WEBVTT - AWS CEO Matt Garman Talks Amazon's Latest Innovations, AI Goals

0:00:02.400 --> 0:00:06.760
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio news.

0:00:07.160 --> 0:00:09.720
<v Speaker 2>It is a joy to be joined by Matt. Thank you,

0:00:09.760 --> 0:00:14.000
<v Speaker 2>remain Matt. Welcome from Las Vegas, Reinvent upon You Chip,

0:00:14.200 --> 0:00:18.320
<v Speaker 2>Ultra Clusters, Ultra Service Chip Development LLM, and we've got

0:00:18.480 --> 0:00:20.959
<v Speaker 2>talk about developer tools. There is a lot to get through.

0:00:21.120 --> 0:00:23.919
<v Speaker 2>Let's start with the ultra cluster. Can you just talk

0:00:23.960 --> 0:00:25.640
<v Speaker 2>to us a little bit about this more than one

0:00:25.720 --> 0:00:29.560
<v Speaker 2>hundred thousand chips that will be put in one single

0:00:29.640 --> 0:00:33.520
<v Speaker 2>era for AI hardware. Why do it? What will it offer?

0:00:34.840 --> 0:00:37.080
<v Speaker 1>Yeah, this is something that we're building together with our

0:00:37.120 --> 0:00:39.960
<v Speaker 1>partner's anthropic and it's actually going to be several hundred

0:00:40.000 --> 0:00:44.120
<v Speaker 1>thousand chips, and they're focused on it's our new Trainium

0:00:44.120 --> 0:00:46.520
<v Speaker 1>two chips, and so we expect that this should deliver

0:00:46.600 --> 0:00:49.519
<v Speaker 1>them five times more compute than they've used in their

0:00:49.600 --> 0:00:51.920
<v Speaker 1>last model training, and so we're really excited about what

0:00:51.960 --> 0:00:55.640
<v Speaker 1>they're able to accomplish, and they're expecting to build much

0:00:55.680 --> 0:00:59.520
<v Speaker 1>bigger and much more capable AI models from that large

0:00:59.520 --> 0:01:00.120
<v Speaker 1>compute clubs.

0:01:00.600 --> 0:01:02.880
<v Speaker 2>Do we have a US location? Do we have an

0:01:02.880 --> 0:01:05.320
<v Speaker 2>exact timeframe other than early twenty twenty five?

0:01:06.480 --> 0:01:08.160
<v Speaker 1>No, No, no location to share with.

0:01:08.640 --> 0:01:11.440
<v Speaker 2>Matt what then, of the technical challenges paint the picture

0:01:11.480 --> 0:01:15.920
<v Speaker 2>for our audience how difficult this is in terms of cooling,

0:01:16.080 --> 0:01:18.400
<v Speaker 2>in terms of energy use. What has been the things

0:01:18.400 --> 0:01:19.440
<v Speaker 2>that you have to overcome with this?

0:01:20.319 --> 0:01:23.080
<v Speaker 1>Yeah, Well, the first core innovation is that we built

0:01:23.120 --> 0:01:25.920
<v Speaker 1>our own chip actually, and so we've built our own chip.

0:01:25.920 --> 0:01:28.880
<v Speaker 1>It's called Trainium two, and we're quite excited about the

0:01:28.920 --> 0:01:31.240
<v Speaker 1>performance that we get out of that. And that's combined

0:01:31.440 --> 0:01:35.600
<v Speaker 1>all in these really large ultra clusters that combine sixty

0:01:35.640 --> 0:01:38.480
<v Speaker 1>four tranium chips all to deliver eighty three petaflops from

0:01:38.520 --> 0:01:40.959
<v Speaker 1>a single node. And so that is the first innovation

0:01:41.240 --> 0:01:45.760
<v Speaker 1>is our Amazon designed custom silicon that gives us really

0:01:45.840 --> 0:01:49.400
<v Speaker 1>unparalleled performance for generative AIA capabilities. And then we build

0:01:49.440 --> 0:01:52.440
<v Speaker 1>these together with high performance networking that we also build

0:01:52.480 --> 0:01:55.040
<v Speaker 1>in house, and then of course they'll be cooling and

0:01:55.080 --> 0:01:58.880
<v Speaker 1>heating and power that we need to go to go

0:01:58.880 --> 0:02:01.280
<v Speaker 1>build the data centers. But it really starts with that

0:02:01.400 --> 0:02:03.840
<v Speaker 1>silicon down at the level, and we innovate on the

0:02:03.960 --> 0:02:06.120
<v Speaker 1>entire stack of AI to make sure that we can

0:02:06.160 --> 0:02:08.240
<v Speaker 1>control everything that goes into those clusters.

