Disruptive CEO Alex Davis says Nvidia has agreed to buy assets from Groq for $20B in cash; leaked email: Jensen Huang says Nvidia is not acquiring Groq
Nvidia has agreed to buy assets from Groq, a designer of high-performance artificial intelligence accelerator chips, for $20 billion in cash …
CNBCDavid Faber
Context & Ripple Effects
The initial account of a $20 billion asset transaction was quickly complicated by Nvidia’s denial and by reporting that the parties had instead reached a non-exclusive licensing arrangement with executive departures to Nvidia. That distinction matters because it separates control of a company from access to its inference technology and leadership.
Later coverage said that roughly 90% of Groq employees were expected to join Nvidia, while Groq said it would remain independent and keep GroqCloud operating. The arc suggests a transaction designed to transfer key capabilities without a conventional corporate acquisition.
First-order effects
Nvidia obtains non-exclusive rights to Groq’s inference technology and adds Groq’s top leadership, expanding its access to low-latency inference expertise without formally taking over Groq.
Groq retains its corporate independence and cloud operation, but the reported movement of executives and employees materially changes where its technical capacity sits.
Second-order effects
The licensing structure lets Nvidia pursue Groq-derived capabilities while leaving Groq able to operate and potentially license or serve customers independently, creating a more complex competitive boundary than a full acquisition would.
For investors and employees, the reported cash payouts and hiring shift value realization toward transaction terms and talent retention rather than an all-shareholder sale of the company.
Third-order effects
If replicated, AI-chip incumbents may increasingly secure specialized IP and teams through licensing-plus-hiring arrangements, blurring the line between partnership, acqui-hire, and acquisition.
That model could make independent accelerator companies less defined by ownership outcomes and more by whether they can retain talent, product roadmaps, and customer access after strategic deals.
The trend:AI infrastructure leaders are using flexible combinations of IP licensing and talent transfers to absorb specialized inference capabilities without always pursuing full-company acquisitions.
Nvidia acquiring Groq for $20B is a big deal. Groq chips are insanely fast at inference, sometimes 10x GPUs. The trick is to put model weights in SRAM instead of HBM to trade memory capacity for speed. Its $/token may lose to GPUs, but for long-wait inference on models like GP…
Wow! NVIDIA just acquired Groq an AI chip maker that enables fast inference It's a smart defensive move - they basically removed a potential competitor from the market They also have solidified their near monopolistic position in the market
This is wild Groq had such a fundamentally different chip than Nvidia's accelerators I wonder why they'd acquire them. Probably diversification? Maybe CPX?
Huge news. Congrats Founder @JonathanRoss321 @sundeep & seed investor @chamath on a huge outcome that will change the game yet again for AI systems in 2026. 🚀🚀 https://www.cnbc.com/...
A $20B datapoint proving that you don't need “win” and others to “lose” to generate an absolutely massive outcome. As much as we'd like to oversimplify, most of venture is NOT zero-sum, it's about building great teams solving real problems. @JonathanRoss321 and the @GroqInc
www.cnbc.com/2025/12/24/n... Well, not sure how easy it is to merge both technologies, but it surely takes out a competitor on the increasingly vital serving side
What a run! 2 years ago, @groq.com was in the news for the ALCF (Argonne Leadership Computing Facility) AI Testbed at Argonne National Laboratory. Time flies. [embedded post]
Nvidia buying AI chip startup Groq's assets for about $20 billion in largest deal on record, according to Alex Davis, CEO of Disruptive, which led the startup's latest financing round in September.