WEBVTT - Nvidia Tumbles on Disappointing Forecast 0:00:00.160 --> 0:00:06.800 Bloomberg Audio Studios, podcasts, radio news. 0:00:09.320 --> 0:00:13.040 This is the Bloomberg Daybreak Asia podcast. I'm Brian Curtis 0:00:13.119 --> 0:00:15.840 along with Doug Krisner. Join us each day for the 0:00:15.880 --> 0:00:19.240 stories making news and moving markets in the Asia Pacific. 0:00:19.480 --> 0:00:21.920 You can subscribe to the show anywhere you get your 0:00:21.920 --> 0:00:25.880 podcasts and always on Bloomberg Radio, the Bloomberg Terminal, and 0:00:25.920 --> 0:00:27.240 the Bloomberg Business App. 0:00:27.520 --> 0:00:31.160 Let's bring in Bloomberg Intelligence Senior tech industry analyst man 0:00:31.200 --> 0:00:34.640 Deep Singh joining me here in the Bloomberg Interactive Brokers 0:00:34.640 --> 0:00:37.720 studio at Bloomberg Headquarters in New York. So you heard 0:00:37.720 --> 0:00:40.920 the conversation that Ed Ludlow had with CEO Jensen Wang. 0:00:41.479 --> 0:00:42.840 What did you think? What's your reaction? 0:00:43.240 --> 0:00:47.159 Results were great, it's still a very solid growth story. 0:00:47.200 --> 0:00:50.680 But when it comes to convincing investors that there are 0:00:50.720 --> 0:00:56.640 more catalysts for upside relative to where expectations are, I 0:00:56.680 --> 0:01:00.560 think a lot of it right now is the on 0:01:00.840 --> 0:01:06.000 how the industry actually deploys generative AI. That's what Jensen 0:01:06.080 --> 0:01:12.399 highlighted that. Look, GENAI is transformational. It's changing how compute 0:01:12.520 --> 0:01:16.160 is done and everyone will use it and we are 0:01:16.200 --> 0:01:18.880 seeing that in their numbers. The data center growth is 0:01:19.120 --> 0:01:24.880 just unprecedented. Question for investors is will we see more 0:01:25.200 --> 0:01:29.680 positive revisions? Will estimates go up by five ten percent tomorrow? 0:01:30.000 --> 0:01:34.600 And the answer is no, simply because right now they 0:01:34.640 --> 0:01:39.440 talked about Blackwell getting delayed to four Q. Obviously they 0:01:39.440 --> 0:01:43.360 are very positive on the ramp, but everything is more 0:01:43.720 --> 0:01:47.040 like pushed out by a quarter and they I mean, 0:01:47.120 --> 0:01:50.120 Jensen was very positive with ed you know in the 0:01:50.200 --> 0:01:54.600 interview about how he expects twenty twenty five to shake out, 0:01:55.120 --> 0:01:58.880 So he's his enthusiasm is intact. It's just when it 0:01:58.920 --> 0:02:02.080 comes to estimates going up, probably investors didn't see enough. 0:02:02.440 --> 0:02:05.280 There was a lot of conversation around the hyperscalers. These 0:02:05.280 --> 0:02:08.080 are the big cloud computing companies that have been investing 0:02:08.120 --> 0:02:10.680 in this technology, and I think the big question for 0:02:10.760 --> 0:02:13.000 the street is whether or not these companies are going 0:02:13.080 --> 0:02:16.240 to begin to demonstrate a return on investment the ROA 0:02:16.440 --> 0:02:18.720 on AI. What do we know about that part of 0:02:18.760 --> 0:02:19.200 the story. 0:02:19.360 --> 0:02:23.800 I think with Hyperscalers, specifically, their cloud business will ramp up. 0:02:23.880 --> 0:02:27.200 And that's what he said, like he's not Nvidia is 0:02:27.240 --> 0:02:30.639 not competing with the hyperscalers. They don't intend to become 0:02:30.760 --> 0:02:35.160 a hyperscaler vendor like a Microsoft, Amazon, or Google. They're 0:02:35.200 --> 0:02:39.840 deploying within Microsoft and Google, and I think that's where 0:02:40.560 --> 0:02:43.839 probably you will see a lot of ramp up from 0:02:43.840 --> 0:02:46.960 the hyperscalers. But that doesn't mean they will continue to 0:02:47.080 --> 0:02:50.480 raise Capex for twenty twenty six and beyond. Ultimately, the 0:02:50.560 --> 0:02:55.639 question for Nvidia investors is is this CAPEX wave sustainable? 