you need to treat Musk's words like you would to China's CCP. you need to see TeraFab in operation and produce real chip before you believe it.
Terafab 21 March 2026
- Thread starter user nl
- Start date
Xebec
Well-known member
Of course - I'm not posting this because I "believe it", but just pointing out - they're building the hype for it.
That said, I trust China's CCP when they talk about introducing new social control measures.
user nl
Well-known member
https://www.digitimes.com/news/a20260330PD217/elon-musk-wafer-fab-ic-manufacturing.html?chid=12
Elon Musk recently announced the launch of the "Terafab" project, aiming to expand compute production capacity to 1 TW per year, which is about 50 times the current global compute supply of around 20 GW. He plans to build advanced wafer fabs first in Austin, Texas. However, with the 1 TW scale revealed, market reactions have been relatively cautious.
With a typical fab producing 50,000 wafers per month, this would require 140-360 new fabs, and at US$35 billion per fab, total investment could reach US$5-13 trillion. This is arguably more difficult than landing on Mars. As a result, Terafab is viewed by some as largely conceptual or speculative at present.
.............................................................................................
Elon Musk recently announced the launch of the "Terafab" project, aiming to expand compute production capacity to 1 TW per year, which is about 50 times the current global compute supply of around 20 GW. He plans to build advanced wafer fabs first in Austin, Texas. However, with the 1 TW scale revealed, market reactions have been relatively cautious.
Skepticism over unrealistic scale
Such a capacity would recreate the global semiconductor industry, and its immediate impact on the foundry market is likely limited. Bernstein analysts point out that, based on current compute architectures like Nvidia racks, the wafer demand for 1 TW of compute equals nearly the entire existing global semiconductor production, which is around 16 million 12-inch wafers per month.With a typical fab producing 50,000 wafers per month, this would require 140-360 new fabs, and at US$35 billion per fab, total investment could reach US$5-13 trillion. This is arguably more difficult than landing on Mars. As a result, Terafab is viewed by some as largely conceptual or speculative at present.
.............................................................................................
Xebec
Well-known member
Website launched - looks like it's intended for both logic and memory:
terafab.ai
...
Scrolling down they mention 3 future chips:
AI5 - Powering FSD & Tesla Optimus
AI6 - Powering Tesla Optimus
D3 - Powering Space
They also show a picture of a Lunar Mass Driver (presumably - a way to shoot moon rocks into space)
Join Us On Our Journey
terafab.ai
Tesla, SpaceX, and xAI are launching the most epic chip-building effort ever - combining logic, memory and advanced packaging under one roof.
Terafab will close the gap between today's chip production and the future's demand - a future among the stars.
...
Scrolling down they mention 3 future chips:
AI5 - Powering FSD & Tesla Optimus
AI6 - Powering Tesla Optimus
D3 - Powering Space
They also show a picture of a Lunar Mass Driver (presumably - a way to shoot moon rocks into space)
how about the 2030 moon landing?
How about Made in China 2025 proposed in 2015?
Chinese leaders know the difference between hype and planning, most westerners do not.
If Elon Musk had even half the ability of ccp , he’d be truly the great..
Xebec
Well-known member
What kind of response are you looking for here?
Why are you going on about "Westerners vs. Easterners" on a post about Terafab announcing it's goals?
Xebec
Well-known member
When I worked at Lockheed Martin, we always enjoyed having discussions around their (at the time) primary financial data center being in Lakeland, FL, because "statistically, hurricanes are less likely to cause damage than [down the highway] in Orlando".