0:02:08.440 --> 0:02:10.040
<v Speaker 2>Can you talk to us a little bit compare and

0:02:10.040 --> 0:02:14.040
<v Speaker 2>contrast here this is about less dependency on Nvidio GPUs

0:02:14.080 --> 0:02:17.440
<v Speaker 2>too many ways, doing an alternative for your clients. What

0:02:17.520 --> 0:02:19.800
<v Speaker 2>sort of cost saving will clients get? What sort of

0:02:19.960 --> 0:02:21.919
<v Speaker 2>energy and efficiency will they get?

0:02:22.800 --> 0:02:24.440
<v Speaker 1>Yeah? Well, first of all, I think we like to

0:02:24.440 --> 0:02:27.000
<v Speaker 1>think of it as a supplement to in video GPUs

0:02:27.000 --> 0:02:29.560
<v Speaker 1>and Vidia has a fantastic product that team has done

0:02:29.560 --> 0:02:32.399
<v Speaker 1>an outstanding job executing, and we think that the vast

0:02:32.480 --> 0:02:34.400
<v Speaker 1>majority of workloads are going to continue to run in

0:02:34.520 --> 0:02:37.600
<v Speaker 1>video processors for a long time. But customers want choice,

0:02:37.639 --> 0:02:39.760
<v Speaker 1>and they want choice that can give them some lower

0:02:39.760 --> 0:02:42.800
<v Speaker 1>cost options. And we think that for certain workloads, for

0:02:42.880 --> 0:02:46.040
<v Speaker 1>many workloads, Trainium two can give customers thirty to forty

0:02:46.040 --> 0:02:50.880
<v Speaker 1>percent cost performance benefits over today's GPU powered instances. And

0:02:50.919 --> 0:02:53.200
<v Speaker 1>so we think that that's a huge win for customers,

0:02:53.280 --> 0:02:55.760
<v Speaker 1>particularly as they're looking to lower the cost of Generator

0:02:55.919 --> 0:02:58.960
<v Speaker 1>I workloads. But we'll be great partners with Nvidia and

0:02:58.960 --> 0:03:01.800
<v Speaker 1>continue to lean in on building great technologies together with

0:03:01.840 --> 0:03:02.639
<v Speaker 1>them for a long time.

0:03:03.040 --> 0:03:06.400
<v Speaker 2>Train two, Trainium two available, TRAININGUM three in the works.

0:03:06.480 --> 0:03:09.359
<v Speaker 2>And all of this is as Nvidia at the moment

0:03:09.720 --> 0:03:12.760
<v Speaker 2>is estimated has about ninety five percent market shared. Do

0:03:12.800 --> 0:03:14.280
<v Speaker 2>you agree with that sort of number they have and

0:03:14.360 --> 0:03:15.840
<v Speaker 2>what sort of area do you think that that will

0:03:15.880 --> 0:03:16.400
<v Speaker 2>come down to.

0:03:17.520 --> 0:03:19.280
<v Speaker 1>Yeah, I mean I think it's probably higher than that.

0:03:19.320 --> 0:03:21.840
<v Speaker 1>I think that the vast majority of workloads in generative

0:03:21.880 --> 0:03:26.840
<v Speaker 1>AI today run on in Nvidia technology and they've absolutely

0:03:26.880 --> 0:03:29.120
<v Speaker 1>been the leaders in that space. But we do think

0:03:29.160 --> 0:03:31.680
<v Speaker 1>we hear from customers that they want choice, and just

0:03:31.760 --> 0:03:34.960
<v Speaker 1>with our processors where we're type partners with Intel and

0:03:34.960 --> 0:03:37.560
<v Speaker 1>with AMD, but we decided to go build a general

0:03:37.560 --> 0:03:40.640
<v Speaker 1>purpose processor are called Graviton, and it's been hugely successful

0:03:40.680 --> 0:03:43.280
<v Speaker 1>with our customers. But we also provide lots of Intel

0:03:43.280 --> 0:03:46.160
<v Speaker 1>and lots of AMD processors in our cloud today and

0:03:46.280 --> 0:03:49.680
<v Speaker 1>those businesses continue to grow. So I expect that our

0:03:49.800 --> 0:03:53.040
<v Speaker 1>usage of Nvidia will continue to grow from our customers

0:03:53.200 --> 0:03:55.800
<v Speaker 1>and that choice is really going to be powerful. And

0:03:55.840 --> 0:03:58.840
<v Speaker 1>as you see this explosion of generative AI usage, I

0:03:58.840 --> 0:04:01.200
<v Speaker 1>think there's going to be plenty of business for multiple

0:04:01.240 --> 0:04:02.520
<v Speaker 1>different people to be successful.

