Nvidia has acquired assets from AI chip startup Groq for about $20 billion, continuing its trend of absorbing potential competitors. The deal closely follows a similar transaction three months earlier, emphasizing the trend towards decentralized AI infrastructure as an alternative to Nvidia’s market dominance.

Groq recently raised $750 million at a $6.9 billion valuation, attracting investment from BlackRock, Samsung, Cisco, and 1789 Capital. Nvidia is acquiring all of Groq’s assets except its cloud computing business in a non-exclusive licensing agreement.

Groq’s CEO Jonathan Ross, along with other senior executives, will join Nvidia as part of the deal. The startup will still operate independently under new chief executive Simon Edwards, while Ross contributes his expertise from Google, where he helped develop the Tensor Processing Unit.

Nvidia’s recent acquisition strategy mirrors a previous $900 million deal with Enfabrica, using licensing structures to potentially avoid antitrust scrutiny. Groq’s Language Processing Unit offers up to 10x better energy efficiency with on-chip SRAM, which could provide Nvidia with valuable technology for its ecosystem.

The acquisition of Groq’s technology by Nvidia comes at a time when Google has unveiled its seventh-generation TPU and released Gemini 3, creating competitive pressure in the AI computing market. While the deal doesn’t directly impact cryptocurrency markets, it highlights the importance of decentralized AI projects like io.net.

Nvidia’s acquisition of Groq’s low-latency technology further solidifies its technical lead in the AI chip market, potentially making it harder for competitors like Cerebras Systems to remain independent. The transaction raises questions about the future of independent AI chip development in the face of Nvidia’s dominance.

Read more at Yahoo Finance: Nvidia Absorbs Another Rival for $20B, Boosting Decentralized AI