0:02:56.000 --> 0:02:58.600 How long of a tail it has and is there 0:02:58.639 --> 0:03:01.040 a cliff three years out? 0:03:01.400 --> 0:03:04.400 Well, you mentioned the Blackwell design, the new chip, the 0:03:04.840 --> 0:03:07.680 GPU that in video will roll out. Apparently there's some 0:03:07.840 --> 0:03:10.920 production issues right now. Is this more so on the 0:03:10.960 --> 0:03:13.440 design side, that's the sense that I'm getting, rather than 0:03:13.480 --> 0:03:14.760 production itself, right. 0:03:14.880 --> 0:03:18.080 Yeah, I mean, look, the complexity of the chips is growing. 0:03:18.160 --> 0:03:21.560 He talked about power, how Blackwell will use the same 0:03:21.600 --> 0:03:25.040 amount of power but be able to compute more than 0:03:25.080 --> 0:03:29.200 the current Hopper architecture. So clearly they're doing a lot 0:03:29.320 --> 0:03:32.080 more on the content side of the chip and it 0:03:32.160 --> 0:03:36.920 requires sophisticated production, which TSMC is doing. But again he 0:03:37.080 --> 0:03:40.480 came across very confident that they will ramp up four 0:03:40.600 --> 0:03:44.520 Q and then if going into twenty twenty five, Blackwell 0:03:44.560 --> 0:03:49.440 will be multi billions in revenue, so that conviction was 0:03:49.480 --> 0:03:53.440 there in what he said during the interview. It's more about, 0:03:53.680 --> 0:03:59.040 you know, will it be enough to drive more upside 0:03:59.080 --> 0:04:03.560 in terms of the incremental buyers CAPEX going up. I mean, 0:04:03.600 --> 0:04:07.720 he talked about there are more companies that are trading models. 0:04:08.200 --> 0:04:12.000 Is just not clear enough, like who are these beyond 0:04:12.040 --> 0:04:16.479 the hyperscaler sovereigns he mentioned sovereigns he did, and those 0:04:16.480 --> 0:04:20.839 are low double digit and billion dollars. But again when 0:04:20.880 --> 0:04:24.400 you compare it to the hyperscalers, it's not very clear 0:04:24.839 --> 0:04:28.560 if the magnitude of those buyers will be the same 0:04:28.600 --> 0:04:29.760 as the hyperscalers. 0:04:29.839 --> 0:04:30.080 End. 0:04:30.240 --> 0:04:33.520 So you mentioned the production side, that's TSMC, the company 0:04:33.560 --> 0:04:36.279 that is the foundry for these Nvidia chips. Do we 0:04:36.320 --> 0:04:38.520 have an indication that TSMC is going to be able 0:04:38.560 --> 0:04:39.880 to keep up with demand as well? 0:04:40.080 --> 0:04:40.359 Yes. 0:04:40.440 --> 0:04:43.679 In fact, he was very clear that supply is growing, 0:04:43.839 --> 0:04:46.880 and it's growing in a way where TSMC is able 0:04:46.920 --> 0:04:52.960 to accommodate in Vidia becoming a larger buyer of the 0:04:52.960 --> 0:04:56.880 manufacturing capacity that TSMC has. So all the things that 0:04:56.960 --> 0:04:59.599 he mentioned of the on the call pointed to a 0:04:59.600 --> 0:05:04.120 steady ramp up more supply of chips but what it 0:05:04.160 --> 0:05:08.800 didn't suggest is where are more upside catalysts. I mean, 0:05:08.839 --> 0:05:12.440 everyone knows about the new architecture, everyone knows about the 0:05:12.560 --> 0:05:15.440 ramp up and supply, So it's more a question of 0:05:15.720 --> 0:05:19.240 is the upside going to come from networking or software 0:05:19.920 --> 0:05:23.080 or cloud. He said they don't intend to become a 0:05:23.080 --> 0:05:26.039 cloud player. So the software business, they said they're going 0:05:26.120 --> 0:05:28.120 to end at a two billion dollar rund rate, So 0:05:28.520 --> 0:05:33.840 that upside for consensus revisions probably that is what is missing. 0:05:33.960 --> 0:05:36.000 One of the things that I didn't hear any discussion 0:05:36.040 --> 0:05:39.359 around was the high bandwidth memory chips. And you need 0:05:39.400 --> 0:05:41.880 those to kind of work with the GPUs, don't you 0:05:41.960 --> 0:05:45.560 when you're creating these cloud infrastructure units. Yes, and we're 0:05:45.560 --> 0:05:48.919 talking about companies like sk Heinez, like Samsung. Do we 0:05:49.040 --> 0:05:51.039 know that these companies are going to be able to 0:05:51.120 --> 0:05:54.400 keep up at the same level that Nvidia is producing chips? 