• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Elon Musk announces Terafab project he claims will be the 'largest chip manufacturing facility ever'

P

polonyc2

Fully [H]
2FA
Joined
Oct 25, 2004
Messages
29,753
Musk says the chip output is expected to reach a whopping 1 terawatt of compute power per year, including logic chips, memory, and advanced packaging...TeraFab is reported to focus on 2nm production, meaning the fab will jumpstart with cutting-edge production from the very beginning

TeraFab's first major product would be the AI5 chips used in Tesla's FSD, Robotaxi, and Optimus robots...the AI5 will account for the 20% ground-based production split, and the remaining 80% of space-based compute will come from the newly unveiled D3 chips

The facility will be built in Austin, Texas, where Tesla is headquartered, and will be a joint effort among all of Musk's companies...as for how fab production would be managed, given the gigantic scale, it is disclosed that 80% of the compute would go to space, since he believes Earth's power constraints make deployment almost impossible...using Starship delivery channels and the power of the Sun, Musk believes he could deploy "terawatts" of computing power in space, which is why he says TeraFab is vital to leading us to a galactic civilization...

https://www.engadget.com/science/el...ip-manufacturing-facility-ever-171718545.html

https://x.com/Tesla/status/2035526423495180779
 
Last edited:
I'm very confused how this is feasible. The chip designing part sure, others have done that, but all the fabbing? Going from nothing to 2nm just sounds like skipping everything that has been a challenge for others in the space for decades (and fiercely guarded by TSMC).
 
Okatis said:
(and fiercely guarded by TSMC).
Click to expand...

TSMC certainly deserves credit where due, but they don't have as many "guarded secrets" as some seem to think. They are still totally reliant on Dutch (ASML) lithography machines for chip production. Musk could simply buy those same machines from ASML if he wanted to, and has more than enough money to build as many fabs as he wants (unlike Intel).
 
GotNoRice said:
TSMC certainly deserves credit where due, but they don't have as many "guarded secrets" as some seem to think. They are still totally reliant on Dutch (ASML) lithography machines for chip production. Musk could simply buy those same machines from ASML if he wanted to, and has more than enough money to build as many fabs as he wants (unlike Intel).
Click to expand...
No you cant simply buy these machines. There is quite a wait. Money is not the only barrier to a modern fab.

And intel isnt broke, they can buy everything tsmc can. Yet they still cant seem to execute nearly as well on their us and previohs fabs as tsmc.

Im not sure how this project works on musks timeframe. Getting a bleedinv edge fab from scratch takes like a decade and a huge sume of money.

And also we have plenty of power on earth, space data centers are not the future. Although we should get around to building more nuke plants at some point.
 
GotNoRice said:
TSMC certainly deserves credit where due, but they don't have as many "guarded secrets" as some seem to think. They are still totally reliant on Dutch (ASML) lithography machines for chip production. Musk could simply buy those same machines from ASML if he wanted to, and has more than enough money to build as many fabs as he wants (unlike Intel).
Click to expand...

Its an incredibly sophisticated it logically simple supply chain. To make the exact same chips as they make in Taiwan in Arizona they needed to adjust all their chemicals because the air and water are different, the seasons are different. Which means the recipes for each design have to be re-worked from scratch to get any yield at all. And then in three years its obsolete. Like the entire facility. You have to hit high yields 24/7 for years to break even. They current tech is pushed to its absolute limit, printing on the same layer four times with different masks and using the interference to get the wire routing right. and then 30 layers of that that all have to work together with no part of your chemical or physical processes messing up. I mean if Elon has a better than EUV tech sitting in his pocket that will be at the beginning of its life and easier to work with, sure, he has a chance of getting ahead. But he needs a piece of tech magic, hes not up against Lockheed or Beoing where best practices were enough to jump him a decade ahead.
 
cdabc123 said:
And also we have plenty of power on earth, space data centers are not the future.
Click to expand...

There are endless stories about data centers opening in areas that had cheap power, and now the power costs in those areas have gone way up. So I guess "plenty" depends on how much you want to pay for power.

The idea of data centers in space sounds dumb at first, but there is actually a lot of logic there. One thing that I think most people aren't really aware of is how much more potent solar is as an energy source in space. Like 13-40 times as much power compared to ground-based solar installations, with the HUGE added benefit of 24/7 sunlight with the right orbit. He can do most of this in-house, and with SpaceX having brought the cost to orbit down so much, it starts to make things possible that weren't practical before.
 
polonyc2 said:
using Starship delivery channels and the power of the Sun, Musk believes he could deploy "terawatts" of computing power in space, which is why he says TeraFab is vital to leading us to a galactic civilization...
Click to expand...

Doc Ock.gif
 
GotNoRice said:
There are endless stories about data centers opening in areas that had cheap power, and now the power costs in those areas have gone way up. So I guess "plenty" depends on how much you want to pay for power.

The idea of data centers in space sounds dumb at first, but there is actually a lot of logic there. One thing that I think most people aren't really aware of is how much more potent solar is as an energy source in space. Like 13-40 times as much power compared to ground-based solar installations, with the HUGE added benefit of 24/7 sunlight with the right orbit. He can do most of this in-house, and with SpaceX having brought the cost to orbit down so much, it starts to make things possible that weren't practical before.
Click to expand...
Yes practical is all dependant on price. Heck plenty of new data centers now just put like 2 dozen portable gas turbines to operate. But thats not a sustainable or environmental solution.

We NEED nuke power if we envision a future with lots of data centers and I have to imagine the demand for computation will not decrease. Unfortunately there hasn't been a good effort for nuclear. And it take like a decade to materialize. Ie were screwed electric prices will go up.

