Blackwell

“Blackwell has 288 gigabytes of HBM capacity around the logic die, and the logic die itself maybe only has like 500 megabytes of SRAM. So it's possibly multiple orders of magnitude, three orders of magnitude difference in density for DRAM versus SRAM.”

“NVIDIA sees the, if they just let the most deep pockets buy all the chips that Maybe hurts them in the long term if that customer ends up accruing a lot of more power. They understand that compute is power today, and so they're quite strategic about how they a…”

“NVIDIA's pumping out five million Blackwell chips this year.”

“If you look at, you know, Hopper to Blackwell to Rubin, and you compare like for like, what is the performance per watt of a beefload-sixteen multiply? It hasn't improved all that much.”

“If you look at the amount of gigawatts that are supposed to come on, and consensus estimates for hyperscalers, they're effectively modeled, and these are gigawatts of Blackwell and Rubin. Rubin being NVIDIA's next chip, Blackwell being the current chip. They a…”

“So on Blackwell, we find that with roughly 15 SMs out of one 48 TMA is able to saturate the NVLink.”

“For example, Cursor is using it to train Composer on tens of thousands of Blackwell GPUs.”

“Why is Colossus charging 12 dollars an hour for Blackwells? It's because they're the only place you can buy 20,000 of them at once, right?”

“And so for the hoppers, the H 100 or H 200 series of GPUs into the Blackwells there was a claim made that it could be 30 times more efficient. And I think the data from, you know, some analysis showed that it was 90 to a hundred times more efficient in terms o…”

“Over time, like, Hopper's now at five percent, but Blackwell's still 10 to 15%. Actually started out higher than that.”

“A lot of different SKUs were canceled, and the reason for that is it was by far the most complex product transition we've ever gone through in technology.”

“One, we will see the first models trained on Blackwell in early, I think the first Blackwell model will come from XAI. And the reason for that is just, it's a, According to Jensen, no one builds data centers faster than Elon.”

“But with Blackwell, and for sure with Rubin, economics are going to dominate the prisoner's dilemma from a decision making and spending perspective, just because the numbers are so big.”

“Those Blackwells are being used for training and there's no ROI on training. Training is you're making the model. The ROI comes from inference.”

“The problem is Blackwell is going to come out now, and the gap between these American frontier labs and Chinese open source is going to blow out because of Blackwell.”

“And then geopolitically, China thought they had the leverage. They're gonna realize, oh, whoopsie daisy, we do need the Blackwells. And unfortunately, they'll probably, for them they'll probably realize that in late 26.”

“If Blackwell comes back to China, which it seems like it probably will happen, that will also be very good because then Chinese open source will be back.”

“Ruben is going to be such an easy, seamless transition relative to Blackwell”

“Like, if you look at the numbers, like, even going from Hopper to Blackwell, like, the performance per watt is about the same. They mostly make the number of transistors go up, and they make the chip size go up, and they make the power usage go up. But even fr…”

“In the three months that ended in August, Oracle lost nearly a hundred million from the rentals of NVIDIA's Blackwell chips.”

“Nvidia created Blackwell, they had a ton of manufacturing problems and challenges with it for their supply chain balance sheets went up for, you know, various companies in the supply chain are building servers and stuff, because they're trying to figure it out…”

“Today, I just was in a data center in Silicon Valley seeing a brand new deployment of NVIDIA GB 200, the Blackwell generation. The rack is 132 kilowatts. It's liquid cooled, and we're headed toward one megawatt rack.”

“All indications are, even though they, you know, more than doubled the theoretical memory bandwidth going from Hopper to Blackwell, the actual percentage of theoretical that you can achieve is, again, going to be less than the previous generation”

“80% of the cost of a GPU data center if you're building Blackwell is capital. Yeah. Right? It's the GPU purchases, it's the networking, it's the physical data center conversion, power conversion equipment. All of this stuff is like 80% of the cost. And then, 2…”

“There's never been a product transition like this in the history of semiconductors going from Hopper to Blackwell, you know, kind of the 2022 generation GPU to this one. The only precedent for this on planet earth is the iPhone.”

“The top four cloud providers in the United States and what the year over year order was, you know, hoppers, 1.3 million hopper GPUs to the top four CPUs or CSPs and comparing like time frame. He said, we've done, you know, 3.6 million Blackwell GPUs so far thi…”

“I think Tranium is not even competitive with Hopper, let alone with where Blackwell is today.”

“People are surprised that, that that Blackwell is such a great leap over Hopper, and the reason for that is because we built Blackwell for inference, and at a, just in time, because all of a sudden thinking is such a big computing load”

“For Blackwell, they assembled two chips. But the surface was so big that the chip started to, you know , wave, like, I don't know the English word, but like, you know, started to bend a bit, which further perpetuated the problem because it then didn't make con…”

“If they can't get their year olds up and produce it, that will be a, you know, concerning thing for NVIDIA and will slow down some of, you know, kind of AI's development because the faster chips in a dense environment is actually, in fact, instrumental and hel…”

“Google actually did this alongside Broadcom you know, and they did it before Nvidia, right? You know, today everyone's freaking out about, or not freaking out, but like everyone's like very excited about Nvidia's Blackwell system, right? It is a rack Of GPUs. …”

“Like with Blackwell, not only is it way, way, way faster, anywhere from 10 to 15 times on really large models for inference, because they've optimized it for very large language models, they've also decided, hey, we're gonna cut our margin, too, somewhat, beca…”

“The cost to make Blackwell is north of two X that of the cost to make Hopper, right?”

“When NVIDIA has introduced a couple of new chips first the Samsung HVM there were heating issues back then with Blackwell again, heating issues came up where that delay that rollout was delayed due to due to heating issues.”

“I actually think you can argue that Elon doing that in a lot of ways kind of saved Nvidia from a tough six month period when Blackwell was delayed because everyone who was waiting for Blackwell and thought it was impossible to make a 100,000 hoppers coherent r…”

“We're going to announce that the P six, which is the Blackwell based instance that's coming early next year.”

“I think we were expecting about two and a half times the compute performance out of the Blackwell chip that you get out of an H-one hundred”

“In Asia, we just, it was 15 to one demand to supply. Four months ago, it was 12 to one.”

“They just told us they sold out of Blackwell chips for 12 months.”

“Our cyber security system today can't run without Without our own agents. We have agents helping to design chips. Hopper wouldn't be possible. Blackwell wouldn't be possible. Ruben don't even think about it.”

“In fact, getting it very efficient may actually drive more use of it to where it'll increase and not decrease. And so I think there's a much broader view you have to take when someone says something's more efficient, it's, and they say it's going to solve all …”