WEBVTT - Tech Left Behind in S&P 500’s Latest Rebound 0:00:02.520 --> 0:00:12.879 Bloomberg Audio Studios, Podcasts, radio news. Bloomberg Tech is live 0:00:12.920 --> 0:00:16.760 from coast to coast with Caroline Hide in New York 0:00:17.040 --> 0:00:20.160 and v Lovelow in San Francisco. 0:00:21.079 --> 0:00:23.840 This is Bloomberg Tech coming up. Big Tech lags as 0:00:23.840 --> 0:00:27.200 investors remain skittish about the AI trade. We dig into 0:00:27.360 --> 0:00:31.280 anxieties and losses coming from the magnificent seven plus earnings 0:00:31.280 --> 0:00:33.720 that still roll in. We speak with Octo CEO Todd 0:00:33.760 --> 0:00:36.600 McKinnon after the software company boosted its full year profit 0:00:36.640 --> 0:00:40.519 Foradcast and TikTok. We'll invest more than thirty seven billion 0:00:40.600 --> 0:00:43.360 dollars to build a data center in Brazil, this first 0:00:43.400 --> 0:00:46.199 project in Latin America. But first we check in on 0:00:46.240 --> 0:00:48.680 these markets that are volatile on this day. Look, we're 0:00:48.720 --> 0:00:50.720 putting back on some of the gains made yesterday on 0:00:50.760 --> 0:00:53.200 the Nastaq one hundred. We're off only by a tenth 0:00:53.200 --> 0:00:56.400 of a percent in video flip flopping between gains and losses. 0:00:56.560 --> 0:00:58.880 But all eyes on some of the key magnificent seven 0:00:58.960 --> 0:01:02.480 names that like magnificent. On the day, Bitcoin so managing 0:01:02.480 --> 0:01:04.959 to power up nine ten percent. We're at ninety two thousand, 0:01:05.040 --> 0:01:07.120 so bit of dip bin and going on a sentiment 0:01:07.200 --> 0:01:10.360 though still remains pretty sour on the asset class. Move 0:01:10.360 --> 0:01:12.320 on and have a little look at what's really under 0:01:12.360 --> 0:01:16.520 fired today. Reports out there that Microsoft potentially having to 0:01:16.520 --> 0:01:19.279 tay you to its salespeople of co pilot of AI 0:01:19.360 --> 0:01:22.240 productivity tools. Maybe you need to pull back on your 0:01:22.280 --> 0:01:25.320 sales targets. There is counter reporting coming from the NBC 0:01:25.440 --> 0:01:27.600 that this isn't currently true, but we're going to dig 0:01:27.640 --> 0:01:30.119 into those details for you. Currently off by one point 0:01:30.120 --> 0:01:32.840 eight percent. All of this speaks to the anxiety that 0:01:32.959 --> 0:01:35.520 AI isn't going to give us the productivity gains many 0:01:35.560 --> 0:01:38.280 and Hope Bailey Nipshew's sent piece to say is here 0:01:38.319 --> 0:01:41.360 with us at the moment, really talking about more broadly 0:01:41.480 --> 0:01:44.160 what we're seeing in this market at the moment. And Bailey, 0:01:44.640 --> 0:01:46.920 are we seeing those sorts of anxieties. 0:01:46.480 --> 0:01:48.200 Well, we are seeing some of the uncertainties that we're 0:01:48.200 --> 0:01:50.200 seeing that play out in real time. The big debate 0:01:50.480 --> 0:01:52.440 really around a lot of the CI trade, and we 0:01:52.480 --> 0:01:55.600 saw this play out throughout November, is how much of. 0:01:55.520 --> 0:01:57.120 This spend is going to be realized? 0:01:57.200 --> 0:01:59.040 Are the dollars going to make sense at the end 0:01:59.040 --> 0:02:01.400 of the day, And if you're seeing companies dial back 0:02:01.440 --> 0:02:03.840 there spending. What does that ultimately mean for the bottom 0:02:03.840 --> 0:02:05.600 lines When you see a report like we've been seeing, 0:02:05.640 --> 0:02:08.320 whether they're conflicting or not with Microsoft, it's a lot 0:02:08.320 --> 0:02:09.280 for investors to chew on. 0:02:09.680 --> 0:02:11.840 And what's been being chewed on is in people who 0:02:11.840 --> 0:02:14.480 have diversified since the bottom in October, since we really 0:02:14.520 --> 0:02:17.760 saw the perhaps end of the significant setting. Tech still 0:02:17.840 --> 0:02:20.720 lagged and we've seen underperformed health, tach, care and other 0:02:20.919 --> 0:02:21.560 key sectors. 