WEBVTT - Amazon's TikTok Bid; Plus, AWS and the State of AI 0:00:00.080 --> 0:00:07.040 Bloomberg Audio Studios, Podcasts, radio News. 0:00:08.920 --> 0:00:13.240 This is Bloomberg Business Week Daily reporting from the magazine 0:00:13.240 --> 0:00:17.840 that helps global leaders stay ahead with insight on the people, companies, 0:00:17.880 --> 0:00:22.760 and trends shaping today's complex economy. Plus global business finance 0:00:22.800 --> 0:00:26.120 and tech news as it happens. The Bloomberg Business Week 0:00:26.239 --> 0:00:31.720 Daily Podcast with Carol Masser and Tim Steneveek on Bloomberg Radio. 0:00:32.840 --> 0:00:34.840 Now, want to get to what was a red sticky 0:00:34.880 --> 0:00:37.360 here at Bloomberg about two hours ago. This is when 0:00:37.400 --> 0:00:39.600 a big headline crosses the Bloomberg terminal. 0:00:39.280 --> 0:00:42.240 And that's that Amazon has submitted an offer to purchase TikTok. 0:00:42.640 --> 0:00:44.479 The company submitted a bid to the White House to 0:00:44.520 --> 0:00:46.960 purchase the app from its Chinese owner. This according to 0:00:46.960 --> 0:00:49.840 a perfect person familiar with the matter. The company did 0:00:49.880 --> 0:00:53.120 send its proposal to Vice president Jad Evance. He's heading 0:00:53.159 --> 0:00:56.200 up efforts to help facilitate a sale of the video platform. 0:00:56.280 --> 0:00:59.280 Commerce Secretary Howard Lutnik two. This, according to the people, 0:00:59.360 --> 0:01:01.960 so with what we know and why Amazon might want 0:01:02.000 --> 0:01:04.760 TikTok with us. As Bloomberg News Technology and e commerce 0:01:04.760 --> 0:01:07.759 reporter Spencer Sober, he follows Amazon very closely been reporting 0:01:07.800 --> 0:01:09.720 on this company for years, he joins us of course 0:01:10.040 --> 0:01:13.399 from our Seattle bureau. Spencer, why would Amazon want TikTok? 0:01:13.480 --> 0:01:14.840 What would it do with TikTok? 0:01:17.200 --> 0:01:20.520 So if you think about it in the most like 0:01:20.560 --> 0:01:23.440 the simplest terms, right, people spend a lot of time 0:01:24.120 --> 0:01:28.160 on TikTok and people spend a lot of money on Amazon, 0:01:28.600 --> 0:01:30.560 and both of them have been trying to kind of 0:01:30.640 --> 0:01:34.160 become more more of the other. So if Amazon were 0:01:34.160 --> 0:01:37.640 to gain We're to buy TikTok, it would be very 0:01:37.680 --> 0:01:42.040 beneficial just in terms of having people's attention where you know, 0:01:42.120 --> 0:01:44.160 right now they've got a lot of wallets share, but 0:01:44.560 --> 0:01:46.840 it's really a site where people dip in, buy what 0:01:46.880 --> 0:01:51.160 they need, dip out, which you know, which has been 0:01:51.280 --> 0:01:53.160 very good for Amazon. But if it could get people 0:01:53.200 --> 0:01:56.920 to linger around and be entertained there, all the better, 0:01:56.960 --> 0:01:59.480 you know, people spend even more and its advertising business 0:01:59.480 --> 0:02:01.320 will be will be that much more lucrative. 0:02:02.360 --> 0:02:04.920 You know, I think about kind of the merge of 0:02:05.320 --> 0:02:09.200 commerce right with TikTok. To me, it makes a lot 0:02:09.240 --> 0:02:09.680 of sense. 0:02:09.720 --> 0:02:10.840 Is that a big part of it? 0:02:10.919 --> 0:02:11.359 Spencer? 0:02:12.760 --> 0:02:15.640 Yeah, exactly. You think about like Amazon is a place 0:02:15.680 --> 0:02:17.560 you go when you're ready to buy something right, You 0:02:17.560 --> 0:02:19.840 don't really you don't really go there to kill time. 0:02:19.919 --> 0:02:22.680 You don't really go there to be inspired or discover things. 