0:05:54.560 --> 0:05:57.000 Yeah, And actually that's a great point. So when you 0:05:57.160 --> 0:06:00.919 look at the gross margins for Nvidia, they seem to 0:06:01.000 --> 0:06:04.280 have peaked last quarter. So last quarter gross margins were 0:06:04.320 --> 0:06:07.960 seventy eight percent. And remember this business right now has 0:06:08.000 --> 0:06:11.679 no seasonality. I mean, it's growing, you know, triple digits 0:06:11.680 --> 0:06:14.960 so obviously you can't factor in any seasonality. This quarter, 0:06:15.240 --> 0:06:18.400 the margins were seventy five percent. They said they can 0:06:18.480 --> 0:06:21.920 maintain that mid seventy percent gross margin. But the reason 0:06:21.960 --> 0:06:25.400 why gross margin may trend down going forward is because 0:06:25.400 --> 0:06:27.359 they have to buy more HBM chips. 0:06:27.600 --> 0:06:28.120 They have to. 0:06:28.120 --> 0:06:32.680 Buy more advance packaging capacity from TSMC. So that's where 0:06:32.720 --> 0:06:36.040 the bill of materials does add up for Nvidia and 0:06:36.120 --> 0:06:38.320 that will start to impact their gross margins. 0:06:38.360 --> 0:06:41.560 Great conversation, Man, Deep, Thank you so much. Bloomberg Intelligence 0:06:41.600 --> 0:06:45.320 senior tech analyst Mandeep Singh joining us here as we 0:06:45.360 --> 0:06:47.960 look at the price action in shares of Nvidia in 0:06:48.080 --> 0:06:52.520 late trading down nearly seven percent. I'm Doug Krisner, and 0:06:52.640 --> 0:07:04.279 this is Bloomberg. We welcome our guest, Mario Morales, Group 0:07:04.440 --> 0:07:10.679 VP of Enabling Technologies and Semiconductors at IDC. He's joining 0:07:10.760 --> 0:07:14.760 us from Silicon Valley, right down the road from San Jose. Mario, 0:07:14.800 --> 0:07:16.800 good of you to make time. What was your reaction 0:07:16.880 --> 0:07:18.800 to what you heard tonight from in Nvidia. 0:07:20.320 --> 0:07:23.080 Well, I think I'm a lot more positive than what 0:07:23.120 --> 0:07:26.480 I've been seeing after the market closed for in video. 0:07:27.000 --> 0:07:30.120 I think when you look at what they basically shared 0:07:30.120 --> 0:07:33.600 with us is that they're still growing pretty phenomenally in 0:07:33.640 --> 0:07:36.480 the data center. I think the other markets that they 0:07:36.560 --> 0:07:40.760 serve are now returning back to more normalized growth. So 0:07:40.800 --> 0:07:42.840 I think those are all positives. I think that the 0:07:42.880 --> 0:07:46.320 market sentiment was a little bit down just because the 0:07:46.360 --> 0:07:48.920 expectations have been so high for the company. 0:07:49.400 --> 0:07:51.200 What do you make of the fact that the company 0:07:51.280 --> 0:07:54.880 is having problems with this new Blackwell chip design? Is 0:07:54.880 --> 0:07:55.800 that a concern to you? 0:07:57.520 --> 0:08:00.200 Not really, because I look at the position and they 0:08:00.200 --> 0:08:03.120 have such a dominant position right now. Even if they're 0:08:03.160 --> 0:08:05.680 late by a quarter or two, it still will not 0:08:05.800 --> 0:08:09.080 change the pecking order in terms of the incumbency that 0:08:09.120 --> 0:08:12.280 they continue to enjoy within the data center market. I 0:08:12.320 --> 0:08:15.200 think what you're looking at nowadays is that these chips, 0:08:15.240 --> 0:08:19.520 whether it's the Black Blackwell platform or any other that's 0:08:19.560 --> 0:08:22.720 in the market, they're very complex chips, and they're basically 0:08:22.760 --> 0:08:26.280 now putting all of these chiplets together in one very 0:08:26.280 --> 0:08:29.960 intricate package, and so you're going to have some yield 0:08:29.960 --> 0:08:33.360 issues initially when you ramp some of these new technologies. 