WEBVTT - OpenAI’s Explosive Growth Continues 0:00:02.520 --> 0:00:15.080 Bloomberg Audio Studios, podcasts, radio news. Bloomberg Tech is live 0:00:15.160 --> 0:00:18.880 from the heart of Silicon Valley with Ed Ludlow in 0:00:19.079 --> 0:00:21.960 Van Francisco. 0:00:23.000 --> 0:00:26.640 This is Bloomberg Tech. I'm Tim Steneveek in for Ed Ludlow. 0:00:27.000 --> 0:00:29.680 Coming up on the program. Open AI is on track 0:00:29.720 --> 0:00:33.279 to generate annualized revenue more than forty billion dollars. That's 0:00:33.360 --> 0:00:36.760 roughly doubling its run rate from the end of twenty 0:00:36.800 --> 0:00:39.960 twenty five. Plus, a new city bill backed by Mayor 0:00:40.080 --> 0:00:44.320 Ziorron Mamdani, calls for Amazon to directly employ their delivery workers. 0:00:44.720 --> 0:00:47.919 What will Amazon do if the bill passes? We discuss 0:00:48.680 --> 0:00:51.400 and Meta's former chief AI scientist is joining a new 0:00:51.479 --> 0:00:54.400 venture firm to invest in AI startups. We'll be joined 0:00:54.400 --> 0:00:58.120 by the firm's leader, Sean Johnson. Well, let's start with 0:00:58.160 --> 0:01:01.960 today's big number, and that is forty billion dollars. Open 0:01:02.000 --> 0:01:04.800 a eyes on track to January analyzed revenue that tops 0:01:04.800 --> 0:01:07.680 that number. That's, according to sources, that would roughly double 0:01:07.720 --> 0:01:09.480 its run rate from the end of twenty twenty five. 0:01:09.520 --> 0:01:11.960 I want to bring in Bloomberg's AI reporter Rachel Matt 0:01:12.040 --> 0:01:16.319 who broke this story. Rachel, what specific products are driving 0:01:16.800 --> 0:01:19.040 this annualized revenue run rate that is such a jump 0:01:19.080 --> 0:01:19.720 from last year. 0:01:20.959 --> 0:01:23.600 So according to what we've we've been told in our 0:01:23.640 --> 0:01:26.440 reporting is actually kind of a range of things. The 0:01:26.440 --> 0:01:29.520 company's consumer business is still by far its largest business, 0:01:29.560 --> 0:01:33.319 and they recently announced that they had passed this one 0:01:33.560 --> 0:01:37.480 billion active weekly user mark, which they've been pushing toward 0:01:37.560 --> 0:01:40.360 for a long time. So they've got more paying users 0:01:40.640 --> 0:01:42.920 on the consumer side. They also have a lot more 0:01:43.200 --> 0:01:46.480 businesses that are using it, and this is driven in 0:01:46.480 --> 0:01:48.960 some part by Codex, which is its coding assistant. 0:01:49.240 --> 0:01:51.120 Okay, well, we can't talk about this as open air 0:01:51.160 --> 0:01:52.600 in a vacuum. How does it compare to what the 0:01:52.640 --> 0:01:55.360 latest is from Anthropic, it's chief rival. 0:01:56.760 --> 0:01:58.880 I mean, they've been that sort of neck and neck 0:01:58.920 --> 0:02:01.480 in a way. Okay, I hate to say like it's 0:02:01.480 --> 0:02:04.040 one versus the other, because I think that it's like 0:02:04.160 --> 0:02:06.120 much much bigger and broader than that, And there are 0:02:06.160 --> 0:02:09.160 a lot of companies here in this ecosystem, but I 0:02:09.160 --> 0:02:12.440 think Inthropic definitely has proved over the last year or 0:02:12.520 --> 0:02:15.720 so that they're a really handy rival to this company, 0:02:15.800 --> 0:02:18.760 and they have especially on the coding front. They in 0:02:18.800 --> 0:02:22.280 some ways really outpaced open Ai. But Opening Eye is 0:02:22.280 --> 0:02:24.840 showing us as revenue is going up that customers are 0:02:24.880 --> 0:02:26.959 also willing to pay for its products as well. 0:02:27.360 --> 0:02:29.720 Okay, we're talking top line here, Do we know anything 0:02:30.240 --> 0:02:32.880 about the bottom line and what expenses are and and 0:02:33.080 --> 0:02:36.000 cost of revenue here and what it's spending. 