0:02:21.720 --> 0:02:23.160 Yeah, if you look so last time, we had an 0:02:23.200 --> 0:02:27.000 all time higher looking back to October twenty eighth best 0:02:27.040 --> 0:02:30.800 performers healthcare, communications, energy. As you mentioned, worst performing group 0:02:30.840 --> 0:02:32.560 is tech. One thing to keep in mind when you 0:02:32.600 --> 0:02:34.960 look at that SMP five hundred info Tech. 0:02:35.040 --> 0:02:36.680 The big names Nvidia, Apple. 0:02:36.440 --> 0:02:40.600 Microsoft are in there, missing Tesla, Alphabet and other names. 0:02:40.600 --> 0:02:42.080 So a little bit of a grain of salt to 0:02:42.080 --> 0:02:45.200 take it with that, But we are certainly seeing investors diversifying, 0:02:45.240 --> 0:02:49.520 flocking to healthcare, Eli, Lilly, Cardinal Health, Biogen all out performing. 0:02:49.520 --> 0:02:51.840 When I talk to portfolio managers, part of the debate 0:02:51.919 --> 0:02:53.920 is we've come so far, so fast in a. 0:02:53.919 --> 0:02:54.519 Lot of these names. 0:02:54.560 --> 0:02:56.480 Yes, and as much as we like to think that 0:02:56.480 --> 0:02:59.840 things only go up eventually, the question is can you 0:03:00.160 --> 0:03:02.720 performing is their alpha generation, and that's why we see some. 0:03:02.680 --> 0:03:03.519 Of that repositioning. 0:03:03.840 --> 0:03:07.200 Repositioning still significant gains on the year if you're thinking 0:03:07.200 --> 0:03:09.880 of an Nvidia, but the momentum traits we've seen on 0:03:09.919 --> 0:03:12.640 the back of this, the fact that quantum stocks have 0:03:12.680 --> 0:03:15.320 done so well, some of the crypto spheres, and they 0:03:15.400 --> 0:03:16.760 still under a lot of pressure right now. 0:03:16.800 --> 0:03:18.840 They're under pressure. There's just a lot of volatility. 0:03:18.880 --> 0:03:20.880 You look at the moves that we saw, whether it 0:03:20.919 --> 0:03:23.800 was a nuclear power or as you mentioned in quantum, 0:03:24.320 --> 0:03:26.880 pretty much NonStop up into the right, whether that was retail, 0:03:26.919 --> 0:03:29.200 whether that was momentum or just hedge funds trying to 0:03:29.240 --> 0:03:32.240 juice their books. We've seen them pull back pretty markedly. Obviously, 0:03:32.280 --> 0:03:34.920 Bitcoin has been quite volatile trying to recoup some of 0:03:34.920 --> 0:03:38.520 those losses. But if you're really kind of macro investor 0:03:38.920 --> 0:03:41.520 or running a book broadly from the equity side, we're 0:03:41.560 --> 0:03:44.119 in December, if you're up quite a bit, maybe take 0:03:44.160 --> 0:03:47.120 some gains off the table and maybe diversify into something 0:03:47.160 --> 0:03:49.720 like a healthcare which could be positioned to benefit or 0:03:49.720 --> 0:03:51.080 at least continue growing. 0:03:51.120 --> 0:03:53.119 If there are concerns around AI. 0:03:53.560 --> 0:03:55.840 And concerns as a macrob of picture. We just saw 0:03:55.840 --> 0:03:58.400 the ADP figures today, I mean jobs market looking a 0:03:58.440 --> 0:04:00.560 little bit weaker as well being lipstials. It's always a 0:04:00.640 --> 0:04:02.240 joy to have him on talking to us through the markets. 0:04:02.240 --> 0:04:04.840 We've got more market context for you now and more 0:04:04.880 --> 0:04:09.120 broadly with Sylvia Deablonski Defiance ETF CEO CIO you've been 0:04:09.160 --> 0:04:11.560 writing it's not all all over, but the pace of 0:04:11.600 --> 0:04:15.200 spending is becoming more selective if upcoming macro data corporates, 0:04:15.600 --> 0:04:19.040 we could be nearing a tradable bottom. As a longer 0:04:19.120 --> 0:04:22.840 term investor, I continue to see these pullbacks as opportunities. 0:04:23.680 --> 0:04:28.159 She joins us. Now so cooperating here, is market sentiment 0:04:28.360 --> 0:04:30.880 not or is market sentiment going to shift to the 0:04:30.920 --> 0:04:31.599 positive again? 0:04:33.200 --> 0:04:33.800 Good warning? 