0:02:23.000 --> 0:02:24.959 But those are things that you do on TikTok. You 0:02:25.040 --> 0:02:27.560 might just kind of scroll through TikTok to kill time 0:02:27.680 --> 0:02:30.560 or follow people you think are funny or entertaining, but 0:02:30.639 --> 0:02:32.560 you also see products on there, you know, And so 0:02:32.639 --> 0:02:37.880 TikTok is really trying to be a destination for discovering products, 0:02:37.880 --> 0:02:41.640 which Amazon never really has been. So that would be 0:02:41.720 --> 0:02:45.320 kind of the you know, the match that Amazon would 0:02:45.320 --> 0:02:45.799 be looking for. 0:02:46.000 --> 0:02:47.800 Not to say though that the company hasn't tried to 0:02:47.840 --> 0:02:50.799 make it a place to discover with the social first 0:02:50.919 --> 0:02:53.440 videos and the option in the app to sort of 0:02:53.480 --> 0:02:56.280 have that content that's generated to try to create this 0:02:56.360 --> 0:02:59.639 social shopping experience and certainly not something that I spent 0:02:59.680 --> 0:03:02.160 a lot of time doing, but Amazon is certainly trying 0:03:02.160 --> 0:03:03.960 to do that. Hey, spencer, before we let you go 0:03:04.040 --> 0:03:09.040 regulatory issues around this, potentially, especially with regard to their 0:03:10.200 --> 0:03:13.440 past contentious relationship that Bezos had with this administration. 0:03:15.160 --> 0:03:18.760 Yeah, well that's what you think about. You know that 0:03:18.800 --> 0:03:21.560 they're under any trust scrutiny, and they you know, there 0:03:21.600 --> 0:03:23.880 was any trust scrutiny about them buying like eye Robot, 0:03:24.040 --> 0:03:27.519 you know, so this would be like a massive, massive acquisition. 0:03:27.720 --> 0:03:30.799 And does that does is there like a new age 0:03:30.800 --> 0:03:34.600 in the Trump administration that would be more favorable to 0:03:34.639 --> 0:03:37.760 these kind of big deals, And especially since the administration 0:03:37.880 --> 0:03:40.720 is directly involved in making this happen with jd Vance 0:03:40.800 --> 0:03:43.040 kind of running running the show and running these bids, 0:03:43.440 --> 0:03:46.960 that that makes it more possible for Amazon to swing 0:03:46.960 --> 0:03:48.240 big on acquisitions. 0:03:49.960 --> 0:03:52.640 One thing I wonder the bid first reported by the 0:03:52.680 --> 0:03:55.200 New York Times. You guys to hear at Bloomberg reporting 0:03:55.200 --> 0:03:58.600 this at that it's not being considered seriously by the administration. 0:03:58.640 --> 0:04:02.960 According to one person in the know, the company has 0:04:03.000 --> 0:04:06.840 declined to comment. How serious Amazon, you know, I mean, 0:04:07.240 --> 0:04:10.360 I'm calling you Amazon because your Amazon spencer. How serious 0:04:10.360 --> 0:04:12.960 spencer should investors be taking this? 0:04:14.360 --> 0:04:18.560 Yeah, that's the million dollar question. And there's a lot 0:04:18.560 --> 0:04:21.240 of eleventh hour bids coming in, like after the news 0:04:21.240 --> 0:04:23.960 broke about Amazon. Now there's a couple others who have 0:04:24.000 --> 0:04:27.400 been reported, like app love in and maybe some consortium 0:04:27.440 --> 0:04:30.880 that includes like OnlyFans. So there's been a lot of 0:04:30.920 --> 0:04:34.560 eleventh hour bids and there can be reasons for companies 0:04:34.560 --> 0:04:36.400 to do that, even if they think they're a long shot, 0:04:36.480 --> 0:04:38.120 you know one, it might just give them kind of 0:04:38.120 --> 0:04:40.640 a seat at the table to figure out what's happening 0:04:40.680 --> 0:04:42.320 and how this thing might be carved up in case 0:04:42.360 --> 0:04:45.440 they end up interesting, you know, competing with TikTok. And 0:04:45.480 --> 0:04:47.359 then another reason is just to try to up the 0:04:47.360 --> 0:04:49.120 price in the event. Hey, if I'm not going to 0:04:49.160 --> 0:04:51.080 get this thing, maybe at least make the person who 0:04:51.080 --> 0:04:51.960 does pay a little more. 0:04:52.080 --> 0:04:54.359 That's what I was thinking, especially if Oracle is a 0:04:54.400 --> 0:04:58.360 serious bidder here, Spencer, maybe this company Amazon just wants 0:04:58.400 --> 0:05:00.640 to say, hey, we want or Oracle is going to 0:05:00.680 --> 0:05:02.880 