0:04:02.640 --> 0:04:05.280
<v Speaker 2>I keep partners Intel. How concerned have you been about

0:04:05.280 --> 0:04:06.040
<v Speaker 2>the change at the top.

0:04:07.600 --> 0:04:09.760
<v Speaker 1>Oh, it's okay. I wish pat the best. You know,

0:04:09.800 --> 0:04:12.800
<v Speaker 1>he's I know Patent, he's been a good partner of ours,

0:04:12.800 --> 0:04:15.120
<v Speaker 1>but we've been partners with Intel for a long time.

0:04:15.160 --> 0:04:18.120
<v Speaker 1>It's been eighteen years since we launched our first Intel

0:04:18.160 --> 0:04:21.560
<v Speaker 1>instance when AWS and EC two first launched, and will

0:04:21.600 --> 0:04:23.760
<v Speaker 1>continue to be great partners with Intel, and they have

0:04:23.760 --> 0:04:26.960
<v Speaker 1>a great technology team there, and we look forward to

0:04:26.960 --> 0:04:29.640
<v Speaker 1>to continuing to roll out the latest technologies from Intel

0:04:29.680 --> 0:04:31.160
<v Speaker 1>for our customers to be able to use, and.

0:04:31.160 --> 0:04:33.160
<v Speaker 2>He've been continuing to roll out, as you say in

0:04:33.279 --> 0:04:36.800
<v Speaker 2>video offerings last time you're on, we're talking about Blackwell.

0:04:36.880 --> 0:04:40.279
<v Speaker 2>Of course, many have been frustrated, perhaps by some slowness

0:04:40.320 --> 0:04:42.480
<v Speaker 2>to market there. When do you anticipate that Blackwell will

0:04:42.520 --> 0:04:46.360
<v Speaker 2>be unfolding? And how difficult has it been to ensure

0:04:46.360 --> 0:04:47.360
<v Speaker 2>the supply side.

0:04:47.120 --> 0:04:47.640
<v Speaker 1>Is there.

0:04:48.960 --> 0:04:49.159
<v Speaker 2>Well.

0:04:49.320 --> 0:04:51.520
<v Speaker 1>They've obviously had some manufacturing things that they're going through,

0:04:51.520 --> 0:04:53.120
<v Speaker 1>but we're very excited about it. I think the early

0:04:53.160 --> 0:04:57.080
<v Speaker 1>returns and the early looks at Blackwell look fantastic. We

0:04:57.160 --> 0:05:00.320
<v Speaker 1>expect almost a two and a half time's gain in

0:05:00.360 --> 0:05:02.640
<v Speaker 1>the compute power that you get from Blackwell that we

0:05:02.680 --> 0:05:04.600
<v Speaker 1>saw from H one hundreds, and so I think it'll

0:05:04.600 --> 0:05:07.520
<v Speaker 1>be a really material jump for customers once we get

0:05:07.520 --> 0:05:09.960
<v Speaker 1>those out, and you know, I think those will be

0:05:10.040 --> 0:05:13.000
<v Speaker 1>early next year, and we're excited about putting them into

0:05:13.160 --> 0:05:14.880
<v Speaker 1>customer's hands, and we'll get them out there as soon

0:05:14.880 --> 0:05:15.640
<v Speaker 1>as they're available.

0:05:16.240 --> 0:05:20.919
<v Speaker 2>Investors, though, they find it wild, this whole frenemy existence

0:05:20.960 --> 0:05:25.000
<v Speaker 2>that's going on. Do you truly think that investors here think, oh,

0:05:25.160 --> 0:05:28.240
<v Speaker 2>we want to see in video dependency as well as

0:05:28.279 --> 0:05:31.680
<v Speaker 2>AWS having its own offerings. Is that something you think

0:05:31.720 --> 0:05:34.680
<v Speaker 2>everyone can swallow? Or do you think ultimately there will

0:05:34.720 --> 0:05:38.080
<v Speaker 2>be a broadening out other than just in video winning all.

