NVIDIA has agreed to pay roughly $20 billion to amass belongings from synthetic intelligence chip startup Groq, marking the corporate’s largest transaction on document and persevering with its technique of absorbing potential opponents earlier than they’ll problem its market dominance.
The chipmaker’s newest licensing deal mirrors an identical transaction simply three months in the past, reinforcing the narrative that decentralized AI infrastructure might provide the one various to Nvidia’s rising dominance.
Sponsored
Sponsored
Threefold Premium in Three Months with Trump Jr. Connection
The deal closed simply three months after Groq raised $750 million at a $6.9 billion valuation—a spherical that included BlackRock, Samsung, Cisco, and 1789 Capital, the place Donald Trump Jr. serves as a companion. Nvidia is buying the entire firm’s belongings considerably, besides its cloud computing enterprise, although Groq framed the transaction as a “non-exclusive licensing agreement.”
Groq CEO Jonathan Ross, a former Google engineer who helped create the search big’s Tensor Processing Unit, will be a part of Nvidia together with president Sunny Madra and different senior executives. The startup will proceed working independently below CFO Simon Edwards as its new chief govt.
A Repeating Playbook
The Groq transaction follows a sample Nvidia established simply three months earlier. In September, the corporate paid over $900 million to rent Enfabrica’s CEO and staff whereas licensing the startup’s know-how. Each offers use licensing constructions fairly than outright acquisitions, doubtlessly avoiding the antitrust scrutiny that blocked Nvidia’s $40 billion bid for Arm Holdings in 2022.
The Kobeissi Letter summarized Nvidia’s strategy bluntly: “We will buy you before you can compete with us.”
Nvidia’s latest technique:
“We will buy you before you can compete with us.”
There has by no means been an organization like Nvidia. https://t.co/wsbuAgIqyM
— The Kobeissi Letter (@KobeissiLetter) December 24, 2025
Sponsored
Sponsored
Technical Edge and Aggressive Strain
Groq’s Language Processing Unit makes use of on-chip SRAM fairly than exterior DRAM, enabling what the corporate claims is as much as 10x higher power effectivity. This structure excels at real-time inference however limits mannequin dimension—a tradeoff Nvidia can now discover inside its broader ecosystem.
The timing is notable. Google not too long ago unveiled its seventh-generation TPU, codenamed Ironwood, and launched Gemini 3, skilled solely on TPUs, to prime benchmark rankings. Nvidia responded on X: “We’re delighted by Google’s success… NVIDIA is a generation ahead of the industry—it’s the only platform that runs every AI model.” When incumbents begin issuing such reassurance statements, aggressive stress is clearly mounting.
Implications for Decentralized AI
Whereas the deal has no direct impression on cryptocurrency markets, it reinforces the narrative driving decentralized AI computing tasks. Platforms like io.web place themselves as alternate options to centralized AI infrastructure.
“People can put their own supply onto a network, whether that’s data centers or yourself with your laptop, contributing your available GPU power, and getting fairly compensated for it using tokenomics,” Jack Collier, io.web’s Chief Development Officer, instructed BeInCrypto. The platform claims enterprise shoppers, together with Leonardo.ai and UC Berkeley, have achieved vital value financial savings.
Nonetheless, the hole between narrative and actuality stays broad. Nvidia’s acquisition of Groq’s low-latency know-how additional extends its technical lead, making it more durable for any various to supply aggressive efficiency.
The transaction additionally raises questions on unbiased AI chip growth. Cerebras Methods, one other Nvidia competitor getting ready an IPO, might finally face comparable stress. Whether or not it could stay unbiased or succumb to Nvidia’s monetary gravity stays to be seen.