bid for this, and we're competitor Oracle, we want them 0:05:02.920 --> 0:05:04.120 to pay more than they're going to pay. 0:05:05.160 --> 0:05:07.160 Yeah, that's that's definitely part of the play, as well 0:05:07.200 --> 0:05:10.480 as well as some visibility into into what's happening, because 0:05:10.480 --> 0:05:12.960 this is still a very complicated process. You need you 0:05:13.040 --> 0:05:15.760 need uh, you need to approval, not only a blessing 0:05:15.760 --> 0:05:18.039 from the White House, but you're gonna need Bike Dance 0:05:18.120 --> 0:05:20.120 to buy in and the Chinese government to buy in. 0:05:20.320 --> 0:05:23.560 So there's a lot of players here, and some you know, 0:05:23.720 --> 0:05:26.400 I imagine anyone, you know, if they think they're a 0:05:26.400 --> 0:05:28.120 long shot. They just be curious about what the heck's 0:05:28.160 --> 0:05:30.000 going on, and putting in a putting in a bid 0:05:30.080 --> 0:05:32.240 might give them more visibility than sitting it out. 0:05:32.400 --> 0:05:35.479 But we should know something at least by Friday, right 0:05:35.560 --> 0:05:37.800 April fifth, because that's the deadline, is that right, third? 0:05:37.839 --> 0:05:38.039 Fourth? 0:05:38.120 --> 0:05:38.360 Fifth? 0:05:38.400 --> 0:05:39.200 Actually Saturday? 0:05:39.240 --> 0:05:40.000 I think it is, right. 0:05:40.320 --> 0:05:42.680 So sometimes you just got up the calendar, open out. 0:05:42.720 --> 0:05:45.400 And that's the and that's the second that's the second deadline. 0:05:45.440 --> 0:05:47.000 So he knows if we'll have another deadline, you know 0:05:47.000 --> 0:05:49.040 what I mean. It's like these these deadlines can come 0:05:49.080 --> 0:05:50.920 and go. But but yeah, that's the that's the date 0:05:50.920 --> 0:05:52.640 for now, and supposedly there might be some news coming 0:05:52.680 --> 0:05:53.240 today about it. 0:05:53.320 --> 0:05:55.000 Yeah, we understand that the president is meeting with his 0:05:55.080 --> 0:05:58.680 advisors about it, right. Thank you so appreciate it. Spencer 0:05:58.800 --> 0:06:01.680 soper As to mention, been following Amazon for so many years, 0:06:01.760 --> 0:06:05.320 understands this company's Bloomberg News Technology and e commerce reporter. 0:06:05.640 --> 0:06:08.400 Joining us out there in Seattle. 0:06:08.200 --> 0:06:11.800 You're listening to the Bloomberg Business Week podcast. Catch us 0:06:11.839 --> 0:06:15.320 live weekday afternoons from two to five pm Eastern Listen 0:06:15.320 --> 0:06:18.880 on Applecarplay and Android Auto with the Bloomberg Business app 0:06:19.040 --> 0:06:21.679 or watch us live on YouTube. 0:06:23.160 --> 0:06:25.400 Okay, so everything we've done for the last couple of 0:06:25.440 --> 0:06:27.800 weeks has certainly been through the lens on the backdrop 0:06:27.880 --> 0:06:30.240 of US tariffs. It's something that touches all of the 0:06:30.279 --> 0:06:35.920 investment in business worlds, including potentially the AI investment world. 0:06:36.160 --> 0:06:39.159 And today at Bloomberg headquarters, our Bloomberg Intelligence team is 0:06:39.240 --> 0:06:42.680 hosting Bloomberg's fourth Generitive AI Conference, and it's an event 0:06:42.720 --> 0:06:46.039 that looks at Jenai posts the Deepseek news earlier this 0:06:46.160 --> 0:06:48.880 year and Carol as a reminder for all of us 0:06:48.920 --> 0:06:51.200 that really shook up the narrative around the AI build out. 0:06:51.400 --> 0:06:54.039 Absolutely, we are featuring a couple of speakers from the event, 0:06:54.080 --> 0:06:56.000 and that includes our next guest, delighted to have with us, 0:06:56.080 --> 0:07:00.599 Dave Brown, vice president AWS Computing and Networking at Amazon Services, 0:07:00.640 --> 0:07:03.400 joining us here in studio. Dave, good to have you here. 0:07:03.520 --> 0:07:04.960 It's great to be here. Thank you for having me. 0:07:05.080 --> 0:07:07.520 There's so much going on there is I do need 0:07:07.560 --> 0:07:09.320 to ask you. We know this isn't your world. Nothing 0:07:09.360 --> 0:07:11.680 happens in a bubble. You were just listening to Spencer. 