Nvidia, every mention
132 scenes, the whole family · ← back to Nvidia
tap a year for its mentions
every year anyone Gavin Baker 99Patrick O'Shaughnessy 54Neil Movva 54Dylan Patel 33Rob Wachen 17Ben Thompson 9Dan Loeb 8Alex Sacerdote 8Gavin Uberti 5Dan Wang 5
Verbatim, from the transcripts: the passages where Nvidia comes up
Ex-NVIDIA Engineer: Why AI Is About to Get 1000x Cheaper
- ▶ 13:17 Neil Movva And the first thing we did was we tried to build the entire LLM software stack around peak GPU efficiency, meaning we're using NVIDIA GPUs. 2 times in the scene
- ▶ 14:21 Neil Movva So NVIDIA, great graphics company, obviously has had market share dominance in GPUs and, and gaming graphics for quite some time. 12 times in the scene
- ▶ 20:39 Patrick O'Shaughnessy And it's a great analogy, and so, step one for what you're trying to do is, like, create the best possible bus on top of NVIDIA GPUs.
- ▶ 20:55 Neil Movva With NVIDIA GPUs, one of the things that they really innovated on and did a great job with is the NVLink interconnect between GPUs. 6 times in the scene
- ▶ 24:24 Neil Movva The problem with SRAM is it takes a lot of area on the silicon die, uh, so if you want to build a large die, like let's say the NVIDIA Blackwell at 800 millimeters square, if you made that whole die SRAM,
- ▶ 24:24 Neil Movva The problem with SRAM is it takes a lot of area on the silicon die, uh, so if you want to build a large die, like let's say the NVIDIA Blackwell at 800 millimeters square, if you made that whole die SRAM,
- ▶ 26:00 Neil Movva Uh, solder them around the main logic die that you get from NVIDIA. 2 times in the scene
- ▶ 26:04 Neil Movva You can now get hundreds of gigabytes, uh, like Blackwell has 288 gigabytes of HBM capacity around the logic die, and the logic die itself maybe only has like 500 megabytes of SRAM. 2 times in the scene
- ▶ 32:06 Neil Movva Something like a matrix multiply, and it's really good to host the, the MLP, the weights, essentially, on the Cerebrus chip, but the GPU has the capacity to scale to really long context lengths, and so you would like to put the attention,…
- ▶ 42:44 Patrick O'Shaughnessy Where are we in that, like, if I just think about a Blackwell or something, and Blackwell is the piece of coal, like, what percent do you think we're at? 2 times in the scene
- ▶ 43:14 Neil Movva The way NVIDIA quotes peak flops is a little optimistic. 3 times in the scene
- ▶ 44:32 Patrick O'Shaughnessy One of the things you hear is that the market for the best chips, Blackwells, let's say, is like a drug market or something right now. 3 times in the scene
- ▶ 45:04 Neil Movva Basically, NVIDIA has a long-term view on, uh, on all their chips. 6 times in the scene
- ▶ 50:46 Neil Movva This is also even different from what we had two years ago, where there was a supply crunch for hopper generation chips in 20, 23, 24.
- ▶ 1:00:02 Neil Movva And if my chips are cheap enough, they're probably not going to be Nvidia racks.
- ▶ 1:04:57 Neil Movva NVIDIA's pumping out five million Blackwell chips this year.
- ▶ 1:08:04 Patrick O'Shaughnessy You talked about some interesting NVIDIA lessons. 2 times in the scene
- ▶ 1:16:31 Neil Movva One of the things I keep coming back to is in this question of NVIDIA. 4 times in the scene
- ▶ 1:16:40 Neil Movva They're always going to reinvent themselves, but like, fundamentally, I think one thing that surprises people is when I tell them that, hey, if you look at, you know, Hopper to Blackwell to Rubin, and you compare like for like, like, what…
- ▶ 1:16:40 Neil Movva They're always going to reinvent themselves, but like, fundamentally, I think one thing that surprises people is when I tell them that, hey, if you look at, you know, Hopper to Blackwell to Rubin, and you compare like for like, like, what…
- ▶ 1:16:40 Neil Movva They're always going to reinvent themselves, but like, fundamentally, I think one thing that surprises people is when I tell them that, hey, if you look at, you know, Hopper to Blackwell to Rubin, and you compare like for like, like, what…
- ▶ 1:19:32 Neil Movva You're building a chip because you think that NVIDIA has made some choices that are difficult for them to change, which is true. 7 times in the scene
- ▶ 1:21:36 Neil Movva A lot of the people in NVIDIA that I mentioned earlier who instilled that, like, love of performance engineering in me, but also my professors in college who I remember, like, my advisor in, like, sophomore year, I was a very impatient…
What Happens When the AI Boom Runs Out of Money · Invest Like The Best
- ▶ 10:55 Ben Thompson To me, this NVIDIA deal is very much
- ▶ 51:43 Ben Thompson There's a reason they're using Google Cloud and NVIDIA chips, but you can certainly imagine a future where, where that's the case.