0:05:40.320 --> 0:05:42.280
<v Speaker 1>Look, I do think it's a partner if you think

0:05:42.279 --> 0:05:45.480
<v Speaker 1>about AWS. We started from the very beginning thinking about

0:05:45.520 --> 0:05:48.760
<v Speaker 1>this partnership mindset, and we built the entire business around

0:05:48.760 --> 0:05:51.640
<v Speaker 1>AWS thinking about how AWS would have services and our

0:05:51.640 --> 0:05:54.279
<v Speaker 1>partners would have services, and that there's plenty of space

0:05:54.320 --> 0:05:57.040
<v Speaker 1>for all of us to really grow and build our businesses.

0:05:57.480 --> 0:06:02.039
<v Speaker 1>And that is true for software providers, true for service providers,

0:06:02.200 --> 0:06:04.560
<v Speaker 1>and it's true for technology providers. And so I think

0:06:04.600 --> 0:06:07.279
<v Speaker 1>that we've proven time and time again over the last

0:06:07.279 --> 0:06:09.880
<v Speaker 1>eighteen years that AWS can have products and our partners

0:06:09.920 --> 0:06:12.880
<v Speaker 1>can have products, and then as we make them all available,

0:06:13.320 --> 0:06:16.000
<v Speaker 1>that the whole, the whole pie gets bigger. And so

0:06:16.040 --> 0:06:17.800
<v Speaker 1>I think there's plenty of opportunity for both and so

0:06:17.960 --> 0:06:19.960
<v Speaker 1>it really isn't. I think it makes for a fun

0:06:20.040 --> 0:06:22.760
<v Speaker 1>narrative that it's either or. But we're great partners with

0:06:22.839 --> 0:06:24.520
<v Speaker 1>n Video. We will continue to be, And this is

0:06:24.560 --> 0:06:26.080
<v Speaker 1>all about making the pie get bigger.

0:06:26.160 --> 0:06:29.320
<v Speaker 2>We love fun narratives. US journalists talk about that pie.

0:06:30.040 --> 0:06:32.400
<v Speaker 2>When it's to do with large language models. You are

0:06:32.480 --> 0:06:36.760
<v Speaker 2>unveiling Nova, you're really saying that the Nova offering compares

0:06:36.880 --> 0:06:39.040
<v Speaker 2>really well to other offerings out there when it comes

0:06:39.040 --> 0:06:42.520
<v Speaker 2>to multimodi multimodal, but also from a cost perspective and

0:06:42.520 --> 0:06:46.960
<v Speaker 2>efficiency perspective. Why invest so much in LLM provisions when

0:06:46.960 --> 0:06:47.960
<v Speaker 2>you already offer the rest?

0:06:49.400 --> 0:06:52.119
<v Speaker 1>Yeah, and again, this is all about giving customers more choice.

0:06:52.120 --> 0:06:54.720
<v Speaker 1>I think we as Amazon were investing in these models

0:06:54.720 --> 0:06:58.320
<v Speaker 1>because we couldn't find the exact right mix of capabilities

0:06:58.360 --> 0:07:01.039
<v Speaker 1>and customizations and things that we needed internally, so we

0:07:01.120 --> 0:07:03.640
<v Speaker 1>started building them. And as we started building these, we

0:07:03.720 --> 0:07:06.240
<v Speaker 1>saw that these models were actually getting quite good. The

0:07:06.240 --> 0:07:08.920
<v Speaker 1>benchmarks were really good, and we were seeing some good

0:07:08.920 --> 0:07:12.240
<v Speaker 1>capabilities come out of there. And our models are quite

0:07:12.240 --> 0:07:15.400
<v Speaker 1>good at some particular areas. They're very good at executing

0:07:15.400 --> 0:07:19.000
<v Speaker 1>agentic workflows. They're really good at pulling knowledge out of

0:07:19.000 --> 0:07:22.680
<v Speaker 1>a RAG index there and they're very very low latency

0:07:22.720 --> 0:07:25.440
<v Speaker 1>and low cost, and so we think that that capability

0:07:25.480 --> 0:07:27.120
<v Speaker 1>is going to be really compelling for a lot of

0:07:27.200 --> 0:07:29.560
<v Speaker 1>use cases. We found use cases inside of Amazon, but

0:07:29.600 --> 0:07:31.880
<v Speaker 1>even inside of Amazon, we use a mix of different models.