0:07:12.000 --> 0:07:14.760 I thought I saw you chuckle at one point. The 0:07:14.840 --> 0:07:17.640 report about Amazon putting in a last minute bit. You know, 0:07:17.800 --> 0:07:20.840 I think about AI, it is about content, it is 0:07:20.880 --> 0:07:23.560 about data, and I just wonder how might a social 0:07:23.600 --> 0:07:26.840 media site like TikTok makes sense in an AI world, 0:07:27.000 --> 0:07:28.800 you know, whether it's at Amazon or somewhere else. 0:07:29.240 --> 0:07:31.800 Well, you know, in terms of Amazon's bid, that smells 0:07:31.800 --> 0:07:35.520 something I'm able to comment about. Obviously, no madmoster corporate 0:07:35.560 --> 0:07:36.880 about this, nothing at all. 0:07:36.960 --> 0:07:38.720 I walked into the studio and I saw the news. 0:07:38.760 --> 0:07:42.520 So but you know, Bikedons has been a really good 0:07:42.520 --> 0:07:46.200 customer of AWS and honestly a customer that's actually used 0:07:46.200 --> 0:07:47.680 a lot of compute and he's now using a lot 0:07:47.680 --> 0:07:49.440 of AI. And one of the things very excited about 0:07:49.440 --> 0:07:51.000 what they've been doing, and unrelated to any of the 0:07:51.040 --> 0:07:53.160 news that you might be reporting, is they've used a 0:07:53.200 --> 0:07:55.840 lot of the customer accelerators that we've that we've built 0:07:55.840 --> 0:07:58.720 at AWS. One of the challenges with AI, and you know, 0:07:58.840 --> 0:08:00.560 is how do we get the cost down? How do 0:08:00.600 --> 0:08:03.200 we make sure that it is affordable for every company 0:08:03.200 --> 0:08:04.680 out there that thinks they could use it in some 0:08:04.760 --> 0:08:07.160 way that would change their product. But really we have 0:08:07.200 --> 0:08:08.600 to be able to get the cost down. And one 0:08:08.600 --> 0:08:10.320 of the ways we can do that is through things 0:08:10.320 --> 0:08:12.440 like you saw with deep seek, improving the models, making 0:08:12.440 --> 0:08:14.680 the model simpler than making them easy to train. But 0:08:14.720 --> 0:08:17.720 the other way is to build different silicon, build silicon 0:08:17.760 --> 0:08:19.520 that's able to get the cost down for customers. And 0:08:19.560 --> 0:08:21.600 so Biketones has been a big user of our custom 0:08:21.680 --> 0:08:24.160 silicon in the very early days of Inferential which improved 0:08:24.160 --> 0:08:26.920 inference models. They've also consumed Graviton, which is our customer 0:08:26.920 --> 0:08:29.400 on chip, and they also using our Trainium chips as well, 0:08:29.440 --> 0:08:30.840 So we work very closely with Byketons. 0:08:30.880 --> 0:08:32.360 From that point of view, how do you make sure 0:08:32.640 --> 0:08:34.960 that you keep a company such as bytans as a 0:08:34.960 --> 0:08:37.720 customer when there is so much competition in this space? 0:08:37.760 --> 0:08:38.640 I mean, think about it from the. 0:08:38.640 --> 0:08:41.960 Perspective of the three biggest players you of course Microsoft, 0:08:42.040 --> 0:08:44.880 Azure and then Google Cloud and we've seen a big investment. 0:08:44.880 --> 0:08:47.480 We'll talk about that for Google just in the last 0:08:47.480 --> 0:08:49.760 couple of weeks. How do you make sure that you 0:08:49.840 --> 0:08:52.160 stay ahead in your view of the pack and keep 0:08:52.200 --> 0:08:53.959 these customers from jumping ship. 0:08:54.200 --> 0:08:55.600 I think there are a few things. I think one 0:08:55.640 --> 0:08:57.760 of it is innovation. You know, the cloud has just 0:08:57.800 --> 0:09:00.760 brought an enormous amount of innovation in to the corporate world. 0:09:00.760 --> 0:09:02.880 But some people might look at this and say, is 0:09:02.920 --> 0:09:04.199 this a commodity at this point? 0:09:04.600 --> 0:09:07.600 I don't think so, because my next point would be execution. 