- ▶ 1:14:08 Patrick O'Shaughnessy One major player and company that we haven't talked about much is Jensen and NVIDIA.
- ▶ 1:15:36 Patrick O'Shaughnessy So curious for your thoughts on, yeah, the gents and NVIDIA specifically. 24 times in the scene
- ▶ 1:19:53 Ben Thompson They start talking about, they tried to come up with these Nemo, they had the Nemo Tron models, but they had all these, they had this thing in 24, I remember, it was the first one where it was like the Rockstar GTC at San Jose, and like… 2 times in the scene
- ▶ 1:21:50 Ben Thompson I think what NVIDIA is hoping for, maybe they wouldn't say this in so many words, but 5 times in the scene
Everyone Is Still Undersizing the AI Market | Eric Vishria · Invest Like The Best
- ▶ 2:26 Eric Vishria This is the same open source model with the same NVIDIA hardware
- ▶ 33:07 Eric Vishria Like, NVIDIA was worth, like, forty billion, not four trillion. 2 times in the scene
- ▶ 1:12:12 Eric Vishria Will NVIDIA do well?
Why the Markets Are Pricing AI Wrong | Gavin Baker · Invest Like The Best
- ▶ 0:21 Gavin Baker The underlying fundamentals are improving, and stocks, NVIDIA is actually, as we record this, 3 times in the scene
- ▶ 5:11 Gavin Baker But there was an unusual amount of one-timers this, this quarter, and if you adjust for that, we went from 28 to 35, and that's, that's a, that's a material acceleration at this scale, and that's really before, like, they start to light up…
- ▶ 13:48 Gavin Baker Rubin being NVIDIA's next chip, Blackwell being the current chip.
- ▶ 20:59 Gavin Baker I think you're going to see NVIDIA bring Nebotron steadily closer to the frontier. 2 times in the scene
- ▶ 28:36 Gavin Baker Well, yeah, and also, like, Nvidia is heavily involved with all of these startups.
- ▶ 38:38 Gavin Baker There's Google with their TPUs, there's AMD, and then there's Nvidia, who's, like, much bigger than everybody else combined. 9 times in the scene
- ▶ 50:40 Gavin Baker but no, but just basically that, um, you know, renting an H 100 for a year would cost 250,000 dollars.
- ▶ 1:10:19 Patrick O'Shaughnessy Micron all of a sudden, you know, would be like a sample answer to the question of someone that becomes as important as Anthropic, OpenAI, Microsoft, Amazon, you know, NVIDIA, SpaceX, yeah.
Sam Altman on AGI, Compute, and Human Agency · Invest Like The Best
- ▶ 7:32 Sam Altman Nvidia has been a tremendous partner.
The Two Harvard Dropouts Who raised $800M to take on NVIDIA · Invest Like The Best
- ▶ 4:34 Gavin Uberti We think we can be much faster than NVIDIA, and Mark's like, no you can't, it will not work.
- ▶ 11:19 Gavin Uberti For example, on Blackwell chips, it can be about 4000 nanoseconds to go point to point.
- ▶ 15:56 Rob Wachen And, you know, the inference side, you know, today people usually think about it as an eight chip cluster, or maybe just, you know, NVL, uh, as the scale up domain, uh, but very quickly, this is going to become thousands of chips and tens…
- ▶ 23:13 Gavin Uberti They got bought by Nvidia for hundreds of millions of dollars.
- ▶ 26:36 Rob Wachen Uh, at that point we just started building our rack team and we brought over Brian Loyler, uh, who built all of NVIDIA's HGX and DGX systems, which is like 80% of their revenue. 2 times in the scene
- ▶ 26:36 Rob Wachen Uh, at that point we just started building our rack team and we brought over Brian Loyler, uh, who built all of NVIDIA's HGX and DGX systems, which is like 80% of their revenue.
- ▶ 26:36 Rob Wachen Uh, at that point we just started building our rack team and we brought over Brian Loyler, uh, who built all of NVIDIA's HGX and DGX systems, which is like 80% of their revenue.
- ▶ 29:23 Rob Wachen If we could find somebody who, like, started at NVIDIA and built the entire rack team through all their different generations, learned all this different stuff, but is still scrappy, still understands the startup culture, but has seen… 4 times in the scene
- ▶ 29:59 Rob Wachen Um, Brian started the HGX and DGX team at NVIDIA, uh, you know, which was, you know, a majority of NVIDIA's revenue, you know, tens of billions of dollars a quarter.
