Cern BasherGuestHerbertHostLet's explain this slide, because this is one of your greatest slides.
This is Q1, Q2, Q3, Q4.
You think that if they are supposed to shoot 20 gigawatts, this is what they would need to do.
That's
In order to hit the number, because each GB300 mini hard data center, 220,000 ships, which is standardized on that number because of... architectural configurations and efficiency, and also they made software that takes advantage of that to have faster compute.
But then you need 1.5 times more for a Vera Rubin same configuration because it needs more power.
But still, in order to hit possible 20 gigawatts, then and you go to VROOM in the 2027, then you need about 24 of those, which means you need to build two or three a year, sorry, a month in order to get to that number, right? Especially if I start with one, right? So then they're going faster and faster according to my analysis.
The question was, do they have enough via Rubens and memory to fill all of them? Elon said maybe no, but all the buildings and the other stuff would need.
But the number of them needs to be about 28 based upon the power usage of a Ruben data center.
if each one's about 600, 700 megawatts, then I need, you know, this 24 to 30 kind of number in order to get to the 20, right? Taking Elon and his word, right? You can always say, you know, give haircuts.
Elon said there's a zero, it's no longer a 0% probability that they will get to 20 gigawatts Right.
But it's not likely, he said.
It's not like we're saying we're going to do it.
But this is, in order for him to do it, this is how many Vera Rubin data centers, mini-hours, you call them a hall.
Yeah.
Is that what they're called? Halls? But they're basically a building, right? So maybe, Cern, you can tell us what one of these looks like, because you need to build that many.
And they're buildings.
Let's explain this slide, because this is one of your greatest slides.
This is Q1, Q2, Q3, Q4.
You think that if they are supposed to shoot 20 gigawatts, this is what they would need to do.
That's
In order to hit the number, because each GB300 mini hard data center, 220,000 ships, which is standardized on that number because of... architectural configurations and efficiency, and also they made software that takes advantage of that to have faster compute.
But then you need 1.5 times more for a Vera Rubin same configuration because it needs more power.
But still, in order to hit possible 20 gigawatts, then and you go to VROOM in the 2027, then you need about 24 of those, which means you need to build two or three a year, sorry, a month in order to get to that number, right? Especially if I start with one, right? So then they're going faster and faster according to my analysis.
The question was, do they have enough via Rubens and memory to fill all of them? Elon said maybe no, but all the buildings and the other stuff would need.
But the number of them needs to be about 28 based upon the power usage of a Ruben data center.
if each one's about 600, 700 megawatts, then I need, you know, this 24 to 30 kind of number in order to get to the 20, right? Taking Elon and his word, right? You can always say, you know, give haircuts.
Elon said there's a zero, it's no longer a 0% probability that they will get to 20 gigawatts Right.
But it's not likely, he said.
It's not like we're saying we're going to do it.
But this is, in order for him to do it, this is how many Vera Rubin data centers, mini-hours, you call them a hall.
Yeah.
Is that what they're called? Halls? But they're basically a building, right? So maybe, Cern, you can tell us what one of these looks like, because you need to build that many.
And they're buildings.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.