0:09:08.120 --> 0:09:11.760 An execution really rarely matters. It's one thing to say 0:09:11.760 --> 0:09:12.400 we have a hard. 0:09:12.559 --> 0:09:13.240 Mean I feel like no. 0:09:13.480 --> 0:09:16.560 For yeah, absolutely executive execution all the time. 0:09:16.559 --> 0:09:17.760 What does it mean in your world? 0:09:17.800 --> 0:09:20.080 Well, what it does mean is, you know, bring running AI. 0:09:20.160 --> 0:09:22.480 For example, so the servers behind AI, whether you're running 0:09:22.480 --> 0:09:25.240 an Invertio GPU or you're running some customs Silicon such 0:09:25.240 --> 0:09:27.640 as our Tranium chip, they're not easy machines to run. 0:09:27.960 --> 0:09:31.240 They're incredibly complicated, they're incredibly hard to make sure that 0:09:31.240 --> 0:09:36.359 they're highly available, and they also need incredibly advanced networking capability. 0:09:36.440 --> 0:09:38.839 So we've had to build network capabilities to be able 0:09:38.880 --> 0:09:42.040 to bring hundreds of thousands of these accelerators together in 0:09:42.080 --> 0:09:45.400 a single cluster. Now at AWS, we spend a lot 0:09:45.440 --> 0:09:47.840 of time making sure that we provide those clusters and 0:09:47.920 --> 0:09:51.600 those those systems to be as highly available as possible 0:09:51.600 --> 0:09:54.040 for our end customer. And so when you choose your technology, 0:09:54.040 --> 0:09:56.560 if you're running on a you know, a solution, whether 0:09:56.559 --> 0:09:58.800 it's something you've built yourself or a provider, and that's 0:09:58.800 --> 0:10:01.000 not working for you. That's what I mean by execution. 0:10:01.320 --> 0:10:03.120 It has to be flawless and it has to work, 0:10:03.240 --> 0:10:04.920 and at AWS we think that's one of the ways 0:10:04.960 --> 0:10:06.520 we're able to differentiate ourselves. 0:10:06.559 --> 0:10:09.400 Are you at all worried about overbuilding? Like obviously with 0:10:09.440 --> 0:10:11.240 Deep Seek, the narrative all of a sudden, like in 0:10:11.280 --> 0:10:13.800 a heartbeat, switched in terms of, well, maybe we can 0:10:13.840 --> 0:10:16.559 do this more cheaply, maybe it doesn't require the buildup 0:10:16.559 --> 0:10:18.439 that we've been talking about for more than two years. 0:10:18.920 --> 0:10:22.319 Are you worried about overbuilding? Are we possibly in a 0:10:22.360 --> 0:10:23.160 bubble when it comes. 0:10:23.000 --> 0:10:26.480 To well, I think from our point of view, Sydney, Well, 0:10:26.480 --> 0:10:28.800 from our point of view, we're not really listening to 0:10:28.880 --> 0:10:30.760 sort of the narrative. I think the narrative is very interesting. 0:10:30.800 --> 0:10:33.120 We've been following it. What we're listening to is customers 0:10:33.400 --> 0:10:36.280 and what a customers want to do with AI. There 0:10:36.360 --> 0:10:38.920 certainly no shortage of customer demand for the chips right, 0:10:38.960 --> 0:10:42.920 whether it's in video chips so uncustomed silicon you've seen partnerships. 0:10:42.520 --> 0:10:45.400 No doubt about that demand. And the demand is smart 0:10:45.440 --> 0:10:47.040 because it's being put to use and it's going to 0:10:47.040 --> 0:10:48.640 be monetized. Yeah, no, worry. 0:10:49.160 --> 0:10:50.280 Well, I mean the other thing is you just have 0:10:50.280 --> 0:10:52.959 to you have to make good investments. And so the 0:10:53.200 --> 0:10:55.040 trick is to say, what are we building in terms 0:10:55.080 --> 0:10:57.160 of customer demand? Not are we building in terms of 0:10:57.160 --> 0:10:59.240 the narrative? And I think from a customer demand point 0:10:59.240 --> 0:11:01.920 of view, it's very very strong. We expect it to 0:11:01.920 --> 0:11:04.760 continue to grow, and we're very comfortable with the investments 0:11:04.800 --> 0:11:05.319 that we're making. 