0:04:33.839 --> 0:04:38.080 Well, you know, barring external factors, which we've seen a 0:04:38.160 --> 0:04:40.560 lot of right throughout the last year, whether it's been 0:04:40.640 --> 0:04:43.160 geo politics, whether it's been tariffs or for things like this, 0:04:43.400 --> 0:04:45.560 you know, if we're on pace here for kind of 0:04:45.600 --> 0:04:47.919 the trend that we have here, and we do have 0:04:48.040 --> 0:04:49.960 this growth of AI, you know, I know there's news 0:04:50.000 --> 0:04:52.520 around spending being cut back and things like this, but 0:04:52.520 --> 0:04:54.920 we're still in the infancy of AI and there has 0:04:55.000 --> 0:04:57.200 already been so much spending that is now going to 0:04:57.279 --> 0:05:00.440 kind of pile into the economy, pile into do these 0:05:00.440 --> 0:05:02.760 different companies, different sectors and things like this. So I 0:05:02.760 --> 0:05:06.080 think it plays out and benefits these these companies over time. 0:05:06.480 --> 0:05:09.159 So I think, you know that coupled with a FED 0:05:09.320 --> 0:05:12.479 that is likely to cut particularly if we have a 0:05:12.520 --> 0:05:15.880 new FED chair in, you know, in with the president 0:05:15.920 --> 0:05:19.240 and kind of you know, knowing where the predictions will 0:05:19.240 --> 0:05:21.600 go on. That that's what the market's really looking for. 0:05:21.720 --> 0:05:23.760 Right we have a little bit of a weaker labor number, 0:05:24.040 --> 0:05:27.760 Consumer spending is holding up. You know, corporate earnings were 0:05:27.960 --> 0:05:30.400 arguably great. You had over eighty percent of names beating 0:05:30.400 --> 0:05:32.600 on the top and bottom line. You know, double digit 0:05:32.760 --> 0:05:35.800 growth continues, positive GDP outlook. I think that we could 0:05:35.800 --> 0:05:38.760 be in for a good market here, a. 0:05:38.760 --> 0:05:40.839 Good market led by the same names. It's interesting that 0:05:40.839 --> 0:05:43.520 Innvidia CEO Jensen Wang is in Capital Hill today. We 0:05:43.600 --> 0:05:46.200 understand he's just been seen entering US House speak at 0:05:46.240 --> 0:05:51.160 Mike Johnson's office. How much will political tailwinds or headwinds 0:05:51.440 --> 0:05:53.360 help will hindle some of these companies be worried about 0:05:53.400 --> 0:05:56.240 access to China, well broadly, Cannon Vinie continue to lead 0:05:56.279 --> 0:05:57.080 this market higher. 0:05:58.480 --> 0:06:01.440 So I think in a way, and Video still has 0:06:01.480 --> 0:06:05.240 the monopoly on AI and chips right, and so I 0:06:05.279 --> 0:06:07.880 think that you know, a lot of the other semiconductor 0:06:07.920 --> 0:06:11.480 companies are starting to do you know better and take 0:06:11.520 --> 0:06:13.280 some of the business away, but for now, and Video 0:06:13.360 --> 0:06:16.240 remains the clear leader. In terms of Washington and politics. 0:06:16.279 --> 0:06:19.000 Of course, all of these things impact these companies. You know, 0:06:19.040 --> 0:06:21.720 they are the trillion dollar clubs. They lead the market 0:06:21.800 --> 0:06:24.359 and policy effects them first. But the other side of 0:06:24.360 --> 0:06:26.719 that argument is when the market falls back, you know 0:06:26.800 --> 0:06:29.000 where do people invest. A lot of times they look 0:06:29.040 --> 0:06:31.719 for defensive names, but you have these high again trillion 0:06:31.760 --> 0:06:34.880 dollar companies, high quality, strong balance sheets that over time 0:06:35.320 --> 0:06:38.320 tend to perform. Investors actually do look to those companies 0:06:38.320 --> 0:06:40.640 for investments. And if you look at today, you know, 0:06:40.720 --> 0:06:43.720 four hundred out of the five hundred names actually participated 0:06:43.360 --> 0:06:46.400 in the bounce off of you know, the last pullback, 0:06:46.440 --> 0:06:49.400 and so you see a little breath expansion there. I do, though, 0:06:49.480 --> 0:06:52.480 think that the AI tech is going to continue leading 0:06:52.480 --> 0:06:53.680 markets for years to come. 0:06:53.960 --> 0:06:58.320 You've