0:02:36.639 --> 0:02:38.800 Ooh, We're still looking into that as far as like 0:02:39.080 --> 0:02:42.280 specific numbers, but what we do know generally, and I 0:02:42.280 --> 0:02:44.840 feel like we've we've talked about this so many times. Yes, 0:02:44.919 --> 0:02:47.120 and it's so frustrating, but it's the same thing over 0:02:47.160 --> 0:02:47.480 and over. 0:02:47.880 --> 0:02:49.559 Compute is really expensive. 0:02:49.880 --> 0:02:50.760 It costs a lot. 0:02:50.960 --> 0:02:54.480 Open Ai has worked very hard to a mass a 0:02:54.520 --> 0:02:58.600 war chest essentially of computing power, and it has that 0:02:58.760 --> 0:03:01.440 at its disposal. And yet the company also will still 0:03:01.480 --> 0:03:04.880 say that its biggest obstacle is getting enough compute. But 0:03:05.240 --> 0:03:07.120 that is going to take up a huge chunk of 0:03:07.160 --> 0:03:10.239 any revenue that it brings in. So other than that, 0:03:10.560 --> 0:03:14.120 it's obviously that's probably by far its largest extense. 0:03:15.400 --> 0:03:18.480 Bloomberg's Rachel Mettz out there on the West coast. Rachel, 0:03:18.480 --> 0:03:20.160 thanks for the update. We're going to keep asking you 0:03:20.400 --> 0:03:22.960 those questions about top and bottom lines at these companies. 0:03:22.960 --> 0:03:26.119 Appreciate your time well. Strategists that City Group and Bank 0:03:26.120 --> 0:03:28.560 of America are finding that the Chips Index is in 0:03:28.639 --> 0:03:32.480 quote bubble land after it rose over too standard deviations 0:03:32.520 --> 0:03:35.960 above its long term trends in real terms, as risks 0:03:35.960 --> 0:03:38.440 develop in the AI trade. Our next guest argues that 0:03:38.560 --> 0:03:44.400 early winners aren't necessarily winners forever, and early winner Nvidia. 0:03:43.960 --> 0:03:45.160 Is at the center of it all. 0:03:45.520 --> 0:03:49.480 Grenadilla, advisory founder at DOO Anna Rathbund, joins us now. 0:03:49.720 --> 0:03:53.120 And you talk about this quote duration mismatch between the 0:03:53.160 --> 0:03:56.800 realities of AI infrastructure and investor expectations. 0:03:56.960 --> 0:03:57.880 How do you quantify that? 0:03:59.280 --> 0:04:03.080 Yeah, you know, we're looking at quarter to quarter numbers, 0:04:03.280 --> 0:04:05.920 and perhaps that's not the right timeframe, even though that's 0:04:05.960 --> 0:04:10.240 what we're used to as equity investors. The hyperscalers and 0:04:10.480 --> 0:04:13.640 semiconductor companies are telling us, you know, look out to 0:04:13.720 --> 0:04:16.400 twenty twenty seven, twenty eight, twenty nine. I mean, they're 0:04:16.440 --> 0:04:19.120 really they're issuing bonds that go all the way out 0:04:19.120 --> 0:04:21.640 to twenty sixty six. I'm not saying that we have 0:04:21.720 --> 0:04:24.760 to wait until twenty twenty six, but we definitely have 0:04:24.839 --> 0:04:28.400 a mismatch in expectations of when we might actually see 0:04:28.440 --> 0:04:32.400 return on some of these investments. The revenue expectations that 0:04:32.480 --> 0:04:35.279 are being promised years out, they have to be able 0:04:35.279 --> 0:04:37.720 to meet it years out, and that takes a lot 0:04:37.800 --> 0:04:41.960 of investment today in numbers that we're totally not used to, right, 0:04:42.040 --> 0:04:44.279 So I think we have to we have to sort 0:04:44.279 --> 0:04:47.279 of adjust our expectations to be a little bit more 0:04:47.320 --> 0:04:49.240 longer term focused than we are today. 0:04:49.400 --> 0:04:51.280 Okay, but how how longer term? 0:04:51.279 --> 0:04:53.240 I mean, at least what we're talking about memory makers, 0:04:53.760 --> 0:04:56.440 we still hear about twenty thirty over and over again. 