0:07:31.920 --> 0:07:34.200
<v Speaker 1>We use models from Anthropic, we use models from Meta

0:07:34.200 --> 0:07:37.120
<v Speaker 1>and Lama, we use models from a bunch of different sources,

0:07:37.200 --> 0:07:40.640
<v Speaker 1>and so I think that hopefully customers see a lot

0:07:40.640 --> 0:07:43.840
<v Speaker 1>of value from these new Nova models, but we also

0:07:43.880 --> 0:07:45.960
<v Speaker 1>expect that customers are going to combine lots of different

0:07:46.000 --> 0:07:47.360
<v Speaker 1>models in lots of different ways.

0:07:47.560 --> 0:07:51.000
<v Speaker 2>You haven't yet got the access to open ai, and

0:07:51.120 --> 0:07:53.840
<v Speaker 2>for obvious reasons, with their relationship with Microsoft, would you

0:07:53.880 --> 0:07:56.760
<v Speaker 2>think that that will at some point unfold.

0:07:57.440 --> 0:08:00.840
<v Speaker 1>Look, I think my view is anytime customers demand something

0:08:00.880 --> 0:08:03.960
<v Speaker 1>that eventually that sorts its way out. And I think

0:08:04.520 --> 0:08:06.760
<v Speaker 1>open Ai obviously has a great set of models as well,

0:08:06.800 --> 0:08:08.720
<v Speaker 1>and I'm sure many of our customers would love to

0:08:08.760 --> 0:08:11.560
<v Speaker 1>have those available in Bedrock as well, and we'd love

0:08:11.560 --> 0:08:14.200
<v Speaker 1>to support those there. And you know, I think I'd

0:08:14.240 --> 0:08:16.280
<v Speaker 1>look at and we have a long term view on this,

0:08:16.520 --> 0:08:19.120
<v Speaker 1>and over the long term, we'd like to make every

0:08:19.200 --> 0:08:22.480
<v Speaker 1>technology available inside of Amazon and inside of AWS to use,

0:08:22.520 --> 0:08:24.640
<v Speaker 1>and that means all the models that are out there,

0:08:24.680 --> 0:08:26.920
<v Speaker 1>we'd love to offer them at Bedrock. All the software

0:08:26.960 --> 0:08:28.680
<v Speaker 1>that's out there, we'd love to offer and make sure

0:08:28.720 --> 0:08:32.200
<v Speaker 1>that it's available inside of AWS. Every service that everyone

0:08:32.240 --> 0:08:35.040
<v Speaker 1>else in third parties are providing we want to make available.

0:08:35.080 --> 0:08:37.200
<v Speaker 1>And that all goes back to that choice story. And

0:08:37.280 --> 0:08:39.760
<v Speaker 1>so in the fullness of time, I believe that we

0:08:39.800 --> 0:08:43.520
<v Speaker 1>would love open AI models inside of AWS. We absolutely

0:08:43.600 --> 0:08:46.200
<v Speaker 1>it's you know, were we listen to our customers and

0:08:46.200 --> 0:08:49.319
<v Speaker 1>if that's something the customers want, we're absolutely game. Obviously,

0:08:49.320 --> 0:08:51.839
<v Speaker 1>with a partnership, it takes multiple sources and there's probably

0:08:51.840 --> 0:08:54.920
<v Speaker 1>some complexities in there, as you allude to, but I

0:08:54.960 --> 0:08:57.360
<v Speaker 1>think in the fullness of time it likely is the case.

0:08:57.800 --> 0:08:59.520
<v Speaker 2>In the here and now. You spelled out on the

0:08:59.520 --> 0:09:03.000
<v Speaker 2>previous that AI is already a multi billion dollar business

0:09:03.400 --> 0:09:05.360
<v Speaker 2>growing at more than one hundred percent. Can you give

0:09:05.400 --> 0:09:08.959
<v Speaker 2>us any more detailed numbers as to how large AAAI

0:09:09.040 --> 0:09:10.199
<v Speaker 2>offering is for AWS?