I do not yet conscribe to the feasibility of datacenters in space. There is some solar power but every watt you take in is a headache to shed via ir in space. There is no way to get around that so there will be no space datacenter near the watt output of stuff on earth. We could however do light computations in space. Like say all starlink sats had a reasonable amount of efficiant computational power. A small distributed data center. Even the latency along in low orbit is much more then data centers on earth.
 
Last edited:
GotNoRice said:
TSMC certainly deserves credit where due, but they don't have as many "guarded secrets" as some seem to think. They are still totally reliant on Dutch (ASML) lithography machines for chip production. Musk could simply buy those same machines from ASML if he wanted to, and has more than enough money to build as many fabs as he wants (unlike Intel).
Click to expand...

You need the ASML machines, but you also need a lot of other stuff. Fabbing requires precision, sustained focus, and consitency. Doesn't sound like a job for Musk. TSMC took 5 years to get production running in AZ from when they announced and they know what they're doing.
 
I would put this into the "not gonna happen" column in terms of him doing this on his own. The more likely thing he'll do is buy someone's working fab and then put his branding on it.
 
toast0 said:
You need the ASML machines, but you also need a lot of other stuff. Fabbing requires precision, sustained focus, and consitency. Doesn't sound like a job for Musk. TSMC took 5 years to get production running in AZ from when they announced and they know what they're doing.
Click to expand...
You also forgot the trained workforce with years of experience in running and maintaining multiple generations of state-of-the-art lithography machines. The talent pool is neither broad nor deep for management and senior staffers who can train the newer people that will fill out the much of skilled workforce.
 
Austins not a bad place if you are looking for work force. The Texas Triangle has a lot of fab manufacturing experience. Though moving between the big cities is proving to be hard for most people currently due to cost of housing in Texas exploding (mostly because of all these companies moving here like this).

Musk says the chip output is expected to reach a whopping 1 terawatt of compute power per year
Click to expand...

This is a strange way of measuring fab output. Wafers Per Day would be a far far more useful metric.
 

Intel joins Elon Musk's TeraFab project — 'Intel is proud to join the Terafab project with SpaceX, xAI, and Tesla to help refactor silicon fab technology'​

News
By Anton Shilov published 10 minutes ago
Intel's abilities to design, fabricate, and package chips to help TeraFab to achieve its goals.

Intel on Tuesday said that it had joined Elon Musk's TeraFab project. The announcement mentions Intel's ability to develop, produce, and package advanced processors in high volumes, which could help Tesla, SpaceX, and xAI to get enough compute performance for next-generation AI and robotics applications.

"Elon has a proven track record of reimagining entire industries," Lip-Bu Tan wrote in his own X post. "This is exactly what is needed in semiconductor manufacturing today. Terafab represents a step change in how silicon logic, memory and packaging will get built in the future. Intel is proud to be a partner and work closely with Elon on this highly strategic project."

https://www.tomshardware.com/tech-i...tesla-to-help-refactor-silicon-fab-technology
 
Marees said:

Intel joins Elon Musk's TeraFab project — 'Intel is proud to join the Terafab project with SpaceX, xAI, and Tesla to help refactor silicon fab technology'​

News
By Anton Shilov published 10 minutes ago
Intel's abilities to design, fabricate, and package chips to help TeraFab to achieve its goals.

Intel on Tuesday said that it had joined Elon Musk's TeraFab project. The announcement mentions Intel's ability to develop, produce, and package advanced processors in high volumes, which could help Tesla, SpaceX, and xAI to get enough compute performance for next-generation AI and robotics applications.

"Elon has a proven track record of reimagining entire industries," Lip-Bu Tan wrote in his own X post. "This is exactly what is needed in semiconductor manufacturing today. Terafab represents a step change in how silicon logic, memory and packaging will get built in the future. Intel is proud to be a partner and work closely with Elon on this highly strategic project."

https://www.tomshardware.com/tech-i...tesla-to-help-refactor-silicon-fab-technology
Click to expand...
wow, beat me by 4 minutes..
 
Marees said:
"Elon has a proven track record of reimagining entire industries," Lip-Bu Tan wrote in his own X post. "This is exactly what is needed in semiconductor manufacturing today. Terafab represents a step change in how silicon logic, memory and packaging will get built in the future. Intel is proud to be a partner and work closely with Elon on this highly strategic project."

https://www.tomshardware.com/tech-i...tesla-to-help-refactor-silicon-fab-technology
Click to expand...
I'm inclined to agree with him. Elon tends to get shit done. If he can build rockets, I'm sure he can figure out EUV. I'm not saying he'll be competitive with TSMC in the next ten years, but he could probably get there.
 
Marees said:
Intel joins Elon Musk's TeraFab project — 'Intel is proud to join the Terafab project with SpaceX, xAI, and Tesla to help refactor silicon fab technology'
News
By Anton Shilov published 10 minutes ago
Intel's abilities to design, fabricate, and package chips to help TeraFab to achieve its goals.

Intel on Tuesday said that it had joined Elon Musk's TeraFab project. The announcement mentions Intel's ability to develop, produce, and package advanced processors in high volumes, which could help Tesla, SpaceX, and xAI to get enough compute performance for next-generation AI and robotics applications.

"Elon has a proven track record of reimagining entire industries," Lip-Bu Tan wrote in his own X post. "This is exactly what is needed in semiconductor manufacturing today. Terafab represents a step change in how silicon logic, memory and packaging will get built in the future. Intel is proud to be a partner and work closely with Elon on this highly strategic project."

https://www.tomshardware.com/tech-i...tesla-to-help-refactor-silicon-fab-technology
Click to expand...
YUP!!! :LOL:
He's so damn predictable. So Intel is going to build this fab for him and he's going to slap his branding on it.
 
You must log in or register to reply here.
Back
Top