0:04:56.440 --> 0:05:00.560 I mean the visibility quarter to quarter is I think 0:05:00.560 --> 0:05:02.480 tough for a lot of people to comprehend. How much 0:05:02.480 --> 0:05:03.279 time do we have to wait? 0:05:04.080 --> 0:05:06.719 Yeah, and that's the uncertainty, right And that's why you 0:05:06.800 --> 0:05:11.720 see the volatility in the markets, And that's basically investors 0:05:11.760 --> 0:05:15.520 expressing that uncertainty. The answer, I think is we don't 0:05:15.560 --> 0:05:19.200 really know, and I'm not entirely sure that the hyperscalers 0:05:19.279 --> 0:05:23.479 know either. Right now, they are in an existential race. 0:05:24.000 --> 0:05:27.880 Let's capture that future compute that we get to sell 0:05:27.960 --> 0:05:30.640 by building data centers and buying in video chips and 0:05:30.680 --> 0:05:33.080 perhaps making our own chips. And that is I think 0:05:33.080 --> 0:05:35.560 how the industry is changing, and frankly, we have. 0:05:35.520 --> 0:05:37.239 To wait and see a sort of field of dreams. 0:05:37.240 --> 0:05:41.040 If we build it, they will come maybe okay, maybe, 0:05:41.160 --> 0:05:43.520 Well we'll help us envision that future and what it 0:05:43.560 --> 0:05:46.920 looks like. Because there's the conversation too happening where there's 0:05:46.920 --> 0:05:49.919 no guarantee on this ROI. Do you buy into that 0:05:50.040 --> 0:05:51.920 or do you think that there will be an ROI. 0:05:51.760 --> 0:05:55.039 On this I think there will be an ROI on it. 0:05:55.120 --> 0:05:58.039 I mean, And this is the other expectation that we 0:05:58.160 --> 0:06:02.560 have perhaps misplaced, is the level of compute. 0:06:02.080 --> 0:06:03.640 Right overall investors. 0:06:03.680 --> 0:06:06.960 Are we're used to looking at the last twenty years 0:06:07.040 --> 0:06:11.200 of Internet and cloud growth, etc. At the AI level, 0:06:11.320 --> 0:06:14.719 the sophistication of these models, and just how much compute 0:06:14.720 --> 0:06:17.120 that it takes. I'm not entirely sure the market has 0:06:17.200 --> 0:06:19.920 digested it, and that's why there's so much money going 0:06:19.960 --> 0:06:20.360 into it. 0:06:20.440 --> 0:06:22.360 I do believe that we will see the ROI. 0:06:22.440 --> 0:06:25.640 It's just as you mentioned before, it's the timing is uncertain, 0:06:25.920 --> 0:06:27.760 and that's why we have to take a long term view. 0:06:28.120 --> 0:06:29.160 Are in your view? 0:06:29.160 --> 0:06:30.960 Are we talking about the right names? Are we looking 0:06:31.040 --> 0:06:33.200 at the right names at least on the public equity side. 0:06:33.200 --> 0:06:35.000 We're going to talk to some venture capitalists later a 0:06:35.040 --> 0:06:37.080 little later this hour. You know, they're thinking about it 0:06:37.120 --> 0:06:39.360 from the perspective of Okay, what are the early companies 0:06:39.400 --> 0:06:41.880 now we should be investing in. But you focus on 0:06:41.880 --> 0:06:44.200 public markets. Are we talking about the right names here? 0:06:45.400 --> 0:06:47.000 I think those are the names that we have to 0:06:47.040 --> 0:06:49.200 talk about because that's where the spend is coming from, 0:06:49.279 --> 0:06:52.000 that's where the circular financing is, that's where the debt 0:06:52.040 --> 0:06:54.560 issuance is coming from. And so if you are a 0:06:54.640 --> 0:06:57.800 public market investors, those names you do have to pay 0:06:57.800 --> 0:06:59.960 attention to. Now, what we have discovered in the last 0:07:00.120 --> 0:07:03.160 two or three years is that sometimes names come up 0:07:03.400 --> 0:07:06.080 out of nowhere, out of the left field in terms 0:07:06.160 --> 0:07:09.800 of being in the front line of AI. So until 0:07:09.840 --> 0:07:12.640 we see those names, I think we are focusing on 0:07:12.720 --> 0:07:15.200 the names that we see today. But we do have 0:07:15.280 --> 0:07:18.280 to think about the risks. You mentioned Nvidia earlier, it 0:07:18.320 --> 0:07:21.040 was an early winner. Now everyone is making their own chips. 0:07:21.320 --> 0:07:25.120 What does that mean for Nvidia? Now Corewave is actually 0:07:25.640 --> 0:07:29.600 dependent on Nvidia and their debt is actually collateralized by 0:07:29.880 --> 0:07:30.680 Nvidia chips. 0:07:30.760 --> 0:07:32.160 What does that mean for Nvidia? 0:07:32.240 --> 0:07:34.400 I mean, I think we have to look at it differently, 0:07:34.440 --> 0:07:37.400 give it a second look in the same names that 0:07:37.520 --> 0:07:38.600 we all talk about today. 