0:08:34.160 --> 0:08:37.480 I know that about forty five percent of Nvidia's total 0:08:37.520 --> 0:08:41.480 business comes from the so called hyperscaleers, the companies that 0:08:41.559 --> 0:08:45.719 invest in cloud computing names like Alphabet. Amazon is in there, 0:08:45.760 --> 0:08:48.600 along with Microsoft, IBM, you could even add Oracle. I 0:08:48.640 --> 0:08:52.480 think if it's forty five percent of the business that 0:08:52.559 --> 0:08:55.720 has been really restricted to companies that have already built 0:08:55.760 --> 0:08:59.840 out their AI infrastructure, isn't it reasonable to think that 0:09:00.080 --> 0:09:03.520 growth is going to begin tapering off because those companies 0:09:03.559 --> 0:09:06.440 really at this point aren't in a position to spend more. 0:09:06.480 --> 0:09:07.199 They have enough. 0:09:09.120 --> 0:09:11.440 You know, when you look at the cloud service provider 0:09:11.480 --> 0:09:16.040 space and how they procure silicon, it tends to be lumpy. 0:09:16.160 --> 0:09:20.400 But the investment in AI has been so strong that 0:09:20.480 --> 0:09:23.880 I think it can continue on. Maybe not necessarily only 0:09:23.920 --> 0:09:26.480 by the cloud service providers, but I think the next 0:09:26.520 --> 0:09:30.040 wave that's going to come will be from telcos and enterprise. 0:09:30.160 --> 0:09:33.400 These are all industries that are very large that are 0:09:33.440 --> 0:09:37.839 just beginning to experiment with AI and want to really 0:09:38.200 --> 0:09:41.880 turn it on in a big way across their enterprises. 0:09:41.920 --> 0:09:44.600 So I see more growth ahead. I think the cloud 0:09:44.640 --> 0:09:48.120 service corriders might take a pause maybe over the coming quarters, 0:09:48.160 --> 0:09:51.200 but I still believe that fundamentally on a longer term basis, 0:09:51.200 --> 0:09:53.760 we're just still you know, maybe in the second inning 0:09:54.000 --> 0:09:56.360 of what we're seeing today in terms of the amount 0:09:56.400 --> 0:09:58.000 of investment that we're seeing in AI. 0:09:58.559 --> 0:10:02.199 So Nvidia designs that they are manufactured at a foundry 0:10:02.240 --> 0:10:06.600 that's operated by TSMC. That's essentially what TSMC does, they're 0:10:06.679 --> 0:10:10.680 the manufacturer. Are you concerned that TSMC may not be 0:10:10.760 --> 0:10:12.120 able to keep up with demand. 0:10:13.679 --> 0:10:16.280 Yeah, I mean that's continued to be the theme right 0:10:16.320 --> 0:10:18.600 for over a year now. I think last year we 0:10:18.640 --> 0:10:21.640 saw a correction in the semic metor space, but when 0:10:21.679 --> 0:10:24.160 you look at TSMC, they tend to buck the trend. 0:10:24.360 --> 0:10:25.160 They're a very. 0:10:26.480 --> 0:10:29.120 Strong player in the founder space. The leader they control 0:10:29.160 --> 0:10:32.560 over sixty percent of the market. Where they're investing heavily 0:10:32.640 --> 0:10:35.959 now is in advanced packaging. So when you hear about costs, 0:10:36.679 --> 0:10:39.080 all of that means is that that's where they're seeing 0:10:39.400 --> 0:10:41.680 the big investments that they need to make in order 0:10:41.720 --> 0:10:45.040 to serve companies like Nvidia and Broadcom and AMD and 0:10:45.120 --> 0:10:49.000 others that are now beginning to use more advanced packaging 0:10:49.000 --> 0:10:50.880 in order to continue to drive integration. 0:10:51.559 --> 0:10:56.599 So the GPUs that are designed by Nvidio manufactured by TSMC, 0:10:57.200 --> 0:10:59.959 they have to be paired with the high bandwidth memory 0:11:00.160 --> 0:11:04.280 chips do they not, and those devices or chips are 0:11:04.679 --> 0:11:08.079 created by companies like Samsung and s k Heinex in 0:11:08.120 --> 0:11:12.440 South Korea. How are those companies How have they been 0:11:12.480 --> 0:11:14.760 able to keep up with the demand. Should we be 0:11:14.840 --> 0:11:18.400 concerned about maybe a little weakness on that side. 0:11:19.200 --> 0:11:22.640 Well, we're not seeing the weakness. In fact, since probably 0:11:22.640 --> 0:11:27.280 September of last year, the memory companies began to change 0:11:27.320 --> 0:11:29.960 their operations in order to get back to profitability, but 0:11:30.400 --> 0:11:34.000 demand returned, especially from the data center, and today we're 0:11:34.000 --> 0:11:37.920 expecting that both DRAM and NAND will grow significantly from 0:11:37.960 --> 0:11:41.040 a revenue basis, close to seventy three percent in terms 0:11:41.040 --> 0:11:43.880 of revenue growth, so very phenomenal growth. A lot of 0:11:43.920 --> 0:11:46.959 that led not just by high bandwidth memory, but also 0:11:47.320 --> 0:11:50.719 DDR five that's being used in the flagship smartphones and 0:11:51.240 --> 0:11:55.280 the upcoming aipcs that we'll see later during the holiday season. 