0:11:05.440 --> 0:11:06.960 So it brings us back to the way that we 0:11:07.160 --> 0:11:11.280 started this conversation. We're speaking to you on Wednesday, April second, 0:11:11.400 --> 0:11:14.040 ahead of President Trump's announcement that's set to come at 0:11:14.080 --> 0:11:17.160 four pm Wall Street time from Washington, DC, So we 0:11:17.200 --> 0:11:18.920 don't know what he's going to say at this point. 0:11:19.679 --> 0:11:21.720 How do you watch an event such as that? How 0:11:21.760 --> 0:11:23.880 do you watch tariffs? How does that affect your world? 0:11:23.920 --> 0:11:24.280 Yeah? 0:11:24.320 --> 0:11:25.840 I think one of the things when you're running what 0:11:25.920 --> 0:11:29.079 is probably one of the largest supply chain supply chains 0:11:29.120 --> 0:11:30.679 in the world, right, we have to think about every 0:11:30.720 --> 0:11:33.200 single component, whether it costs a few cents or it's 0:11:33.360 --> 0:11:34.120 millions of dollars. 0:11:34.200 --> 0:11:36.040 Within our You're talking about. 0:11:35.800 --> 0:11:38.880 The aw aws, not the retails, not the retail session 0:11:38.880 --> 0:11:39.840 it's certainly effective. 0:11:39.880 --> 0:11:40.680 The retail side is. 0:11:40.640 --> 0:11:42.440 Certainly helpful to us, and it's taught us a lot 0:11:42.440 --> 0:11:44.720 about how to run an effective supplies right and so 0:11:44.760 --> 0:11:46.840 on the AWS side. Running a supply chain, we've got 0:11:46.880 --> 0:11:49.600 to think about every single possible outcome, you know, the 0:11:49.600 --> 0:11:51.280 sort of things that you're talking about now with regards 0:11:51.320 --> 0:11:54.040 to track tariffs or you know, global uncertainty, whatever it 0:11:54.120 --> 0:11:55.360 might be, we're going to think about how they would 0:11:55.360 --> 0:11:58.640 affect our supply chain. And we've put mitigating steps in place, 0:11:58.760 --> 0:12:01.000 and we're thinking about these as we to make sure 0:12:01.000 --> 0:12:03.760 that any outcome means that we can still serve our customers. 0:12:03.800 --> 0:12:06.200 What are some of those mitigating steps that you've put 0:12:06.240 --> 0:12:07.559 in place. There's als of things you can do from 0:12:07.559 --> 0:12:08.880 a supply chain point of view. A lot of the 0:12:08.920 --> 0:12:12.360 sil can manufacture ourselves. We always think about multiple multiple suppliers, 0:12:12.400 --> 0:12:14.440 like are you stockpiling this stuff right now? I don't 0:12:14.440 --> 0:12:16.680 believe we're stockpiling. But there's many many things that we're 0:12:16.679 --> 0:12:18.400 doing to ensure that we're going to be. 0:12:18.400 --> 0:12:21.559 Okay, such as. 0:12:20.400 --> 0:12:23.160 Such as making sure that we have multiple locations for service, 0:12:23.240 --> 0:12:25.400 such and making sure that we have a cost structure 0:12:25.440 --> 0:12:28.560 that we think is maintainable for our end customer. It's 0:12:28.640 --> 0:12:30.200 just it's supply chain technology based. 0:12:30.320 --> 0:12:33.439 Is most of your supply chain safe in terms of 0:12:33.520 --> 0:12:35.000 let's say it's an all out trade war. 0:12:35.120 --> 0:12:37.280 The reality is I think in the IT space that 0:12:37.320 --> 0:12:40.240 we're where depending on global supply chain, not just a 0:12:40.360 --> 0:12:43.040 WS but the world. But our chips are made most 0:12:43.080 --> 0:12:45.400 of them are manufactured in the Far East, most of 0:12:45.440 --> 0:12:47.920 the you know, the technology and cloud comes from the US, 0:12:48.400 --> 0:12:50.839 and the realities we're living in a global supply chain 0:12:50.840 --> 0:12:52.720 and we're going to have to take whatever the announcement 0:12:52.800 --> 0:12:54.839 might be later today and understand how that affects the 0:12:54.840 --> 0:12:57.400 supply chain that we've built out to be as redundant 0:12:57.440 --> 0:12:58.440 as it possibly can be. 0:12:58.840 --> 0:13:00.000 But we're going to have to see what happens. 0:13:00.480 --> 0:13:03.079 Is there a spectrum in terms of the cost hit 0:13:03.120 --> 0:13:05.760 that you think you could take as a result of 0:13:05.800 --> 0:13:06.400 these tariffs. 