actually got a lodge cap etf that is, excluding 0:06:58.560 --> 0:07:01.800 magnificent seven, how is that performed of late and how 0:07:01.839 --> 0:07:03.440 many people are wanting to put money into it? 0:07:04.320 --> 0:07:07.040 So interestingly enough, we've actually seen an increase in flows 0:07:07.040 --> 0:07:09.720 into that ETF. And you know, it may sound like 0:07:09.720 --> 0:07:12.200 a contradiction what I just said, it absolutely isn't. I 0:07:12.200 --> 0:07:16.120 think that the average investor has exposure to MAG seven everywhere, right, 0:07:16.160 --> 0:07:18.760 whether it's in ETFs or they hold the single name stocks. 0:07:19.120 --> 0:07:21.240 And so when you're when the market pulls back and 0:07:21.280 --> 0:07:24.160 you're looking to diversify it and broaden your exposure, we 0:07:24.240 --> 0:07:26.240 take out those names because we're saying, hey, you already 0:07:26.280 --> 0:07:28.840 have them, hang on to them, but participate in the 0:07:29.040 --> 0:07:31.640 in the in the breath and the broadening of the 0:07:31.760 --> 0:07:33.960 rallies in the years to come. And so you know, 0:07:34.120 --> 0:07:37.360 very much like an equal which strategy x MAC excludes 0:07:37.400 --> 0:07:39.520 the max haven't. But but what's you know we think 0:07:39.600 --> 0:07:42.480 very favorable about that is you're still letting the winners win, right, 0:07:42.560 --> 0:07:44.400 so you still have market. 0:07:44.240 --> 0:07:46.080 Caps rallying in that fund. 0:07:46.120 --> 0:07:47.880 And can you know do well over time. 0:07:48.120 --> 0:07:50.640 You really have created a lot of ETFs that have 0:07:51.000 --> 0:07:54.720 been galvanized by this whole broadening out of the Trade 0:07:54.800 --> 0:07:58.559 AI and Power Infrastructure ETF. For example, But I'm also 0:07:58.640 --> 0:08:00.960 interested in the quantum side of things because there's an 0:08:00.960 --> 0:08:04.760 interesting story. Key Nobel Laureate, winner of the Physics Prize, 0:08:04.800 --> 0:08:06.880 has just been talking about how China looks like it's 0:08:07.000 --> 0:08:10.080 really pulling ahead just a nanosecond behind the US in 0:08:10.160 --> 0:08:14.280 terms of quantum compute scale and indeed innovation. How much 0:08:14.320 --> 0:08:16.080 have people be wanting to get into the quantum side 0:08:16.080 --> 0:08:16.560 of the equation? 0:08:17.760 --> 0:08:20.800 Well, I think, you know, what you just mentioned will 0:08:20.880 --> 0:08:23.080 be a big reason, and we'll put a lot of 0:08:23.080 --> 0:08:25.600 focus on the quantum names again for investors, because of 0:08:25.600 --> 0:08:29.160 what happens when we think that, you know, other governments 0:08:29.400 --> 0:08:31.920 are beating us in AI or technology or things like this. 0:08:32.000 --> 0:08:35.319 The US government tends to invest right, and American corporations 0:08:35.360 --> 0:08:38.200 tend to invest and catch up and try to win 0:08:38.240 --> 0:08:40.120 the race, right. So I think that just you know, 0:08:40.160 --> 0:08:42.560 on a basic level, you've seen some good news around quantum, 0:08:42.559 --> 0:08:46.000 whether it's been HSBC, the advanced bond pricing models that 0:08:46.000 --> 0:08:47.720 have come out of it. You know, you've seen some 0:08:47.760 --> 0:08:50.760 good results around and kneeling from d Wave. You know, 0:08:50.800 --> 0:08:52.880 some of the top quantum companies have done quite well 0:08:52.880 --> 0:08:55.160 for investors. But I think you know, as this continues 0:08:55.200 --> 0:08:57.440 and it becomes more commercial, you will see flows coming 0:08:57.480 --> 0:08:59.680 into the space, and we think it's a longer term 0:08:59.720 --> 0:09:01.200 hole for investors. 