0:11:55.720 --> 0:11:58.000 So I think in general that the demand for memory 0:11:58.040 --> 0:12:02.120 will still remain robut this year and likely into twenty 0:12:02.160 --> 0:12:02.720 twenty five. 0:12:03.200 --> 0:12:05.480 One of the things that is simply stunning when you 0:12:05.480 --> 0:12:08.240 look at any area of high technology is the rate 0:12:08.320 --> 0:12:13.320 of change, how quickly technology evolves. Give me a sense 0:12:13.320 --> 0:12:15.880 of what we're seeing now with not only a company 0:12:16.000 --> 0:12:19.400 like in Nvidia, but broadly when it comes to companies 0:12:19.440 --> 0:12:22.960 that are involved in artificial intelligence. Is some of this 0:12:24.640 --> 0:12:27.240 equipment going to be obsolete a year from now? Do 0:12:27.280 --> 0:12:27.680 you think? 0:12:29.000 --> 0:12:29.240 Well? 0:12:29.760 --> 0:12:32.520 You definitely. When you look at the SEMAC, nintor space, 0:12:32.559 --> 0:12:36.000 it's always followed Moore's law, which means that that performance 0:12:36.080 --> 0:12:39.760 or power changes every eighteen to twenty four months. Now 0:12:39.840 --> 0:12:43.440 you add AI to this, and when you start looking 0:12:43.440 --> 0:12:46.559 more at the large language models, they're actually changing every 0:12:46.600 --> 0:12:49.880 three to four months. So we've seen an acceleration of 0:12:50.080 --> 0:12:53.240 change that's happening. And you look at companies like in 0:12:53.320 --> 0:12:57.320 Nvidia or AMD, they're trying to go and keep up 0:12:57.320 --> 0:12:59.440 with that change. And I think that's going to be 0:12:59.640 --> 0:13:03.119 the big challenge for these companies because they've been operating 0:13:03.920 --> 0:13:06.960 with More's law in mind. But these but the market 0:13:07.040 --> 0:13:10.080 for AI is moving so much faster. So now AMD 0:13:10.360 --> 0:13:14.520 and Nvidia and Intel now have to basically introduce products 0:13:14.679 --> 0:13:17.160 every single year now, and you're going to have some 0:13:17.240 --> 0:13:20.199 hiccups because of that, because the cadence is now changing 0:13:20.240 --> 0:13:22.360 in order to keep up with this demand and this 0:13:22.440 --> 0:13:26.520 appetite that the world has for AI, especially around infrastructure. 0:13:26.720 --> 0:13:29.560 What is your sense of the competition that in Vidia 0:13:29.720 --> 0:13:32.199 may face in the future. Right now, there is clearly 0:13:32.320 --> 0:13:35.679 very little. The company is essentially a monopolist when it 0:13:35.679 --> 0:13:38.760 comes to these GPUs. But can you imagine a world 0:13:38.800 --> 0:13:41.839 where Nvidia begins to look over at shoulder a bit? 0:13:43.640 --> 0:13:46.040 I think so. I think over the coming year, you're 0:13:46.040 --> 0:13:49.640 definitely going to see AAMD stepping up their game. They 0:13:49.760 --> 0:13:52.559 you know, back in computext in June they announced an 0:13:52.559 --> 0:13:56.040 annual cadence for their own accelerators with the M three 0:13:56.120 --> 0:13:58.120 hundred beginning to fray, and then you got the three 0:13:58.200 --> 0:14:00.680 twenty five and three fifty. So you're going to see 0:14:00.720 --> 0:14:04.320 AMD now introducing an annual cadence of accelerators in order 0:14:04.679 --> 0:14:07.520 to try to catch up to Nvidia. But I think 0:14:07.840 --> 0:14:12.280 in Nvidia's own customers continue to chip away little by 0:14:12.320 --> 0:14:15.480 little in terms of the instances and services that they're 0:14:15.520 --> 0:14:18.080 turning on with their own silicon, And so when I 0:14:18.120 --> 0:14:21.240 look at the competition, I think it's going to come 0:14:21.320 --> 0:14:24.920 more so from the customers. Initially, so a company like 0:14:25.000 --> 0:14:31.840 Google and Microsoft and AWS all developing their own chips 0:14:32.240 --> 0:14:34.840 and using some of those ships with their infrastructure, and 0:14:34.880 --> 0:14:37.880 then on a longer term basis. Broadcom is still a 0:14:37.920 --> 0:14:40.720 significant player here that not a lot of people talk about. 