0:07:38.680 --> 0:07:40.080 Yeah, well, we'll find out a couple of weeks what 0:07:40.160 --> 0:07:42.520 Nvidia says August twenty six is when they're scheduled to 0:07:42.560 --> 0:07:45.760 report their most recent quarter. And before we let you go, 0:07:45.920 --> 0:07:48.520 you mentioned circular financing. Certainly that gets a lot of attention, 0:07:48.640 --> 0:07:52.160 but increasingly we're hearing about concentration risks and worries that 0:07:52.200 --> 0:07:55.880 some of these hyperscalers are really have two customers, anthropic 0:07:56.160 --> 0:07:59.360 and open AI. What's your view on concentration risk. 0:08:00.280 --> 0:08:03.760 Yeah, concentration risk is something that these companies need to 0:08:03.840 --> 0:08:06.320 work on. I mean, we heard from Cerebras saying that 0:08:06.880 --> 0:08:09.680 they actually have a huge concentration risk, right, But they're 0:08:09.720 --> 0:08:12.560 actually new to the public markets. So I think we're 0:08:12.600 --> 0:08:15.360 going to have to expect that to reduce the risk. 0:08:15.760 --> 0:08:17.360 If at least they're not going to be able to 0:08:17.360 --> 0:08:20.120 tell us when we're going to get the ROI tell 0:08:20.200 --> 0:08:22.480 us that they are actually reducing the risk on their 0:08:22.520 --> 0:08:23.400 business side. 0:08:24.040 --> 0:08:26.840 Yeah, well, I guess the question that I have is 0:08:26.840 --> 0:08:29.880 who would take the place of those large customers at 0:08:29.880 --> 0:08:30.400 this point? 0:08:30.880 --> 0:08:32.640 Well, right, and that's is Algeum this. 0:08:33.200 --> 0:08:36.439 Yeah, that's the conundrum of the circular financing, right, because 0:08:36.679 --> 0:08:39.600 you know, it takes so much capital and size in 0:08:39.720 --> 0:08:42.680 order to build this AI ecosystem that there are a 0:08:42.720 --> 0:08:45.719 few players, right, and so I mean, I think we're 0:08:45.760 --> 0:08:47.160 going to be able to see in about a year 0:08:47.280 --> 0:08:49.720 or two. I mean, we have the hyperscalers trying to 0:08:49.760 --> 0:08:54.520 reduce their dependence on the number one supplier, and I 0:08:54.559 --> 0:08:56.720 think the companies are going to continue to make those 0:08:56.760 --> 0:08:59.960 innovative changes to their business model. And so this is, 0:09:00.120 --> 0:09:02.360 unfortunately is also a wait and see because of the 0:09:02.480 --> 0:09:04.880 nature of the AI ecosystem and how much money it 0:09:04.920 --> 0:09:05.800 takes to build it out. 0:09:06.120 --> 0:09:06.360 MANA. 0:09:06.440 --> 0:09:10.920 Rathbund of Grenadilla Advisory joining us this morning on Bloomberg Tech. Anna, 0:09:10.960 --> 0:09:12.640 thanks so much, do appreciate your time. 0:09:13.200 --> 0:09:13.360 Well. 0:09:13.400 --> 0:09:17.880 Coming up, applied Materials beats expectations investors, though, oh they're 0:09:17.920 --> 0:09:20.320 not so impressed. Shares down by about four point six 0:09:20.360 --> 0:09:22.520 percent right now. We're going to break down the chip 0:09:22.559 --> 0:09:27.880 equipment giants outlook next. I'm also watching shares a broadcom 0:09:28.640 --> 0:09:32.360 that owns VMware. Shares a broadcom down this morning, trading 0:09:32.400 --> 0:09:34.440 and dragging the S and P five hundred lower. It 0:09:34.440 --> 0:09:38.560 follows a report that VMware security vulnerability being actively exploited 0:09:38.600 --> 0:09:42.000 across several countries. Broadcom shares off their worse levels, but 0:09:42.080 --> 0:09:46.120 still down five point three percent. This is Bloomberg Tech. 0:09:47.679 --> 0:09:54.840 AMD has raised four point seventy five billion dollars in 0:09:54.880 --> 0:09:57.520 its biggest ever US dollar bond sale. It joins a 0:09:57.520 --> 0:10:00.440 wave of borrowing tied to the AI boom, the chip 0:10:00.440 --> 0:10:04.480 maker ramping up spending as demand for computing capacity surges. 