0:09:10.280 --> 0:09:14.040
<v Speaker 1>Now, You're right, it is a multi billion dollar business

0:09:14.040 --> 0:09:16.920
<v Speaker 1>for us, and it's really growing rapidly. Take Bedrock, which

0:09:16.960 --> 0:09:20.000
<v Speaker 1>is our platform that lots of customers are building all

0:09:20.000 --> 0:09:24.120
<v Speaker 1>of their production AI applications on top of bedrock has

0:09:24.200 --> 0:09:26.720
<v Speaker 1>grown the number of users five x in the last

0:09:26.800 --> 0:09:29.040
<v Speaker 1>year alone, so it is really a rocket ship growth

0:09:29.040 --> 0:09:31.920
<v Speaker 1>of people building this. I think what's really exciting is

0:09:31.920 --> 0:09:34.080
<v Speaker 1>that people are using bedrock not just to do proof

0:09:34.120 --> 0:09:37.160
<v Speaker 1>of concepts, but they're using it to really deeply integrate

0:09:37.200 --> 0:09:40.960
<v Speaker 1>with their own enterprise data and to launcher production applications,

0:09:40.960 --> 0:09:44.560
<v Speaker 1>which I think is a particularly telling sign of where

0:09:44.600 --> 0:09:45.960
<v Speaker 1>the technology is going, where.

0:09:45.840 --> 0:09:48.240
<v Speaker 2>It's evolved to and clients willing to spend it a

0:09:48.280 --> 0:09:49.880
<v Speaker 2>time of a lot of uncertainty that we've got a

0:09:49.960 --> 0:09:53.360
<v Speaker 2>new administration coming in, do you see any concerns in

0:09:53.440 --> 0:09:58.680
<v Speaker 2>terms of well geopolitics, your road supply chain relationship with TSMC,

0:09:58.800 --> 0:09:59.360
<v Speaker 2>for example.

0:10:01.320 --> 0:10:03.360
<v Speaker 1>No, I mean, I think, look, we've we've worked with

0:10:03.559 --> 0:10:08.000
<v Speaker 1>various administrations all throughout the last eighteen years that AWS

0:10:08.000 --> 0:10:10.080
<v Speaker 1>has been around, and we're really excited to work with

0:10:10.120 --> 0:10:14.120
<v Speaker 1>this administration. You know. I think that everybody is interested

0:10:14.160 --> 0:10:18.280
<v Speaker 1>in ensuring that we have technology continuity and and there's

0:10:18.320 --> 0:10:20.839
<v Speaker 1>a ton of opportunity here like that this is how

0:10:20.880 --> 0:10:23.600
<v Speaker 1>we drive the economy, and I think all administrations are

0:10:23.600 --> 0:10:26.160
<v Speaker 1>interested in continuing to make sure that we're able to

0:10:26.240 --> 0:10:29.400
<v Speaker 1>drive the economy forward, and and AWS is a big

0:10:29.520 --> 0:10:31.960
<v Speaker 1>enabler of that, and so we're excited to work with

0:10:31.960 --> 0:10:35.640
<v Speaker 1>the new administration. I think we're we're obviously anxiously watched

0:10:35.679 --> 0:10:37.880
<v Speaker 1>all of the geopolitical events all around the world to

0:10:37.880 --> 0:10:40.320
<v Speaker 1>make sure that we're trying to insulate our customers as

0:10:40.400 --> 0:10:44.319
<v Speaker 1>much as possible from any eventuality. But but you know,

0:10:44.360 --> 0:10:48.040
<v Speaker 1>I think we're we're cautiously optimistic that we're in a

0:10:48.080 --> 0:10:48.640
<v Speaker 1>good spot.

0:10:48.800 --> 0:10:50.720
<v Speaker 2>You might have been cautiously watching what was happening in

0:10:50.720 --> 0:10:54.560
<v Speaker 2>South Korea today amid your own event any exposure there.

0:10:54.400 --> 0:10:58.400
<v Speaker 1>Matt, Sorry I missed that I was on stage, so

0:10:59.520 --> 0:11:01.440
<v Speaker 1>just I'm from the stage right here to talk to you.

0:11:01.520 --> 0:11:04.720
<v Speaker 2>So okay, I understood, But that in terms of the

0:11:04.760 --> 0:11:08.160
<v Speaker 2>South Korean political destabilization, nothing that was a worry there.

0:11:09.320 --> 0:11:11.200
<v Speaker 1>No, I mean, look, we have a we have a

0:11:11.320 --> 0:11:13.679
<v Speaker 1>region in South Korea, and we have many of our

0:11:13.720 --> 0:11:17.120
<v Speaker 1>customers that operate in South Korea and and will continue

0:11:17.120 --> 0:11:20.440
<v Speaker 1>to support them and and operate with with with any

0:11:20.600 --> 0:11:23.000
<v Speaker 1>with ever whichever government is in operation.

0:11:23.960 --> 0:11:26.120
<v Speaker 2>We thank you so much for your time rushing off

0:11:26.120 --> 0:11:28.680
<v Speaker 2>a stage A w S ce O Matt Garman