0:09:01.880 --> 0:09:05.000 What hasn't been a longer term held just in recent 0:09:05.080 --> 0:09:07.240 days has been micro strategy. You have a lot of 0:09:07.280 --> 0:09:10.760 ETFs giving either you long exposure to micro micro strategy 0:09:10.960 --> 0:09:14.480 now called strategy or indeed leverage positioning for it, but 0:09:14.520 --> 0:09:16.920 also to go short as well. So you're kind of 0:09:16.920 --> 0:09:21.000 offering both sides of the pile here. What do you 0:09:21.040 --> 0:09:24.080 make of the fact that maybe Michael Sailor isn't a 0:09:24.120 --> 0:09:27.160 complete hoddler for the longer term, that maybe at some 0:09:27.200 --> 0:09:28.760 point they might have to sell bitcoin. 0:09:29.840 --> 0:09:31.400 Yeah, I mean I think that that's you know, that's 0:09:31.400 --> 0:09:33.839 a business decision for their firm in terms of how 0:09:33.960 --> 0:09:37.160 best to to you know, generate ROI and and be 0:09:37.240 --> 0:09:40.120 profitable over time. But you know what you just mentioned, 0:09:40.320 --> 0:09:42.200 it is true. We have long funds, we have short funds, 0:09:42.200 --> 0:09:45.320 and for us, we're just trying to democratize hedge fund 0:09:45.400 --> 0:09:48.920 like products for investors that you know, are tactical and 0:09:49.200 --> 0:09:52.840 knowledgeable about these markets and look for short term exposure opportunities. 0:09:52.840 --> 0:09:55.559 So those products are simply there. We don't necessarily have 0:09:55.600 --> 0:09:58.000 an opinion on you know, their view of that. We 0:09:58.040 --> 0:09:59.920 want to give them the right tool to do that. 0:10:00.080 --> 0:10:02.319 But I think you know, if Bitcoin rallies, you'll see 0:10:02.360 --> 0:10:04.400 strategy do well. If it doesn't, you'll see it go 0:10:04.520 --> 0:10:06.600 the other way, and perhaps the flows we'll also match 0:10:06.600 --> 0:10:08.480 in terms of long and short leverage funds. 0:10:08.760 --> 0:10:12.200 Slovia jabonskiin talking about what has been quite the volatile 0:10:12.240 --> 0:10:16.520 trade for certain retail of late Defiance ETF CEO CIO, 0:10:16.960 --> 0:10:19.360 it's always great to catch up with you. Coming up, 0:10:19.559 --> 0:10:22.080 we're going to hear from AWS CEO mcgarmond at his 0:10:22.240 --> 0:10:24.800 take on the AI power demand, the race with Nvidia 0:10:25.000 --> 0:10:27.559 so much more of course trainium three. This will be 0:10:27.679 --> 0:10:28.319 a big tech. 0:10:31.440 --> 0:10:34.439 The desire and the hunger out there for more power 0:10:34.480 --> 0:10:38.880 and more compute is almost insatiable. And so the more 0:10:38.960 --> 0:10:42.680 we can take an existing power footprint, an existing set 0:10:42.679 --> 0:10:45.200 of capabilities and bring more and more compute into that 0:10:45.240 --> 0:10:49.920 for customers to build cool applications and cool environments and 0:10:50.000 --> 0:10:52.000 to get value from that, that's we're focused on. 0:10:52.040 --> 0:10:52.199 Then. 0:10:52.240 --> 0:10:54.320 So we're going to be pushing that envelope as fast 0:10:54.360 --> 0:10:56.760 as we as we possibly can to get those new 0:10:56.800 --> 0:10:58.320 and new capabilities out to customers. 0:10:58.480 --> 0:11:00.960 The pitch for Trainium in both the training and infants 0:11:01.040 --> 0:11:04.360 use case is that it's a great deal, you know, 0:11:04.520 --> 0:11:07.680 cost effective performance. At the same time, you went on 0:11:07.720 --> 0:11:10.840 stage and said AWS is quote by far the best 0:11:10.920 --> 0:11:15.320 place to run in video GPUs. How are both if possible? 0:11:15.520 --> 0:11:18.120 Well, I mean that both both are possible because that 0:11:18.240 --> 0:11:22.920 is a great environment to run accelerators and compute in. 0:11:23.360 --> 0:11:26.160 And so we've been working for fifteen plus years with 0:11:26.240 --> 0:11:29.640 the in Nvidia team and Jensen and team to deliver 0:11:29.840 --> 0:11:33.360 outstanding capabilities for our customers. And for when you're running 0:11:33.480 --> 0:11:36.360 a large cluster of Nvidia GPUs, people will tell you 0:11:36.440 --> 0:11:38.800 AWS is the best place you get the best performance, 0:11:38.840 --> 0:11:42.319 the most stable cluster, the best capabilities out there and 0:11:42.520 --> 0:11:44.920 