0:10:04.880 --> 0:10:08.000 AD says the proceeds will be used for general corporate purposes, 0:10:08.040 --> 0:10:10.719 potentially including debt repayment, with eight hundred and seventy five 0:10:10.760 --> 0:10:11.520 million dollars. 0:10:11.320 --> 0:10:13.240 Of bonds due next month. 0:10:14.280 --> 0:10:17.400 A Fly Materials delivered a better than expected forecast investors. 0:10:17.400 --> 0:10:20.600 Though hard to impress, it's the biggest US chip maker 0:10:20.760 --> 0:10:23.360 of equipment, I should say. It says that customers are 0:10:23.400 --> 0:10:26.400 pushing it to ramp up production even faster. Bloomberg's I 0:10:26.440 --> 0:10:29.520 and King covers all things semiconductors, and he joins US now. 0:10:29.840 --> 0:10:31.240 Ian SO shares. 0:10:31.240 --> 0:10:33.120 We're up about one hundred and eight percent going into 0:10:33.120 --> 0:10:36.640 this print yesterday. Expectations fair to say, we're pretty high. 0:10:37.160 --> 0:10:39.200 Looks like a pretty good print across the board. What 0:10:39.200 --> 0:10:40.320 are investors upset. 0:10:40.120 --> 0:10:45.200 About Yeah, whether the actually upset about buy anything in 0:10:45.200 --> 0:10:47.640 particular isn't really clear. I Mean, all of the numbers 0:10:47.640 --> 0:10:50.360 look very clean, but really what we've got is yet 0:10:50.400 --> 0:10:52.720 another case of a company comes out and says, hey, 0:10:52.840 --> 0:10:55.400 guess what our innings are going to be hundreds of 0:10:55.440 --> 0:10:58.680 millions of dollars better than you had expected. And the 0:10:58.720 --> 0:11:01.000 reaction they get is, yeah, great, we know that. What 0:11:01.040 --> 0:11:02.720 are you going to do for us next And so 0:11:02.800 --> 0:11:04.640 a lot of the conversation on the call was about 0:11:04.679 --> 0:11:05.160 next year. 0:11:05.400 --> 0:11:08.840 Okay, well, let's talk about that, because the question that 0:11:08.880 --> 0:11:10.559 I think a lot of people have is about how 0:11:10.559 --> 0:11:13.920 this firm is somewhat of a barometer for the entire industry. 0:11:14.200 --> 0:11:16.120 Can you talk a little bit about visibility that we 0:11:16.240 --> 0:11:20.080 got from applied Materials into what the rest of the 0:11:20.120 --> 0:11:21.400 year and what next year looks like. 0:11:22.120 --> 0:11:24.000 Yeah, no, that's a very good question, and this is 0:11:24.000 --> 0:11:26.040 one of the reasons why you would concentrate on a 0:11:26.040 --> 0:11:29.000 company that does something a little bit more esoteric. So 0:11:29.080 --> 0:11:32.360 what the CEO told us in a conversation, you know 0:11:32.400 --> 0:11:35.559 after the numbers hit, was look, I've got eight quarters 0:11:36.000 --> 0:11:38.920 of guidance from my customers because it takes so long 0:11:39.440 --> 0:11:42.240 to make this equipment, to then ship it, to then 0:11:42.320 --> 0:11:44.839 install it and get it up and running in these factories. 0:11:45.120 --> 0:11:47.160 You know, we have to plan over the horizon and 0:11:47.160 --> 0:11:50.080 guess what, they're giving me even more visibility than they 0:11:50.120 --> 0:11:52.800 normally would. So he's saying, look, next year, things are 0:11:52.840 --> 0:11:55.719 going to get better. So the call was all about, well, 0:11:55.720 --> 0:11:57.960 how much better? Tell us exactly how much better? And 0:11:58.520 --> 0:12:01.760 you know he's saying, look a lot better. So I 0:12:01.800 --> 0:12:04.120 don't really know what more of a conversation, well that 0:12:04.200 --> 0:12:06.600 you could have as a management team at this point, Ian. 0:12:06.440 --> 0:12:08.120 Before we let you go, we have thirty seconds. You've 0:12:08.120 --> 0:12:10.240 been doing this a long time and covering these types 0:12:10.280 --> 0:12:13.040 of companies a long time through different cycles, So I'm 0:12:13.040 --> 0:12:15.640 just curious contextualize the moment for us. Have you seen 0:12:15.640 --> 0:12:16.240 anything like this? 0:12:17.400 --> 0:12:21.040 Never the extent that we are, you know, sort of 0:12:21.320 --> 0:12:24.839 getting in front of ourselves and projecting growth. We've never 0:12:24.880 --> 0:12:26.880 seen anything like this. We've seen ups and downs with 0:12:26.880 --> 0:12:29.160 a fool, but never anything to this extent. 