Nvidia (NVDA.US) spent $20 billion to “gather” the strongest rivals, Wall Street: Target price rose to $300

The Zhitong Finance App learned that Wall Street analysts are generally optimistic about the latest deal between Nvidia (NVDA.US) and AI inference chip company Groq. Among them, the Cantor agency believes that the deal has both “offensive” and “defensive” strategic significance, reaffirming Nvidia as the “preferred stock”, maintaining the “gain” rating and providing a target price of $300; while Bank of America pointed out that although Nvidia's acquisition of Groq at a high price was unexpected, it successfully turned a potential ASIC technology threat into its own competitive barrier. From a long-term perspective, this layout is worth remarkable, so it maintains a “buy” rating and a target price of $275.

On Wednesday local time, Groq announced that it has signed a non-exclusive license agreement with Nvidia, authorizing the latter to use its inference technology. Under the agreement, Groq founder Jonathan Ross, President Sonny Madera, and other team members will join Nvidia to drive and expand the implementation of this licensed technology. According to reports, Nvidia will invest around $20 billion to acquire Groq-related assets.

Cantor reaffirmed Nvidia as a “preferred stock” and maintained an “overweight” rating and a target price of $300.

“On Christmas Eve, Nvidia announced the acquisition of Groq's IP and talent for $20 billion (which can be considered a 'talent merger and acquisition'). Overall, we believe this acquisition is both offensive and defensive. On the offensive side, we learned that Nvidia has been partnering with Groq for specific inference acceleration. We speculate that Nvidia saw a real opportunity and thought it would be more beneficial to have Groq as an internal team rather than an external partner,” said the Cantor analyst team led by C.J. Muse.

The bank's analysts also highlighted strong talent intake (including the CEO of Groq, who was one of the core developers of Google (GOOGL.US) TPU).

On the defensive side, analysts pointed out that Nvidia dominates the field of AI training and “time-based reasoning,” and that Groq's low-latency, energy-efficient inference technology is incorporated into Nvidia's complete system stack (which is expected to be CUDA compatible in the future), which will help it expand its share in the inference market, especially in the next stage of AI infrastructure construction — real-time workloads such as robots and autonomous driving.

“Taken together, we believe this acquisition will only strengthen Nvidia's full-stack system capabilities and overall leadership position in the AI market, and further broaden its competitive moat,” the Muse team concluded.

Bank of America Securities maintains Nvidia's “buy” rating and a target price of $275.

Bank of America Securities analysts led by Vivek Alia said the Groq deal was surprising and expensive, but it had strategic value.

They listed two main points:

1. Breaking news on Christmas Eve involving hardware different from GPUs — language processing units (LPUs) — that Nvidia is good at shows that although GPUs dominate in the field of AI training, demand for inference is rapidly exploding, and more specialized chips may be needed.

2. Different hardware will increase the complexity of future GPU/LPU roadmaps and pricing, but Nvidia can use its balance sheet and platform position to provide customers with more choices and conceptually mitigate competitive threats from Groq and other dedicated ASIC chips.

The bank's analysts pointed out that there are still problems to be solved, but in the long run, this potential deal is a positive factor.

“In the long run, if this potential Groq deal finally lands, its strategic significance could be comparable to Nvidia's April 2020 acquisition of Mellanox — the latter is now the foundation for its network/AI expansion moat,” Arya's team said.

According to information, GroQ is a star startup that specializes in AI inference chips. It was founded in 2016 and is headquartered in California, USA. Founder Jonathan Ross was a core R&D member of Google's self-developed AI chip TPU (Tensor Processing Unit) project, and some former Google TPU team members also joined GroQ along with him.

As the core developer of Google's first-generation tensor processing unit (TPU) project, Jonathan Ross was deeply involved in designing chips optimized for AI. In 2016, he led 7 of the 10 core members of Google's TPU team to leave their jobs and founded Groq. At the time, he discovered that traditional computing architectures (such as CPU/GPU) could not efficiently handle modern AI tasks, and this perception prompted him to start a company that breaks through traditional limitations.

In February 2024, GroQ launched a new AI chip which claims to have achieved “the strongest inference on the surface” — running large models on GroQ is 10 times or more faster than Nvidia GPUs. In November 2025, the latest statements from the US White House and the US Department of Energy showed that 24 top artificial intelligence companies have signed agreements with the US government to join the “Genesis Project,” and Nvidia and Groq are among them.

GroQ's core product is an LPU (language processing unit). This type of chip is mainly used to speed up the speed of big language models to complete inference related tasks, and is regarded by the outside world as an alternative to Nvidia's GPU. Currently, Groq has partnered with Meta to provide inference acceleration for its Llama API; collaborated with IBM to integrate its AI inference platform; and signed a huge agreement with Saudi Aramco to build a large-scale AI inference data center.