broad scale. And so why folks like OPENINGI and others 0:11:45.120 --> 0:11:48.080 are running in AWS and we have that choice. And 0:11:48.160 --> 0:11:50.000 so for others that want to be able to take 0:11:50.040 --> 0:11:53.000 advantage of Trainium. And there's some uses cases that are 0:11:53.000 --> 0:11:55.240 best for TRAININGUM, there's other use cases where in video 0:11:55.280 --> 0:11:57.240 GPUs are going to be your best option. We want 0:11:57.240 --> 0:11:59.360 to have all of those available and so we think 0:11:59.400 --> 0:12:01.880 that if we can continue to push the envelope on 0:12:01.920 --> 0:12:04.720 what TRAINUM can deliver for customers and make sure that 0:12:04.760 --> 0:12:08.240 we are supporting the latest and greatest from everything that 0:12:08.240 --> 0:12:10.959 the awesome team in Nvidia is delivering. That's going to 0:12:11.000 --> 0:12:12.440 be the best outcome for our customers. 0:12:13.000 --> 0:12:16.400 The plan for AWS is to basically double capacity by 0:12:16.520 --> 0:12:19.319 end of twenty twenty seven to round eight gigawatts, so 0:12:19.440 --> 0:12:22.400 you have a sense of how you apportion that capacity 0:12:22.920 --> 0:12:26.079 in how Silicon and serve a design to Trainium versus 0:12:26.280 --> 0:12:27.360 and video GP is. 0:12:28.200 --> 0:12:29.760 We're just going to keep pushing as fast as we 0:12:29.800 --> 0:12:32.400 can and we'll see where customer demands drives us as 0:12:32.440 --> 0:12:35.920 we go. And as you said, we're massively adding capacity. 0:12:35.960 --> 0:12:38.360 In the last year alone, we've added three point eight 0:12:38.400 --> 0:12:40.960 gigawatts of capacity, and we'll continue to add more and 0:12:41.000 --> 0:12:43.800 more as over the next couple of years, and we'll 0:12:43.880 --> 0:12:46.080 let customer demands drive us a little bit on what 0:12:46.080 --> 0:12:49.120 they're looking for and what they want, and that's what 0:12:49.160 --> 0:12:50.920 we always listen to and that's what we'll continue. 0:12:50.679 --> 0:12:51.080 To listen to. 0:12:51.600 --> 0:12:53.960 The focus with Trainium in the time i've been able 0:12:53.960 --> 0:12:56.560 to interact with you and talk about again not just 0:12:56.600 --> 0:12:59.680 the accelerator, but at the server design level, there's a 0:12:59.679 --> 0:13:02.720 lot of benefits the customer. When does that benefit start 0:13:02.720 --> 0:13:05.880 accrueing to AWS in terms of profitability, Like, if it's 0:13:05.920 --> 0:13:09.640 such a good financial proposition, you must be able soon 0:13:09.720 --> 0:13:11.280 to say we're making a lot of money on this. 0:13:11.440 --> 0:13:11.600 Yeah. 0:13:11.640 --> 0:13:14.080 Well, you're already seeing some of the benefits of crew 0:13:14.080 --> 0:13:16.720 You see things like bedrock growing really really rapidly, and 0:13:16.760 --> 0:13:19.080 you see training empowering that under the covers. And we 0:13:19.120 --> 0:13:23.200 announced today that more than half of all tokens and 0:13:23.280 --> 0:13:26.240 inference done in bedrock are done on TRAININGUM two servers 0:13:26.320 --> 0:13:28.960 under the covers, And so you're already seeing that benefit come. 0:13:29.240 --> 0:13:31.800 You see the models that we're building and Nova and 0:13:31.800 --> 0:13:34.280 Nova too start to get better and better over time 0:13:35.080 --> 0:13:37.520 and be accelerated by Trainum. And so we really think 0:13:37.559 --> 0:13:39.920 that there's a whole bunch of dimensions on which both 0:13:39.960 --> 0:13:43.000 our customers, our partners, and our own products are going 0:13:43.040 --> 0:13:44.679 to get accelerated all from TRAINUM. 