0:12:29.440 --> 0:12:31.720 Ian King, It's always great to chat with you. Thanks 0:12:31.720 --> 0:12:33.839 for getting up early. Do appreciate that's Ian King out 0:12:33.840 --> 0:12:37.199 in our San Francisco bureau. Well Chinese aifirm z dot 0:12:37.240 --> 0:12:39.800 Ai is upgrading its flagship model as it looks to 0:12:39.800 --> 0:12:42.719 close the gap with US rivals Open Ai and Anthropic. 0:12:43.200 --> 0:12:46.720 The new GLM five point three will focus on stronger 0:12:46.760 --> 0:12:50.839 coding capabilities and stronger performance gains over its previous model. 0:12:51.160 --> 0:12:53.720 The company plans to release the new LLM weights within 0:12:53.760 --> 0:12:58.640 two weeks, allowing developers to download and customize it. Sanford's 0:12:58.679 --> 0:13:01.920 AI Index finds a striking divide and attitudes toward AI 0:13:02.000 --> 0:13:05.199 in China in the US. But the question what's driving 0:13:05.240 --> 0:13:08.800 that gap and could China's enthusiasm give its AI industry 0:13:08.800 --> 0:13:11.600 and edge. Bloomberg's minmen Low explaance. 0:13:12.840 --> 0:13:15.800 This Stanford University AI Index shows that more than eighty 0:13:15.800 --> 0:13:19.200 percent of Chinese are optimistic towards AI, compared to less 0:13:19.240 --> 0:13:22.720 than forty percent in the US. And Edelmann has a 0:13:22.800 --> 0:13:26.120 similar survey showing the same trend that more than seventy 0:13:26.160 --> 0:13:29.679 percent of Chinese trusts AI compared to less than forty 0:13:29.679 --> 0:13:32.520 percent in the US. And again, why is there such 0:13:32.520 --> 0:13:37.280 a huge gulf in attitudes between Chinese and Americans? One 0:13:37.320 --> 0:13:41.120 reason could be people's experience with technology in China over 0:13:41.120 --> 0:13:44.960 the last two decades. It's still in a rapid development phase, 0:13:45.000 --> 0:13:49.839 and people's experience of tech is one of empowerment and transformation. 0:13:50.200 --> 0:13:53.520 We're talking about rural residents getting to sell their goods 0:13:53.600 --> 0:13:57.679 through e commerce channels, elite frogging credit cards to become 0:13:57.720 --> 0:14:01.800 a cashless society, experiencing the convenience of door to door 0:14:01.840 --> 0:14:05.920 delivery and very cheap efficient logistics for example. But in 0:14:05.960 --> 0:14:09.520 the US that's already a norm, right. People are familiar 0:14:09.520 --> 0:14:12.760 with big box retail, they have reliable mail, and the 0:14:12.960 --> 0:14:16.640 experience of the tech explosion has been one of addiction, 0:14:17.280 --> 0:14:21.280 data privacy, violation, over concentration of power in big tech. 0:14:21.600 --> 0:14:24.640 So that's one reason, and the other concern is around 0:14:24.840 --> 0:14:27.400 whether there is enough guard rails that has been put 0:14:27.400 --> 0:14:30.320 in place. And in China, these tech films are still 0:14:30.400 --> 0:14:34.080 under the thumb of the government. They're constantly worried about crackdowns. 0:14:34.280 --> 0:14:38.080 And when the government pushed this Common Prosperity campaign, tech 0:14:38.080 --> 0:14:41.680 films were rushing to pledge billions of un to support that. 0:14:41.800 --> 0:14:43.880 While in the US it's a different story where you 0:14:43.960 --> 0:14:47.200 have a revolving door where tech executives raise funds to 0:14:47.280 --> 0:14:50.360 run political campaigns, where billions are being poured into lobbying. 