0:13:06.720 --> 0:13:09.200 That's not something that I think we've been discussing. I mean, 0:13:09.200 --> 0:13:11.360 I think there will be things we'll have to think through, 0:13:11.400 --> 0:13:13.120 but I don't have any sort of spectrum in mind. 0:13:13.160 --> 0:13:15.520 What about thinking about raising prices for customers as a 0:13:15.520 --> 0:13:16.800 result of higher inputs. 0:13:16.520 --> 0:13:18.920 You know, that's that's something I think a similar situation 0:13:18.960 --> 0:13:21.160 would be the sort of the energy prices increases that 0:13:21.200 --> 0:13:23.360 we saw, you know, during the COVID time there was 0:13:23.559 --> 0:13:26.840 energy shortage and slightly off the COVID in Europe, and 0:13:27.080 --> 0:13:28.640 that was a big question back then, as you know, 0:13:28.760 --> 0:13:31.199 is able is going to raise prices because we've we've 0:13:31.200 --> 0:13:34.600 seen energy prices spike in Europe. Other providers did raise 0:13:34.640 --> 0:13:36.560 prices at that time, and from an als point of view, 0:13:36.559 --> 0:13:38.520 we were able to actually not raise prices on our 0:13:38.520 --> 0:13:41.920 customers and ride out that increasing commodity prices and we 0:13:41.920 --> 0:13:43.600 were able to get back to, you know, a better 0:13:43.640 --> 0:13:46.080 cost structure for ourselves after this, and that's really what 0:13:46.080 --> 0:13:47.800 we're looking to do. We'll have to see what happens 0:13:47.840 --> 0:13:50.880 later today and see how supply, how supply chain holds up. 0:13:51.040 --> 0:13:53.400 I'm also curious about deep Seek in terms of, you know, 0:13:53.520 --> 0:13:56.840 what might be a more efficient AI model, how that 0:13:56.960 --> 0:13:59.599 is affecting the design of your own chips at aws 0:13:59.640 --> 0:14:02.800 and kind of how you think about how AWF offers 0:14:02.880 --> 0:14:04.600 compute and compute services and so on. 0:14:04.760 --> 0:14:08.240 Yeah, deep Seek itself, you know, it was good technology. 0:14:08.240 --> 0:14:10.160 It came out. I think it surprised a lot of folks. 0:14:10.559 --> 0:14:14.440 We're expecting to see these kinds of significant improvements step FUNCTIONINGY, 0:14:14.480 --> 0:14:17.160 you weren't shocked by deep seek, Well, certainly the news 0:14:17.200 --> 0:14:19.120 it wasn't something that everybody knew about. But what I 0:14:19.160 --> 0:14:20.840 wasn't just shocked about is that we're going to see 0:14:20.880 --> 0:14:24.080 these these order the magnitude improvements in models. I'm expecting 0:14:24.120 --> 0:14:26.840 more like the technology is so new still and we 0:14:26.920 --> 0:14:29.400 are going to see massive improvements. And I think you 0:14:29.400 --> 0:14:30.840 know the one thing is when you do see a 0:14:30.880 --> 0:14:33.560 cost reduction in these models, what it does do is 0:14:33.600 --> 0:14:36.880 it unlocks companies to be able to deploy more and 0:14:36.920 --> 0:14:38.680 so why it's cheaper to do it just means that 0:14:38.720 --> 0:14:39.560 you can do more of it. 0:14:39.600 --> 0:14:40.960 And that's one of the reasons why we're not. 0:14:40.920 --> 0:14:42.920 Worried about some of the investments we've made because we 0:14:42.960 --> 0:14:45.200 can see the demand. We know if the price goes down, 0:14:45.400 --> 0:14:46.080 our customers are. 0:14:46.000 --> 0:14:46.440 Going to do more. 0:14:46.480 --> 0:14:48.200 They're going to change the user experience for more of 0:14:48.240 --> 0:14:50.640 their customers, and so you know, that's what really excites us. 0:14:50.560 --> 0:14:52.320 About the opportunity really cool stuff. 0:14:52.360 --> 0:14:54.680 Great to talk with someone like yourself who's kind of 0:14:54.760 --> 0:14:56.800 front row seat to all that's going on in AI. 0:14:57.040 --> 0:14:58.600 Dave, thank you so much ye, thank you for your time. 0:14:58.680 --> 0:14:59.120 Dave Brown. 