chips with a focus on low-latency, high-performance inference technology, saw its latest funding round value the firm at around $6.9 billion earlier in 2025. Nvidia’s acquisition price represents a substantial premium over that valuation, underlining the importance the chip giant places on extending its AI infrastructure beyond traditional graphics processing units (GPUs).
Under the agreement, most of Groq’s assets will be folded into Nvidia’s operations, while the startup’s early-stage cloud business — GroqCloud — is expected to remain independent. Key executives from Groq, including founder and CEO Jonathan Ross and other senior engineers, will join Nvidia to help integrate and scale the acquired technology within Nvidia’s broader AI platform.
The deal highlights Nvidia’s broader strategy of consolidating advanced AI hardware technologies and talent to maintain its industry lead. Analysts note that securing Groq’s inference capabilities will allow Nvidia to offer more competitive, energy-efficient solutions for AI applications that require rapid, real-time responses, such as conversational AI services and large-scale enterprise deployments.