0:14:50.640 --> 0:14:53.400 So again that concern around guard rails and there's just 0:14:53.440 --> 0:14:56.880 not enough trust that the government is protecting the interests 0:14:56.920 --> 0:15:00.320 of residents. So why does this matter When some people 0:15:00.360 --> 0:15:03.640 are more willing to try new technology? That means market 0:15:03.680 --> 0:15:07.040 adoption rate can be much faster. And while China legs 0:15:07.080 --> 0:15:10.000 behind in some of the foundational technology like the chips 0:15:10.040 --> 0:15:13.160 and the hardware, Chinese firms could have an advantage in 0:15:13.240 --> 0:15:17.240 having that more futile ground to test new products and 0:15:17.400 --> 0:15:22.600 to allow much faster adoption of new AI products. Mimnlo 0:15:22.600 --> 0:15:24.320 Bloomboo News, Hong Kong. 0:15:25.600 --> 0:15:28.840 Well so breaking news crossing the Bloomberg terminal. Luigi Mangioni 0:15:28.920 --> 0:15:32.800 to plead guilty to killing of healthcare CEO. Once again, 0:15:32.840 --> 0:15:36.720 Luigi Mangioni to plead guilty to US charges. This according 0:15:36.920 --> 0:15:40.040 to Luigi Mangione's lawyer, will continue to bring you updates 0:15:40.080 --> 0:15:43.720 on that story as it continues to break well. Staying 0:15:43.800 --> 0:15:47.320 with China's Ai boom, Deepseek is sharply raising prices for 0:15:47.400 --> 0:15:51.080 its flagship V four models. Peak hour rates will more 0:15:51.080 --> 0:15:54.720 than quadruple starting August sixteenth. This is the startup prepares 0:15:54.760 --> 0:15:58.600 for a potential IPO and puts greater focus on profitability. 0:15:59.000 --> 0:16:01.400 Even after the hike, deeps Seek says its models remains 0:16:01.400 --> 0:16:06.560 significantly cheaper than major rivals. Well, coming up, how close 0:16:06.600 --> 0:16:10.120 are we really to a world full of humanoid robots? 0:16:10.200 --> 0:16:27.000 We've got some answers. Next, this is Bloomberg. The promise 0:16:27.080 --> 0:16:32.000 of embodied AI think humanoid robots, drones, autonomous vehicles seems 0:16:32.040 --> 0:16:34.640 to be coming into its own and increasingly entering the 0:16:34.680 --> 0:16:37.840 real world. Bloomberg Tech Europe's Tom McKenzie spoke with Google 0:16:37.880 --> 0:16:41.440 Deep Mind set of Robotics Carolina Parata about whether physical 0:16:41.440 --> 0:16:44.480 AI is finally having its quote chat GPT moment. 0:16:45.400 --> 0:16:47.600 I would say not yet, but you can see it. 0:16:47.680 --> 0:16:49.480 You can see it, and you can see the lights. 0:16:51.200 --> 0:16:53.240 In my opinion depends what you call the TGIC moment 0:16:53.320 --> 0:16:55.720 for robotics. In my opinion of ChiPT moment for robotics 0:16:55.760 --> 0:16:59.720 would be one where everybody can experience what it's like 0:16:59.760 --> 0:17:02.480 to interact with the robot, and the robot responds to 0:17:02.520 --> 0:17:04.879 what you want, and that's what you want in a 0:17:04.880 --> 0:17:08.040 completely new environment or a completely new setting. So that's 0:17:08.080 --> 0:17:09.840 what it means to me. I don't think we're there yet, 0:17:10.840 --> 0:17:12.320 but we're getting there. 0:17:12.440 --> 0:17:13.640 What gets us there? 0:17:13.680 --> 0:17:16.760 What are the technological barriers that need to be overcome 0:17:16.920 --> 0:17:17.760 to get to that point? 0:17:17.920 --> 0:17:20.560 Yeah, I mean this is precisely what we're invested on. 0:17:20.760 --> 0:17:23.960 So our goal is to build the intelligence layer that 0:17:24.080 --> 0:17:27.680 can power any robot and make it really intelligent, so 0:17:27.720 --> 0:17:30.280 that it's includive for humans to use, so that it 0:17:30.320 --> 0:17:32.639 can do a broad range of tasks, so that it 0:17:32.680 --> 0:17:36.080 can make the robot smart enough to understand an environment, 0:17:36.359 --> 0:17:39.480 reason about it, and then take action to complete a 0:17:39.560 --> 0:17:41.840 completely new task that you might have asked to do. 0:17:42.400 --> 0:17:44.919 That's what we're going after. I think that this is 0:17:45.080 --> 0:17:48.639 there's been significant progress over the last three four years. 0:17:48.720 --> 0:17:52.560 I mean, before robots didn't even understand what an object 0:17:52.680 --> 0:17:55.359 was and what their environment that were just simply the 0:17:55.920 --> 0:17:59.840 points in space without any meaning. We've actually introduced around 0:18:00.119 --> 0:18:02.840 if you need to lllms and blms into robots that 0:18:03.000 --> 0:18:06.520 taught them but naturally like understand to understand natural language. 