0:14:40.720 --> 0:14:42.960 But when you think about a lot of the designs 0:14:42.960 --> 0:14:46.480 of the APIs that are today serving the cloud service providers, 0:14:46.640 --> 0:14:49.080 a lot of those acs are designed in collaboration with 0:14:49.160 --> 0:14:51.760 Brock them, so I would call them the number two player. 0:14:51.960 --> 0:14:55.680 And then AMD coming in being able to leverage their 0:14:55.720 --> 0:15:00.040 CPU entrenchmen, especially in the data center, and hopefully that 0:15:00.080 --> 0:15:05.480 will help accelerate their own entrance into the GPU market 0:15:05.520 --> 0:15:07.040 for training and infancy. 0:15:07.600 --> 0:15:10.960 Mario very quickly in about thirty seconds. A big semiconductor 0:15:10.960 --> 0:15:14.080 industry a conference next week in Taiwan. What do you 0:15:14.120 --> 0:15:18.280 think the big issues are for the semiconductors supply chain 0:15:18.520 --> 0:15:19.640 in Asia right now? 0:15:20.840 --> 0:15:23.640 I think that we're still seeing a slower recovery in 0:15:24.040 --> 0:15:27.600 the spending and I'm talking about capital spending of the 0:15:27.760 --> 0:15:31.320 major companies. Even though TSMC has increased their numbers, the 0:15:31.720 --> 0:15:34.360 rest of the industry is still a lot more cautious 0:15:34.640 --> 0:15:36.360 around spending. So that's going to be one of the 0:15:36.440 --> 0:15:39.760 key things. How much of the tools are still making 0:15:39.800 --> 0:15:42.440 their way into China. That is also becoming something that 0:15:43.040 --> 0:15:47.280 governments are watching more closely. And then how long can 0:15:47.320 --> 0:15:50.600 we continue to sustain the growth that is needed to 0:15:50.600 --> 0:15:52.080 believe it there billion dollars. 0:15:52.200 --> 0:16:01.640 Mario Morales, this is Bloomberg. We bring in our guest, 0:16:01.760 --> 0:16:05.040 Dan Newman. He is the CEO of the Futuram Group, 0:16:05.160 --> 0:16:08.320 joining us from Austin, Texas. Dan, pleasure to have you 0:16:08.360 --> 0:16:10.440 on the program. I hope you had some time to 0:16:10.480 --> 0:16:13.400 go through what we heard from Nvidia tonight. The stock, 0:16:13.440 --> 0:16:15.360 as I mentioned, was down more than seven percent in 0:16:15.360 --> 0:16:17.920 the late US trading. Do you think this is an overreaction? 0:16:19.560 --> 0:16:21.520 Yeah, thanks for having me, and I have spent some 0:16:21.560 --> 0:16:24.960 time and actually listened to a really compelling interview that 0:16:25.040 --> 0:16:29.280 was done Bloomberg with Ed Ludlow. I think the overall 0:16:29.600 --> 0:16:32.240 reaction is a little bit of what I would have 0:16:32.320 --> 0:16:36.040 expected based on them just doing enough. And as crazy 0:16:36.080 --> 0:16:40.080 as this sounds, just doing enough was beating raising and 0:16:40.120 --> 0:16:44.000 being really solid and consistent in what the company is delivering. 0:16:44.120 --> 0:16:46.400 The problem is, over the last several quarters, we've come 0:16:46.440 --> 0:16:50.680 to expect these massive beats, huge revenue upside, these guides 0:16:50.720 --> 0:16:53.760 that look phenomenal, and this sort of endless upgrade cycle 0:16:54.160 --> 0:16:57.000 and this just was good. And you know what I 0:16:57.080 --> 0:17:00.160 said coming into this, it has to be great or 0:17:00.160 --> 0:17:00.800 there's only. 0:17:00.640 --> 0:17:04.080 Downside the price to perfection story. I think we can 0:17:04.119 --> 0:17:06.680 agree on that, right. But are you troubled at all 0:17:06.720 --> 0:17:09.880 by the word that in Nvidia is having problems with 0:17:09.920 --> 0:17:11.560 that new Blackwell chip design. 