0:13:45.000 --> 0:13:47.400 Every time you come onto the program, I always offer 0:13:47.440 --> 0:13:49.600 the audience opportunity to pose a question to you. There's 0:13:49.600 --> 0:13:51.880 a lot of interests in AWS, right, many of your 0:13:52.120 --> 0:13:55.880 customers span global technology. Actually most of the questions were 0:13:55.920 --> 0:13:59.880 about anthropic That wasn't much said on stage. I think 0:14:00.000 --> 0:14:02.920 people are trying to understand what is the benefit and 0:14:03.040 --> 0:14:07.960 advantage AWS office tanthropic while they are ramping trainum through 0:14:07.960 --> 0:14:11.520 Project Raineer, but also ramping their TPU allocations as well. 0:14:12.040 --> 0:14:14.200 Well, Look, our partners are an anthropic. Our partnership with 0:14:14.200 --> 0:14:16.679 them is incredibly strong and it's never been stronger, and 0:14:17.920 --> 0:14:19.920 we do a ton of collaboration with them, and as 0:14:19.920 --> 0:14:22.680 I mentioned through Project rain here, it's a huge collaboration 0:14:22.800 --> 0:14:25.480 there to go build their current generation models and all 0:14:25.520 --> 0:14:28.600 their models run today and launch on day one on 0:14:28.720 --> 0:14:30.920 top of Trainium and on top of AWS. 0:14:31.200 --> 0:14:34.400 AWS CEO Matt Garman along with their own ed Ludlow. 0:14:34.480 --> 0:14:38.720 Let's discuss this growing AI chip race and the big 0:14:38.760 --> 0:14:40.680 tech names. Who's when, who's losing? Joan Feeney is the 0:14:41.120 --> 0:14:43.200 perfect person to ask partner and portfolio manager over it 0:14:43.240 --> 0:14:47.760 advises capital management. So did Amazon manage to solidify its 0:14:47.800 --> 0:14:50.960 prowess when it comes to vertical integration, cloud and ship. 0:14:50.880 --> 0:14:54.880 Offering, Well, they seem to be doing a really excellent 0:14:54.920 --> 0:14:57.360 job of that. And what's interesting about what Match has 0:14:57.440 --> 0:15:00.840 said is they're going to let customer demand drive you know, 0:15:00.960 --> 0:15:04.920 how they build out these different capabilities for AI applications, 0:15:05.280 --> 0:15:09.280 which tells us that they are going to be different 0:15:09.280 --> 0:15:13.960 models purposed to different types of workloads. So everybody's thinking, oh, 0:15:14.000 --> 0:15:16.920 and Vidia is the market dominant player right now. They 0:15:16.960 --> 0:15:20.120 have massive market share, and they're treating it like it's 0:15:20.240 --> 0:15:24.000 one sort of big thing that's homogeneous. And I think 0:15:24.040 --> 0:15:25.600 what people are going to realize over time is that 0:15:25.600 --> 0:15:28.160 they're going to be different types of models for different workloads, 0:15:28.400 --> 0:15:32.160 and different chips will be useful for developing such different 0:15:32.200 --> 0:15:35.400 models and for running those models, for doing the inference. 0:15:35.680 --> 0:15:37.720 So I think there's a lot of room for different 0:15:37.800 --> 0:15:42.280 chip players here, and we should expect more of the 0:15:42.320 --> 0:15:45.480 in house chips like what Amazon does in collaboration with Marvel, 0:15:46.000 --> 0:15:48.960 or what Google does in collaboration with Broadcom, to continue 0:15:49.000 --> 0:15:50.960 to rise in prominence, as we've been seeing in the 0:15:50.960 --> 0:15:51.840 recent news coverage. 0:15:51.880 --> 0:15:54.000 I mean, yeah, let's talk about Marvel out with earnings 0:15:54.000 --> 0:15:57.400 out with also a key purchase of Celestial AI. That's 0:15:57.400 --> 0:15:59.080 all about lasers, and we can talk in lasers a 0:15:59.080 --> 0:16:02.200 little bit more. But are we seeing a six really 0:16:02.240 --> 0:16:05.360 starting to put a concern towards in video? Should they 0:16:05.400 --> 0:16:07.640 be threatened by this custom built model? 0:16:08.640 --> 0:16:11.600 Well, no company likes any kind of competition right, So 0:16:11.840 --> 0:16:14.320 I'm sure in video would just be happier if they 0:16:14.400 --> 0:16:15.000 didn't exist. 0:16:15.000 --> 0:16:17.480 But clearly, look, ASEX are very powerful. 