0:14:59.120 --> 0:15:01.320 He's the vice president of a WS Computing and Networking 0:15:01.360 --> 0:15:02.720 at Amazon Web Services. 0:15:04.800 --> 0:15:08.600 This is the Bloomberg Business Week Podcast. Listen live each 0:15:08.600 --> 0:15:12.360 weekday starting at two pm Eastern on Applecarplay and Android 0:15:12.400 --> 0:15:15.280 Auto with the Bloomberg Business app. You can also listen 0:15:15.360 --> 0:15:18.280 live on Amazon Alexa from our flagship New York station, 0:15:18.800 --> 0:15:21.400 Just Say Alexa played Bloomberg eleven thirty. 0:15:23.080 --> 0:15:25.040 All right, now, let's get to some of the news 0:15:25.040 --> 0:15:26.920 that certainly has caught our attention. This has to do 0:15:27.080 --> 0:15:30.400 with New York City Mayor Eric Adams, and that is 0:15:30.400 --> 0:15:34.440 that a federal judge dismissed corruption charges against Eric Adams, 0:15:34.480 --> 0:15:37.040 bringing a permanent end to the case and capping an 0:15:37.080 --> 0:15:43.280 extraordinary series of events that so discord within the Justice Department. 0:15:43.920 --> 0:15:46.040 With more on the judges ruling, we turned to Bloomberg's 0:15:46.120 --> 0:15:49.080 legal reporter ABA Benny Morrison, who joins us from New York. 0:15:49.440 --> 0:15:52.400 What exactly was this ruling? Is that it everything thrown out? 0:15:54.000 --> 0:15:56.480 It certainly looks like it. The judge really took his 0:15:56.720 --> 0:15:59.960 time considering all the arguments that have been flying out 0:16:00.560 --> 0:16:02.800 in the courthouse over the past few months, and he 0:16:02.880 --> 0:16:07.080 reached the decision today that everyone had been anticipating that 0:16:07.160 --> 0:16:10.040 he was going to dismiss the case against Eric Adams 0:16:10.080 --> 0:16:13.400 without prejudice. Now that last bit is important because that 0:16:13.560 --> 0:16:16.640 means this case can't be revived on the track. You 0:16:17.000 --> 0:16:20.240 may remember that the Justice Department in Washington had wanted 0:16:20.280 --> 0:16:23.640 to dismiss the case with prejudice, which would have given 0:16:23.680 --> 0:16:27.080 it the option to bring back the case against Adams 0:16:27.160 --> 0:16:30.400 if they felt the need to, and this led to 0:16:30.440 --> 0:16:35.760 concerns that Washington wanted to maintain leverage over New York's mayor, 0:16:36.120 --> 0:16:38.640 which didn't seem to sit well with the judge. 0:16:39.240 --> 0:16:44.440 It's interesting too, because in that ruling, the judge criticized 0:16:44.440 --> 0:16:46.600 the effort by President Trump's Justice partn't to leave the 0:16:46.680 --> 0:16:48.480 refiling option open. 0:16:48.800 --> 0:16:49.400 What is that? 0:16:51.160 --> 0:16:56.720 Yes, he referred to this situation as unprecedented, and it 0:16:56.800 --> 0:17:00.320 was certainly something that I haven't reported on before. Judge 0:17:00.320 --> 0:17:03.760 said that it was very clear that everything here smacks 0:17:03.800 --> 0:17:07.000 of a bargain dismissal of the indictment in exchange for 0:17:07.160 --> 0:17:11.760 immigration policy concessions. But he warned that dismissing the case 0:17:11.800 --> 0:17:15.800 without prejudice would create that unavoidable perception that the mayor's 0:17:15.880 --> 0:17:19.680 freedoms depended on his ability to carry out the president's wishes, 0:17:20.040 --> 0:17:24.040 so that's why he decided to dismiss the case without prejudice. 0:17:24.320 --> 0:17:28.720 So basically, the mayor can do whatever he wants right 0:17:28.880 --> 0:17:30.919 in terms of even whatever he agreed to maybe with 0:17:31.119 --> 0:17:34.640 President Trump, because he doesn't have to worry about every filing. 0:17:35.640 --> 0:17:36.040 That's right. 0:17:36.080 --> 0:17:38.480 He doesn't have this case hanging over his head anymore. 0:17:39.240 --> 0:17:41.800 So some may argue that he doesn't have the pressure 0:17:41.880 --> 0:17:43.960 of having to do what Washington wants him to do. 0:17:44.960 --> 0:17:48.199