0:17:13.560 --> 0:17:16.800 I'm not really troubled because it seems he has the 0:17:16.880 --> 0:17:20.320 situation under control. He talked about the update to the mask, 0:17:20.400 --> 0:17:23.000 which was going to improve the yield, which in the 0:17:23.720 --> 0:17:27.000 in the near term means probably an uplift from Hopper. 0:17:27.520 --> 0:17:31.440 These Hopper chips are still very significant, and they're still 0:17:31.440 --> 0:17:35.160 selling in volume. Amazon and Microsoft are still building out 0:17:35.160 --> 0:17:38.320 significant infrastructure with them, so is x and Tesla and others. 0:17:38.560 --> 0:17:40.359 I think where this is going to head, though, is 0:17:40.600 --> 0:17:44.240 when does that ramp, that one year cycle continue to 0:17:44.320 --> 0:17:46.760 drive these triple digit growth rates. And we're starting to 0:17:46.760 --> 0:17:49.400 see that slow down. And I think whether we see 0:17:49.400 --> 0:17:51.480 this ship into the fourth quarter or even if it 0:17:51.520 --> 0:17:54.040 falls into the first quarter, what we end up having 0:17:54.080 --> 0:17:56.720 happen is we've got an entire world that's kind of 0:17:56.760 --> 0:17:59.680 watching this stock and saying triple digit revenue growth, triple 0:17:59.680 --> 0:18:01.600 digit earnings growth, and now we're starting to see that 0:18:01.680 --> 0:18:05.000 quarter on quarter you're on youar it's decelerating and then 0:18:05.080 --> 0:18:09.040 bigger questions about AI digestion and when does this start 0:18:09.080 --> 0:18:11.600 to really find its way into companies like Salesforce, as 0:18:11.680 --> 0:18:14.240 Numbers and others that reported today that didn't get as 0:18:14.280 --> 0:18:14.720 much attention. 0:18:14.920 --> 0:18:17.000 Yeah, you're right about that. It's a good point too. 0:18:17.040 --> 0:18:20.560 But let's get back to Nvidia and the approved fifty 0:18:20.600 --> 0:18:24.160 billion dollars in additional stock buyback. Is that a good 0:18:24.200 --> 0:18:27.960 way to kind of repurpose or redeploy capital. 0:18:29.359 --> 0:18:31.760 I think if the company is looking at the situation 0:18:31.840 --> 0:18:35.000 as hey, there may be some some hesitancy, there may 0:18:35.040 --> 0:18:37.399 be some sellers. We saw some selling pressure into the 0:18:37.480 --> 0:18:40.240 numbers today. The company wants to be armed, to be 0:18:40.280 --> 0:18:43.080 able to lower the flow and to be able to 0:18:43.200 --> 0:18:46.000 take advantage of a lower share price as it expects 0:18:46.040 --> 0:18:49.400 that this Blackwell and then followed by Rubin will continue 0:18:49.400 --> 0:18:52.240 to drive this acceleration. You know, we did hear from 0:18:52.359 --> 0:18:56.200 Mark Luckerberg center of a Chai, you know, Satia Nadella. 0:18:56.320 --> 0:18:58.840 There was really no indication that they plan to slow 0:18:58.880 --> 0:19:03.159 their capex spending. And that means good things for Nvidia 0:19:03.280 --> 0:19:05.280 because despite the fact that all of them are making 0:19:05.320 --> 0:19:09.440 investments or have been making investments in somewhat competitive offerings. 0:19:09.480 --> 0:19:13.040 In the current moment, these companies are building the bridges 0:19:13.119 --> 0:19:16.160 and roads and tunnels with Nvidio, and if they want 0:19:16.200 --> 0:19:18.520 to be competitive, they don't want their Blockbuster moment, they 0:19:18.520 --> 0:19:20.960 don't want to have their BlackBerry moments. So I think 0:19:21.000 --> 0:19:22.919 they are going to keep investing. And I think that 0:19:22.960 --> 0:19:25.160 bude really well for Justin Wong in video. 0:19:25.400 --> 0:19:27.040 So I hear what you're saying when it comes to 0:19:27.040 --> 0:19:30.880 the cloud computing companies, those so called hyperscalers talking about 0:19:31.320 --> 0:19:34.600 their intention to continue with the current level of cap x. 0:19:34.920 --> 0:19:38.840 But won't that largely be determined by whether or they 0:19:39.440 --> 0:19:42.320 or not they get a significant return on their investment. 