- ▶ 29:59 Rob Wachen Um, Brian started the HGX and DGX team at NVIDIA, uh, you know, which was, you know, a majority of NVIDIA's revenue, you know, tens of billions of dollars a quarter.
- ▶ 42:44 Rob Wachen Why is Colossus charging 12 dollars an hour for Blackwells? 2 times in the scene
- ▶ 43:31 Patrick O'Shaughnessy So if I think about a rack like this versus, I don't know, a set of Blackwells or something, or Rubens or whatever's coming next,
- ▶ 48:25 Gavin Uberti That, if you think about chip-to-chip latencies, on a NVIDIA product, you're looking at 4000 nanoseconds to go from one chip to another. 2 times in the scene
- ▶ 56:34 Rob Wachen It's NVIDIA, right? 2 times in the scene
- ▶ 1:18:35 Rob Wachen And that's why actually for our first gen product, we built it on a different supply chain than the Rubens. 2 times in the scene
Why the AI Boom Is Just Getting Started · Invest Like The Best
- ▶ 20:52 Alex Sacerdote When we were buying Nvidia in 2023, we were paying four times earnings.
- ▶ 51:02 Alex Sacerdote They're on the same, you know, they've got to be working with NVIDIA for three or four generations in advance. 2 times in the scene
- ▶ 59:34 Alex Sacerdote And then you also have to have the holistic view, because if you don't have conviction, so every, you know, every time with NVIDIA over the last four years, it's like, oh, they had a great year. 3 times in the scene
- ▶ 1:12:30 Alex Sacerdote Um, and then in their hedge fund portfolio, even if it's long bias, they're not going to have 15% in NVIDIA and all these other things. 2 times in the scene
Uber CEO on AI, Autonomous Vehicles, and the Future of Transportation · Invest Like The Best
- ▶ 24:19 Dara Khosrowshahi So we've got over 30 partnerships now with incredible, uh, partners like a Waymo to a Neuro and Lucid to an NVIDIA that is building not just compute and, and sensors, but also a software driver now to
Legendary Investor Dan Loeb on AI, Credit, & Third Point’s $25B Strategy · Invest Like The Best
- ▶ 3:58 Dan Loeb And I think that all changed when NVIDIA reported its March results three years ago. 3 times in the scene
- ▶ 20:32 Dan Loeb Nvidia in after Q one, three years ago had this monster quarter people piled on. 3 times in the scene
- ▶ 42:29 Dan Loeb I mean, you can still buy Nvidia. 2 times in the scene
Watts, Wafers, and the Future of AI Infra | Gavin Baker · Invest Like The Best
- ▶ 16:27 Gavin Baker A Blackwell rack weighs, you know, 3000 pounds. 3 times in the scene
- ▶ 25:58 Gavin Baker If Taiwan Simi did what Jensen wanted, I think NVIDIA could sell two trillion dollars of GPUs.
- ▶ 31:02 Patrick O'Shaughnessy So like Terra fab is going to be pumping out Nvidia or whatever GPs, whatever chips, like
- ▶ 33:36 Gavin Baker Take it away partially from Broadcom and NVIDIA continuing to make aggressive choices.
- ▶ 42:11 Patrick O'Shaughnessy Like we talked a lot about Nvidia and, you know, their, their sort of relationship with the SMC and Intel and all these sorts of things. 4 times in the scene
- ▶ 49:41 Gavin Baker What rapid technological change has done with the disaggregation of pre-fill and inference is mean that, you know, you can put a Cerebra system or Grok LPUs that NVIDIA acquired.
- ▶ 49:52 Gavin Baker Effectively in front of a hopper, or even an ampere, use that hopper and ampere for prefill, and extend the useful life of that GPU until it melts. 2 times in the scene
- ▶ 49:52 Gavin Baker Effectively in front of a hopper, or even an ampere, use that hopper and ampere for prefill, and extend the useful life of that GPU until it melts. 2 times in the scene
- ▶ 57:53 Gavin Baker The reality is like if, if a company like Nvidia or AMD were to ever really, really use one of these other foundries, that foundry would get better really quickly.
- ▶ 1:03:27 Gavin Baker I think it's hard to square like the valuation of something like Nvidia, which is still, you know, in, in, in early April was essentially as cheap as it gets relative to the market, like in the last 10 or 12 years or whatever it is, and… 2 times in the scene
- ▶ 1:09:48 Gavin Baker It's just interesting, and it just means this NVIDIA effect we discussed is even more powerful than maybe I'd imagined, but I'm very curious to see what the Pareto frontier looks like literally in five days after Google's announced its new… 3 times in the scene
Inside Anthropic's $100 Billion Al Compute Commitment | CFO Krishna Rao · Invest Like The Best
- ▶ 2:56 Krishna Rao You know, we use three different chip platforms, so we are customers of Amazon's Tranium chip, Google's TPUs, and NVIDIA's GPUs. 2 times in the scene
- ▶ 4:14 Patrick O'Shaughnessy Am I thinking about this in the right way that like something like CUDA that has been a part of Nvidia's story for a long time now, that allows you to do a lot with the underlying actual hardware.
