Tesla Considers Building ‘Tera Fab’ to Meet Future Chip Needs

//php echo do_shortcode(‘[responsivevoice_button voice=”US English Male” buttontext=”Listen to Post”]’) ?>

An upbeat Elon Musk, with his newly approved trillion-dollar pay package, talked to a very select group of investors at the Tesla annual shareholders meeting yesterday in Austin, Texas, and told the audience he had chips on his brain and planned to build a “Tera fab” that could potentially produce a million wafer starts per month to meet chip demand for Tesla alone.

Musk said that all he could think about was chips at the moment. He also said that with his AI5 chip and the planned AI6 chip, he did not think existing suppliers (TSMC and Samsung) could meet demand. As a result, he would consider some collaboration with Intel and also building Tesla’s own “Tera fab” to meet demand for Tesla’s own products.

“The thing is that we actually have agreed to buy all the chips that are made from the fab. So, it’s basically a money printing machine for TSMC and Samsung. It’s like literally the faster you make the chips, the faster we send them money. But it’s still not going just fast enough,” Musk said during the shareholder meeting.

“So that’s why I think, as far as I can see, the only option is to go build some like very big chip fab. And then you go to solve memory and packaging too. But otherwise, you just tap out at whatever the chip production rate is. And so, I guess, Tera would be—you’d want to say it’s got to be at least 100,000 wafer starts per month size fab. And maybe that would be one of ten in a complex. So, it would ultimately be one million wafer starts per month,” he added.

To this, he got a giggle from the audience, to which he responded, “Yeah. Exactly. You can tell when it gets the giggle factor. That’s probably a good sign that we’re onto something special. But I wouldn’t be surprised if long term, it’s like a million wafers a month.”

Musk expects huge demand coming from the Tesla AI5 chip for the company’s AI software stack. Currently, this would be made in four places: TSMC Taiwan, Samsung Korea, TSMC Arizona and TSMC Texas.

Possible deal with Intel?

However, the challenge, he said, is how to make enough chips. “I have a lot of respect for the Tesla partners, TSMC and Samsung. Maybe we’ll do something with Intel. We haven’t signed any deal, but it’s probably worth having discussions with Intel. But even when we extrapolate the best-case scenario for chip production from our suppliers, it’s still not enough,” Musk said.

This is the reason he believes that Tesla may need to build a Tera fab. While he did not give a timeline, he alluded to plans for a fab in the future.

“I can’t see any other way to get to the volume of chips that we’re looking for. So, I think we’re probably going to have to build a gigantic chip fab. Got to be done.”

Optimus driving chip demand

The key driver for demand would be from areas like robotics. Early in the shareholder meeting, Musk danced along with two Optimus humanoid robots at the sides of the stage and enthused about the transformation they would make to humanity, in terms of personalized healthcare, tackling crime and generally growing the world economy by a factor of ten.

Elon Musk believes the opportunity from the company’s Optimus robots will be huge, and said it’s going to be the biggest product of all time by far, bigger than cell phones, “bigger than anything.” (Source: Tesla)

Talking about the opportunity from Optimus, he said, “I think it’s going to be the biggest product of all time by far. So, like, bigger than cell phones, bigger than anything. I guess the way to think about it is that every human on earth is going to want to have their own personal R2-D2 C-3PO. So, who wouldn’t?”

On their usefulness in healthcare, Musk said, “With humanoid robots, you can actually give everyone amazing medical care. In terms of Optimus, [it] will be more precise. Optimus will ultimately be better than the best human surgeon with a level of precision that is beyond human. So, I think that’s a pretty wild concept to say, okay, you…there’s always people always talked about eliminating poverty, but actually, Optimus will actually eliminate poverty. Optimus will actually give people incredible medical care.”

And how is it useful in crime? Well, Musk’s vision is that if somebody has committed a crime, a more humane form of containment of future crime would be to give the offender a free Optimus robot, which would be programmed to just follow you around everywhere and stop you from doing crime.

“It’s just going to stop you from committing crime. That’s really it. You don’t have to put people in like prisons and stuff, I think.”

Musk added that everyone would want an Optimus robot, and that there could be tens of billions of Optimus robots on earth. He said it would launch the fastest production ramp of any product ever, with a one-million-unit production line in Fremont and a 10-million-production line in Austin.

AI5 chip: a third of the power of Blackwell and 10% of the cost

With chips on his brain, Musk talked quite a bit about the Tesla chip design expertise and how the AI5 chip was specially designed for Tesla’s needs.  He said he would put a lot of time into the Tesla chip design in enabling a functional robot.

Tesla AI5 chip improvements over it’s AI4 chip. (Source: Tesla)

 “In order to have a functional robot, you have to have a great AI chip. And it needs to be an inexpensive chip, and it needs to be very power efficient,” he said. “So, we believe the AI5 chip will be probably about a third of the power of, say, something like a Nvidia Blackwell, which is a great chip, for roughly comparable performance and much less than ten percent of the cost. So, this is a chip that is very much optimized for the Tesla AI software stack. It’s not meant to be a general-purpose chip. It’s meant to be an amazing chip for the Tesla AI software.”

Musk added that he thought Tesla is unique in using an integer-based system.

“Integer operations are fundamentally more efficient than floating point operations. So, we can do floating point, but the vast majority of our inference is done in integer, which is, if you’re familiar with sort of logic gates, the simplicity of integer, integer is much more power efficient, much more silicon efficient,” he said at the shareholder meeting. “But you actually have to train for integer inference, which everyone else is training for floating point. That’s kind of a niche technical detail, but it’s actually very important.”

A convincing plan?

While some members of the audience may have giggled with a certain sense of disbelief on Musk’s plans to build a Tera fab or fabs with a capacity to do a million wafter starts per month, it is not beyond imagination to envision a scenario that AI demand is so huge and driven by applications like his robots that he will need more capacity than TSMC and Samsung can offer now. A collaboration with Intel on its latest process technology could be a solution, but if Musk really believes that there is going to be demand for billions of robots, one can see why he is putting it out there now that Tesla will need to build its own fabs.

We will be watching to see when there is a timeline put on this—there may already be one in Musk’s head, we will just have to wait to see when he decides the time is right to reveal more details of funding, locations and targets.

The full video recording of the Tesla 2025 shareholders meeting can be found here.

See also:

Samsung’s Multi-Billion Tesla Deal

Infineon OktoberTech 2025: Humanoid Robotics, Edge AI and Power

Tesla’s wireless-power “dream” gets closer to reality—maybe

Will Robots Use Reasoning At The Edge?

Inside Robotics Power: Safety, Batteries, and Market Growth

India’s Addverb Brings AI and Edge Computing to Warehouse Robotics