Nvidia Unveils New Chip to Defend Its AI Dominance

The tech giant is partnering with Groq to boost its inference capabilities and fend off competition in the fast-growing AI market.

Mar. 16, 2026 at 7:34pm

Got story updates? Submit your updates here. ›

Nvidia CEO Jensen Huang announced a new product that combines Nvidia's chips with technology from startup Groq. The goal is to improve the speed and cost-efficiency of AI inference, an area where Nvidia has faced growing competition from rivals like Google and Cerebras. Nvidia's dominance in the AI chip market has been threatened as companies shift towards inference-focused workloads, and the Groq partnership is Nvidia's attempt to regain its edge.

Why it matters

Nvidia's chips power the majority of AI development and deployment today, but the company's lead is being challenged as the focus shifts to inference workloads that require different chip architectures. Nvidia's ability to maintain its market share will be crucial to its continued dominance in the booming AI industry.

The details

Nvidia unveiled a new product that combines its chips with technology from startup Groq, which specializes in chips optimized for AI inference. This partnership aims to make Nvidia's inference capabilities more cost-effective and faster, addressing a weakness compared to rivals like Google's tensor processing units and Cerebras' specialized chips. Over the past year, AI companies have shifted more of their workloads towards inference, which puts a premium on chips that can generate data quickly and inexpensively. Nvidia's traditional chips have lagged behind competitors in this area, leading some major customers like OpenAI and Meta to turn to other providers.

  • Nvidia announced a $20 billion licensing agreement with Groq in December 2025.
  • Nvidia unveiled the new Nvidia-Groq product at its annual GTC conference in San Jose on March 16, 2026.

The players

Nvidia

The most valuable publicly traded company in the world, known for its dominance in the AI chip market.

Jensen Huang

The CEO of Nvidia who has positioned the company's chips as the 'Swiss Army knife of artificial intelligence'.

Groq

A startup that makes chips custom-built for AI inference, which Nvidia has partnered with to boost its own inference capabilities.

Google

A major competitor to Nvidia in the AI chip market, with its own tensor processing units that excel at inference workloads.

Cerebras

An upstart chip company that makes specialized chips for running AI, posing a challenge to Nvidia's dominance.

Got photos? Submit your photos here. ›

What they’re saying

“A.I. is able to do productive work, and therefore the inflection point of inference has arrived.”

— Jensen Huang, CEO, Nvidia

“They're going to thread this together to protect their moat.”

— Daniel Newman, CEO, Futurum Group

“It's a brilliant supply-chain move.”

— Umesh Padval, Managing Partner, Seligman Ventures

What’s next

Nvidia will need to win over more customers with its new Nvidia-Groq inference-focused product in order to maintain its dominant market share as the AI industry continues to evolve.

The takeaway

Nvidia's partnership with Groq is a strategic move to defend its leadership position in the AI chip market as the industry shifts towards more inference-heavy workloads. The company's ability to adapt and innovate will be crucial to staying ahead of competitors and retaining its position as the driving force behind the AI revolution.