- ▶ 33:14 Patrick O'Shaughnessy And this is true at different levels, whether it be, you know, the mythos pricing that is quite high because it's so powerful, the cost of an H 100, you know, the rental price of a cost of an H 100 is, well, you know, it looks like a smile…
- ▶ 38:29 Krishna Rao So we're fortunate in that we have really great partners in Amazon, in Google, in Microsoft, but also with Broadcom and NVIDIA as well.
The Supply and Demand of AI Tokens | Dylan Patel Interview · Invest Like The Best
- ▶ 25:08 Dylan Patel A 100,000 blackwells is equivalent to hundreds of thousands of prior generation chips.
- ▶ 28:27 Dylan Patel There is such demand for these tokens and such limitations on compute, you know, and we see this with H 100 prices skyrocketing and the useful life of these GPUs continue to extend.
- ▶ 30:52 Patrick O'Shaughnessy Each 100 prices look like this. 3 times in the scene
- ▶ 31:05 Dylan Patel Um, there's a 100 clusters that are re-signing for another couple of years.
- ▶ 31:33 Dylan Patel With, you know, Nvidia still charging 75 or whatever percent gross margin. 2 times in the scene
- ▶ 37:09 Patrick O'Shaughnessy What about, like, CPUs or ASICs or things that start to pop out as both opportunities and bottlenecks beyond just, like, NVIDIA's GPU dominance?
World's Top Researcher on AI, LLMs, and Robot Intelligence · Invest Like The Best
- ▶ 1:12:27 Sergey Levine And there was another moment even earlier on that maybe is, like, even more minor where, when I was in college, I got an internship at NVIDIA that, uh, really got me to, like, experience some cool, uh,
Inside Dan Sundheim's Bets on Anthropic, OpenAI, and SpaceX · Invest Like The Best
- ▶ 26:09 Dan Sundheim Uh, and I think there's a lot of interest, uh, from NVIDIA and, uh, other chip companies to make sure that their customer base is diversified. 2 times in the scene
Why The Laws of Startup Physics Have Changed | Ben Horowitz Interview · Invest Like The Best
- ▶ 54:42 Ben Horowitz Jensen has like this, like very defined, you know, agree with her or not, but it's like this view of who he is, what the company is, and so forth, that's kind of gone across eras, um, 3 times in the scene
GPUs, TPUs, & The Economics of AI Explained | Gavin Baker Interview · Invest Like The Best
- ▶ 0:37 Patrick O'Shaughnessy We talk about NVIDIA, Google and its TPUs, the changing AI landscape, the changing math and business models around AI companies.
- ▶ 6:33 Gavin Baker And the reason for that is, you know, after XAI figured out how to get, um, 200,000 hoppers coherent, 3 times in the scene
- ▶ 7:04 Gavin Baker So everything in AI has a struggle between Google and NVIDIA and Google has a TPU. 4 times in the scene
- ▶ 7:15 Gavin Baker You know, Nvidia has the full stack and Blackwell was delayed. 5 times in the scene
- ▶ 7:24 Gavin Baker The first iteration of that was the Blackwell 200.
- ▶ 10:02 Gavin Baker Since Hopper came out of the scaling law for pre-training, and it held, and that's great, because all these scaling laws are multiplicative, so now we're going to apply these two new, um, reinforcement learning, verified rewards, and test…
- ▶ 10:31 Gavin Baker It's like, almost like imagine, like Hopper is like, you know, it's like a World War II era airplane, and it was by far the best World War II era airplane. 5 times in the scene
- ▶ 10:49 Gavin Baker Because Blackwell was such a complicated product and so hard to ramp, Google was training Gemini three, 9 times in the scene
- ▶ 11:34 Gavin Baker NVIDIA is not worth trillions because they're the low-cost producer of AI accelerators. 4 times in the scene
- ▶ 12:40 Gavin Baker And by the way, when Hopper came out, it took six to 12 months for it to really outperform Ampere, which was the generation before.
- ▶ 13:32 Gavin Baker And so now that we know pre-scaling loss we're holding, we know that these Blackwell models are going to be really good. 3 times in the scene
- ▶ 13:43 Gavin Baker So the GB 200 was really, really, it was really hard to get it going. 3 times in the scene
page 1 of 2 · 100 scenes per page · newest episode first next →