Story Content

According to IO Fund lead technology analyst Beth Kindig, Nvidia acquiring AI chip company Groq in a deal worth 20 billion US dollars was a significant step towards propelling its next wave of revenue generation. Kindig highlighted in a recent analysis that integrating Groq's advanced inference technology, and especially the Groq 3 LPX racks, will result in a token boom, reducing inference expenses and increasing output per second by up to 15 times together with Nvidia Vera Rubin GPUs.

This architecture is guaranteed to improve up to 35x more throughput per megawatt, which can fix issues fundamentally with memory bandwidth, and it can do AI processing at lower cost and faster. Kindig compared the acquisition with the transformative culture of Nvidia in 2020, acquiring Mellanox, by saying that reduced prices will enable an expanded use of tokens, which is a direct correlation to increased revenue and margin in the current wave of AI monetization.

The deal (organized as a technology licensing deal with Groq talent coming to Nvidia) was announced in late 2025 and will make the company control inference markets. Kindig considers that this fact triggered the ambidextrous projections of Nvidia, such as a possible $1 trillion revenue potential by 2027, and continues to lead the AI market in unpredictable times.