Why even free ASICs lose to NVIDIA when power is the limit

Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

1. 50:26

   our competitors building ASICs, they could literally, all their chips are cheaper already today, but they could literally price them at zero. Our objective is they could price them at zero, and you would still buy an NVIDIA system because the total cost of operating that system power data center land etc the intelligence out is still a better bet than buying a chip even if it's given to you for free because the land power and shell is already 15 billion dollars right yeah so we've taken a crack at kind of the math on that but walk us through your math because i think for people who don't spend as much time here that they it just doesn't compute how could it possibly be that you were pricing your competitors chips at zero given the expense of your chips and it still is a better bet there's two ways to think about it um one way is um uh let's just think about it

2. 51:15

   from a perspective of revenues yes okay so everybody's power limited and let's say uh you were able to secure two more gigawatts of power well that two gigawatts of power you would like to have translate to revenues yes so your performance or tokens per watt was twice as high as somebody else's token per watt, because I did deep and extreme co-design, and my performance was much higher per unit energy, then my customer can produce twice as much revenues from their data center.

3. 51:57

   And who doesn't want twice as much revenue? revenues and and and if somebody gave them a 15 discount you know the difference between our gross margins was called 75 points and somebody else's gross margins call it 50 to 65 points it's not so much as to make up for the 30 times difference between blackwall and hopper let's pretend hopper hopper's an amazing chip an amazing system let's pretend somebody else's asic is hopper yeah blackwalls 30 times so you've got to give up 30x revenues in that one gigawatt it's too much to give up so even if they gave it to you for free you use you only have two gigawatts to work with your opportunity cost is so insanely high you would always choose the best perf per one so i heard this from one of the cfos at one of the hyperscalers that given the performance improvement, right, that's coming out of your chips, again, precisely to that point,

4. 52:58

   tokens per gig, and power being the limiting factor, right, that they had to upgrade to the new cycle. So, when you look ahead at Rubin, at Rubin Ultra, at Feynman, does that trajectory continue? We're building, what, six, seven chips a year now? Yeah. As part of that system. That's That's right, and that system, software is everywhere, and it takes the integration and the optimization across all of those six, seven chips to deliver on the 30X Blackwell.

5. 53:30

   Now, imagine I'm doing this every single year. Bam, bam, bam, bam, bam, bam. And so, if you build one ASIC in that soup of ASICs, in that soup of chips, and we're optimizing across that, you know, it's a hard problem to solve. This does bring me back to where we started about the competitive moat. We've been covering this and investors for a while.