Elon Musk Says Nvidia's Next-Gen Rubin Chips Won't Be Operational At Scale Soon As Tesla Works On In-House AI Hardware

[Daniel Nenni]

Tesla Inc. CEO Elon Musk says that Nvidia Corp's new chipset would take some time before it can be scaled.

Nvidia's Rubin Chip​

On Monday, influencer Sawyer Merritt shared a video that detailed Nvidia's new Vera Rubin chips' capabilities, showcased at the Consumer Electronics Show (CES) 2026.

The new architecture has no fans or cables and promises up to five times more powerful performance than the preceding Blackwell chip and each rack would contain 72 Rubin chips. The chips, according to Nvidia CEO Jensen Huang, were two GPUs connected together.

NEWS: Nvidia has released a new video of its next-generation Rubin chips, which are 5X more powerful than their predecessor, Blackwell. The new Rubin chips are already in production.

CEO Jensen Huang: "This is a Rubin pod. 1,152 GPUs and 16 racks. Each rack has 72 Rubins." pic.twitter.com/5NDkKmj64y

— Sawyer Merritt (@SawyerMerritt) January 5, 2026

Nvidia Would Need 9 Months, Says Elon Musk​

The SpaceX CEO shared his thoughts in a response to Merritt's post. He shared that despite the impressive technology, it would take "another 9 months" before the technology becomes "operational at scale" and the software can function smoothly.

It will take another 9 months or so before the hardware is operational at scale and the software works well

— Elon Musk (@elonmusk) January 6, 2026

Elon Musk On Nvidia's Self-Driving Exploits​

Nvidia also showcased its new self-driving technology, Alpamayo, at the event, which touted a vision-language-action approach and was hailed by Huang as the "ChatGPT moment" for self-driving.

Musk shared that the technology's distribution would be "super hard" to solve for Nvidia. His comments were echoed by Tesla's AI Chief, Ashok Elluswamy, too. However, the Tesla CEO ultimately wished Nvidia the best and hoped that it succeeded in developing self-driving car tech.

Tesla’s In-House Chips​

Meanwhile, Tesla is also working on developing AI chips. Musk had said that the developments would enable Tesla to produce more chipsets than any of its competitors. He had also announced a major hiring drive at the company to help in its chip-building efforts.

https://www.yahoo.com/finance/news/elon-musk-says-nvidias-next-160016588.html

 

[Daniel Nenni]

This is the type of competition we need to keep semiconductor technology moving at a break neck speed. Bless billionaires and their massive egos!

 

[Xebec]

FWIW, What Nvidia is offering isn't strictly a direct competitor to Tesla's FSD. Per Jensen, they're offering a HW/SW stack that customers can use all of or part of, but it's still up to the car maker using Nvidia still have to fully finish developing the software to make autonomous driving work.

This was Jensen's response at a Q&A as to the difference between Tesla FSD and (Nvidia) Alpamayo:

"Alpamayo was designed around a different idea. The first difference is that NVIDIA doesn’t build self-driving cars—we build the full stack and the technology for everybody else to build self-driving cars. And we build—like we do for humanoid robotics—three computers: the training computer, the simulation computer, and the robotics computer, which is the self-driving car computer. We have software stacks across all of that.

Our customers can use all of it, some of it, or parts of it—whatever makes sense for them. And so we’re working with the entire industry—Tesla for their training system, Waymo for the car computer, and XPeng. Nuro—who I think just announced they’re going into the robotaxi business—with Lucid and Uber; and NVIDIA is part of that.

So our system is really quite pervasive because we’re a technology platform provider—that’s the primary difference. There’s no question in our mind that, of the billion cars on the road today, in another 10 years’ time, hundreds of millions of them will have great autonomous capability. This is likely one of the largest, fastest-growing technology industries over the next decade.

And the last thing we do is: we open-source everything. If a customer would like to use the model that we train, they’re welcome to do that. If they would like to use our model technology but train it themselves, we even help them do that. We’re not a self-driving car company—we just want to enable the world’s autonomous industry. Everything that moves should be autonomous.”


Source (warning: Ads) - https://www.teslarati.com/nvidia-ce...ns-difference-between-tesla-fsd-and-alpamayo/

 

[FIvr]

This is the type of competition we need to keep semiconductor technology moving at a break neck speed. Bless billionaires and their massive egos!

Good to see you are taking a break from your bubble pep talk.