0:16:17.960 --> 0:16:22.320 That stands for applications specific integrated circuit right, So these 0:16:22.840 --> 0:16:26.160 I believe these kind of xbus er TPUs are going 0:16:26.200 --> 0:16:28.000 to be designed and going to be put to use 0:16:28.440 --> 0:16:31.160 for the type of applications in AI for which they're 0:16:31.200 --> 0:16:36.520 best designed. So it's clearly the case that in Vidia 0:16:36.680 --> 0:16:39.840 is going to lose share over time in the big 0:16:39.840 --> 0:16:43.200 world of AI compute. But nobody should be surprised by 0:16:43.240 --> 0:16:45.480 set by that, and the pie is still growing so 0:16:45.520 --> 0:16:48.320 massively they can't meet demand. They're not unlikely to be 0:16:48.400 --> 0:16:50.920 able to meet demand for some years to come, and 0:16:50.960 --> 0:16:53.200 so there's a lot of room for a sick approaches 0:16:53.280 --> 0:16:56.200 to play roles in various types of models. 0:16:56.200 --> 0:16:57.040 So we should expect that. 0:16:57.200 --> 0:16:59.640 Before we talk more lasers, I want to go back 0:16:59.680 --> 0:17:02.120 to where we sort of started, the fact that you're 0:17:02.120 --> 0:17:05.720 getting specific chips for specific applications but also specific models. 0:17:06.000 --> 0:17:07.919 And I think there's some threat in the market today 0:17:07.960 --> 0:17:11.600 around Microsoft's are the part of the flywheel, because everyone 0:17:11.680 --> 0:17:14.399 thinks Microsoft got very strong cloud offering, and clearly we've 0:17:14.440 --> 0:17:17.040 seen as you not be able to fill demand with 0:17:17.119 --> 0:17:20.119 its supply. We're now we're just basically about the applications, 0:17:20.200 --> 0:17:23.280 about companies being willing to pay more for these AI 0:17:23.400 --> 0:17:27.080 agents and actually the use of AI. What are you 0:17:27.119 --> 0:17:29.120 hearing in the market. How much does that impact your 0:17:29.119 --> 0:17:29.919 sector of chips? 0:17:30.880 --> 0:17:35.000 There is a lot of sporadic i should say, information 0:17:35.119 --> 0:17:38.399 about how much AI agents and AI is being used 0:17:38.440 --> 0:17:42.320 in different settings. So, right, there was a story today 0:17:42.320 --> 0:17:46.280 about how Microsoft has reduced its sales targets for getting 0:17:46.359 --> 0:17:49.600 AI agents you know, to be adopted by their customers, 0:17:49.640 --> 0:17:53.120 and so that as people concerned, you know, Salesforce similarly 0:17:53.240 --> 0:17:56.040 is trying to roll out right their AI agent capabilities 0:17:56.080 --> 0:17:58.280 and some people are wondering whether that's going to happen. 0:17:58.520 --> 0:18:01.520 But there are so many other applications in addition to 0:18:01.720 --> 0:18:05.360 those kind of smart helper AI agents, and I think 0:18:05.400 --> 0:18:09.400 the earlier in the week announcement with Nvidia in synopsis 0:18:09.800 --> 0:18:12.280 is a good example of that. Right, the industrial uses, 0:18:12.320 --> 0:18:15.320 whether it's these digital twins basically being able to model 0:18:15.359 --> 0:18:18.280 out a whole factory or a whole building, or you know, 0:18:18.320 --> 0:18:21.920 any other physical presence in simulation. You need this kind 0:18:21.920 --> 0:18:25.320 of computational power. You need AI and so that kind 0:18:25.359 --> 0:18:28.200 of collaboration I think really points to lots of other 0:18:28.240 --> 0:18:31.239 applications for AI, and then you go to biotech and 0:18:31.240 --> 0:18:32.040 pharma and. 0:18:33.480 --> 0:18:34.080 Computation. 0:18:34.400 --> 0:18:37.240 We've just been held back for so many years by 0:18:37.280 --> 0:18:40.080 its limitations from the CPU side, and now that we 0:18:40.119 --> 0:18:43.600 have these GPU based accelerators and innovations in fabric like 0:18:43.640 --> 0:18:47.560 what Celestial is bringing to Marvel and what Broadcom already does, 0:18:48.560 --> 0:18:52.159 is speeding up all of these computations, making farmer applications 0:18:52.160 --> 0:18:53.600 possible across all industries. 0:18:54.000 --> 0:18:56.960 And some would say limitations have been in the fabrication 0:18:57.040 --> 0:19:00.360 in the United States energy just making sure that we've 0:19:00.359 --> 0:19:04.800