0:19:42.359 --> 0:19:46.880 The so called ROI on AHI has to be the key, right. 0:19:48.160 --> 0:19:50.520 You would think so. But if you listen to Mark 0:19:50.600 --> 0:19:54.000 Zuckerberg's comments, he was pretty clear that he was willing 0:19:54.080 --> 0:19:57.160 to bet even if that ingestion period, even if that 0:19:57.600 --> 0:20:00.760 consumption period is longer than we expect. And we know 0:20:00.840 --> 0:20:02.960 that he's done that before with Reality Labs, and the 0:20:03.000 --> 0:20:05.280 market sort of pushed back at him until he had 0:20:05.320 --> 0:20:08.119 to pivot on his as capex investments on that side. 0:20:08.200 --> 0:20:11.600 But AI is not virtual reality. It is very clear 0:20:11.640 --> 0:20:15.040 that this is going to change every industry from autonomous 0:20:15.119 --> 0:20:19.119 driving to drug discovery to these large language models that 0:20:19.119 --> 0:20:21.760 are going to change software and the way we sell, 0:20:21.800 --> 0:20:24.080 the way we buy, the way we watch. And so 0:20:24.160 --> 0:20:26.640 when you put all these things together, I think these 0:20:26.640 --> 0:20:29.280 companies are willing. They have great balance sheets, most of 0:20:29.320 --> 0:20:32.240 them are strong cash positions. They're willing to make the bet. 0:20:32.520 --> 0:20:35.119 But I agree with you, I think we need to 0:20:35.160 --> 0:20:39.840 see more measurable returns downstream from these software companies. Who 0:20:39.880 --> 0:20:43.360 else is making money besides Nvidia and these cloud companies 0:20:43.359 --> 0:20:45.639 have not been really clear. I think Azure amy Hood 0:20:45.640 --> 0:20:47.639 gave the best numbers so far, but you're talking to 0:20:47.680 --> 0:20:49.840 single digit percent about Azure revenue. 0:20:50.160 --> 0:20:54.360 TSMC is obviously the foundry that produces these chips. I mean, 0:20:54.760 --> 0:20:57.280 are you confident that TSMC can keep up with a 0:20:57.359 --> 0:20:58.760 level of production required. 0:21:00.119 --> 0:21:03.880 Well, TSMC is continuing to talk to how it will 0:21:03.880 --> 0:21:07.480 expand its capabilities. We've heard about some of the you know, 0:21:07.560 --> 0:21:11.320 the availability of its three danimeter being really sold out 0:21:11.359 --> 0:21:14.719 through twenty twenty five. We of course they're hearing about coosts, 0:21:14.760 --> 0:21:17.840 you know, the packaging technology, and that was part of 0:21:17.920 --> 0:21:20.480 what was going on with some of the potential delays. 0:21:20.800 --> 0:21:24.359 This new cost l technology and TSMC has the market. 0:21:24.400 --> 0:21:27.919 They have a huge percentage of this AI movement, and 0:21:27.960 --> 0:21:30.439 of course a lot of Nvidia's business, and they're going 0:21:30.520 --> 0:21:33.880 to have to invest substantially. TSMC is a clear winner here, 0:21:34.359 --> 0:21:36.040 and you know, while we are waiting on some of 0:21:36.080 --> 0:21:38.800 these others, the sks and microns, and they're building the 0:21:38.920 --> 0:21:42.080 HBM side this partner ecosystem, and Jenson said this on 0:21:42.119 --> 0:21:44.040 the interview with that ed it was very clear that 0:21:44.119 --> 0:21:47.720 this supply partnership is monumental in terms of the company's 0:21:47.720 --> 0:21:50.720 ability to continue to hit its growth. TSMC has given 0:21:50.760 --> 0:21:54.160 us no reason to doubt its ability. But this technology 0:21:54.200 --> 0:21:56.199 keeps getting more complex and that puts a lot of 0:21:56.240 --> 0:21:58.760 stress on TSMC to deliver well. 0:21:58.800 --> 0:22:02.920 You mentioned the high band with memory chip manufacturers sk Heinix, 0:22:03.119 --> 0:22:07.080 Samsung in South Korea. Are they kind of the weak 0:22:07.160 --> 0:22:09.119 link in this chain, which is to say, if they 0:22:09.160 --> 0:22:11.600 don't keep up with the level of output that in 0:22:12.119 --> 0:22:15.560 Nvidia is striving for that the overall move in the 0:22:15.560 --> 0:22:20.119 build out of some of these AI infrastructure models, I 0:22:20.119 --> 0:22:21.480