0:18:06.560 --> 0:18:08.920 It also taught them how the reason about their environment 0:18:09.200 --> 0:18:10.880 and then how do you take action in a way 0:18:11.280 --> 0:18:15.400 that was touris a particular semantic task or a particular 0:18:15.480 --> 0:18:17.600 meaningful task for the robot that didn't exist before. 0:18:17.760 --> 0:18:20.800 Humanoids get a lot of headlines, they get a lot 0:18:20.840 --> 0:18:23.879 of pause. Do you think that's distracting? Do you think investors, 0:18:24.600 --> 0:18:27.440 the media, others. Do you think we're still overly obsessed 0:18:27.480 --> 0:18:28.880 with humanoids. 0:18:29.359 --> 0:18:31.280 That's a great question. I mean, we be live in 0:18:31.320 --> 0:18:33.600 a world where there will be many different robotypes. I 0:18:33.640 --> 0:18:36.359 actually think I'm very practical about this. I think that 0:18:36.400 --> 0:18:37.879 there's going to be a lot of tasks where a 0:18:37.880 --> 0:18:39.879 different form factor that is not a humanoid is the 0:18:39.920 --> 0:18:42.760 right thing. However, I do think humanoids are going to 0:18:42.800 --> 0:18:44.800 become useful and they are going to be an important form 0:18:44.880 --> 0:18:48.560 factor for a few reasons. First, I think the world 0:18:48.680 --> 0:18:51.600 is made for humans. So if you want to immediately 0:18:51.600 --> 0:18:55.320 deploy a robot in an environment that is human centric, 0:18:55.400 --> 0:18:57.440 then a humanoid is probably going to be an easy 0:18:57.480 --> 0:19:01.720 one to deploy. Actually, there's an interesting research aspect to this, 0:19:02.200 --> 0:19:04.760 which is if we want robots to learn from humans 0:19:05.359 --> 0:19:08.160 and the robot has a human form factor, it's actually 0:19:08.160 --> 0:19:11.040 easier because it can now translate what the human is 0:19:11.080 --> 0:19:13.960 doing when it's say cooking a recipe, into what the 0:19:14.320 --> 0:19:16.320 robots should do in order to do the same task. 0:19:16.560 --> 0:19:19.800 So that's a learning perspective that I find interesting. And 0:19:19.840 --> 0:19:21.560 then The other piece is that my goal is to 0:19:21.600 --> 0:19:24.720 really show that we can get to AGA in the 0:19:24.720 --> 0:19:27.480 physical world. What does that mean. It really means that 0:19:27.520 --> 0:19:29.919 the robot can do anything the human can. So I 0:19:29.920 --> 0:19:32.240 feel like humanoids are going to enable us to show 0:19:32.920 --> 0:19:35.200 and really believe that we have gotten to that level. 0:19:35.320 --> 0:19:37.840 So that is another reason why I like it. But 0:19:38.080 --> 0:19:40.400 we care about all kinds of robots, and we're interested 0:19:40.400 --> 0:19:43.000 in enabling bringing intelligence to all robots. 0:19:43.840 --> 0:19:47.120 That was Google d minds Carolina Parata speaking with Bloomberg's 0:19:47.160 --> 0:19:49.680 Tom McKenzie. You can check out the full interview in 0:19:49.720 --> 0:19:52.240 the latest episode of Bloomberg Tech Europe as Tom looks 0:19:52.240 --> 0:19:55.400 into whether we are on the cusp of a robot reality. 0:19:55.520 --> 0:19:58.840 You can find it on YouTube and Bloomberg dot Com. Well, 0:19:58.880 --> 0:20:00.560 let's take a look at more at tex stories with 0:20:00.680 --> 0:20:03.520 Bloomberg's You Hire annd Now You Hire a Take it away? 0:20:03.840 --> 0:20:06.360 Hi, Tim, It's time now for talking tech. First up, 0:20:06.640 --> 0:20:09.520 Francis Highest Court blocked a ban on social media for 0:20:09.600 --> 0:20:13.160 children under the age of fifteen, calling the ruling unconstitutional 0:20:13.480 --> 0:20:16.960 for harming youth's freedom of speech and communication. It really 0:20:17.000 --> 0:20:20.240 shows how global efforts to ban kids from social media 0:20:20.280 --> 0:20:25.440