WEBVTT - Siemens CEO Roland Busch Talks AI

0:00:02.520 --> 0:00:07.040
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio news.

0:00:14.480 --> 0:00:17.800
<v Speaker 2>AI is moving beyond chatbots and onto the factory floor,

0:00:17.920 --> 0:00:22.280
<v Speaker 2>helping manufacturers boost productivity, with industrial AI becoming a bigger

0:00:22.320 --> 0:00:25.160
<v Speaker 2>part of the conversation. Joining us now is Semens CEO

0:00:25.480 --> 0:00:28.680
<v Speaker 2>Roland Bush, who recently met with European Commission President Ursula

0:00:28.720 --> 0:00:32.000
<v Speaker 2>Vonderleon and other tech leaders to discuss AI's role in

0:00:32.080 --> 0:00:34.560
<v Speaker 2>Europe's industrial future. But I would note as well, Roland,

0:00:35.400 --> 0:00:37.960
<v Speaker 2>you were the first conversation I had this year. We

0:00:38.040 --> 0:00:41.600
<v Speaker 2>started twenty twenty six with Jensen Wang talking about the

0:00:41.640 --> 0:00:46.000
<v Speaker 2>transition into the physical world. Just very simply to start.

0:00:46.560 --> 0:00:50.840
<v Speaker 2>Why Semens is bedding so much on industrial AI.

0:00:52.360 --> 0:00:57.400
<v Speaker 1>Actually, big ass Siemens is really geared for this moment

0:00:57.680 --> 0:01:02.280
<v Speaker 1>to bring AI into the real world because its industrially.

0:01:02.280 --> 0:01:05.319
<v Speaker 1>I why you need a couple of things. Number one

0:01:05.400 --> 0:01:08.200
<v Speaker 1>is you need a technology stack, and the stack includes

0:01:08.240 --> 0:01:11.960
<v Speaker 1>hardware and software. Both is super relevant. Number two is

0:01:12.160 --> 0:01:14.319
<v Speaker 1>domain or how we know how to build things we

0:01:14.319 --> 0:01:16.800
<v Speaker 1>are on the shop floor. The next one is data.

0:01:17.040 --> 0:01:19.600
<v Speaker 1>We have a lot of data, our own data, data

0:01:19.640 --> 0:01:23.200
<v Speaker 1>shared with our customers and partners, and we have the

0:01:23.280 --> 0:01:27.080
<v Speaker 1>trust of our customers. Because when the AI hits the

0:01:27.120 --> 0:01:31.000
<v Speaker 1>real world. Pallucination is not an option, so you need

0:01:31.240 --> 0:01:34.200
<v Speaker 1>hardcore results. They should work. So we have all this

0:01:34.760 --> 0:01:36.240
<v Speaker 1>and we can bring it now to the real world

0:01:36.280 --> 0:01:38.919
<v Speaker 1>and husband at least we have great partners like in media,

0:01:39.000 --> 0:01:42.760
<v Speaker 1>like some others who are with us in making this

0:01:42.880 --> 0:01:44.839
<v Speaker 1>transformation for our customers.

0:01:45.600 --> 0:01:48.120
<v Speaker 2>I asked you at CES in January, give me a

0:01:48.160 --> 0:01:52.040
<v Speaker 2>case study, an example. There's some evidence of AI in

0:01:52.080 --> 0:01:55.280
<v Speaker 2>a factory anywhere in the world. Fast forward to June.

0:01:55.800 --> 0:01:58.920
<v Speaker 2>Is it real now? Is there to please? Can point

0:01:58.920 --> 0:01:59.800
<v Speaker 2>to please?

0:02:00.160 --> 0:02:03.640
<v Speaker 1>It's really now, and we can talk about the design face,

0:02:03.920 --> 0:02:07.080
<v Speaker 1>the manufacturing face, and the operation phase when your assets

0:02:07.080 --> 0:02:10.120
<v Speaker 1>are in the field. But you talked about an example

0:02:10.160 --> 0:02:14.320
<v Speaker 1>on the on the shop floor. Actually, we now launched

0:02:14.360 --> 0:02:17.480
<v Speaker 1>our iding engineering agent. What is it. It is an

0:02:17.520 --> 0:02:21.960
<v Speaker 1>agent which programs an industrial PC for you. So if

0:02:21.960 --> 0:02:24.600
<v Speaker 1>you have a cutting chop on the shop floor and

0:02:24.639 --> 0:02:28.240
<v Speaker 1>the cutting is not precise anymore, now an engineer has

0:02:28.280 --> 0:02:31.400
<v Speaker 1>to go there reprogram your PLC and so on. Iing

0:02:31.440 --> 0:02:34.639
<v Speaker 1>engineering agent does it for you. You say, I mean,

0:02:35.040 --> 0:02:38.799
<v Speaker 1>my cutting machine is not person they're not working precisely.

0:02:39.760 --> 0:02:43.320
<v Speaker 1>Iron engineering agent analyzes the problem, breaks the problem down,

0:02:43.440 --> 0:02:46.760
<v Speaker 1>looks for all the data, the machines the job which

0:02:46.800 --> 0:02:52.520
<v Speaker 1>has to be done, analyzes it, develops a code, runs it,

0:02:52.840 --> 0:02:56.040
<v Speaker 1>compiles it. If it doesn't work, it compiles it over again,

0:02:56.120 --> 0:02:59.000
<v Speaker 1>fixes it until you can say now go and you

0:02:59.080 --> 0:03:02.760
<v Speaker 1>push a bottom and then PLC is running with the

0:03:02.800 --> 0:03:06.320
<v Speaker 1>same position as before. So we are brought about fifty

0:03:06.360 --> 0:03:11.600
<v Speaker 1>percent higher productivity while the quality of your programming is

0:03:11.639 --> 0:03:14.400
<v Speaker 1>increasing by eighty percent. So this is real factor. And

0:03:14.440 --> 0:03:17.800
<v Speaker 1>now we're introduced using another function where electrical design comes

0:03:17.800 --> 0:03:21.040
<v Speaker 1>also into this into the place. Well, there's a real

0:03:22.120 --> 0:03:26.239
<v Speaker 1>and this also not available in the market. So far

0:03:26.760 --> 0:03:28.760
<v Speaker 1>we brought it to the market and ready to expand

0:03:28.760 --> 0:03:29.919
<v Speaker 1>to other use cases.

0:03:29.639 --> 0:03:33.960
<v Speaker 2>To either by geography or bi sector. Is there a

0:03:34.080 --> 0:03:38.000
<v Speaker 2>market where there's just more speed here? You know, we

0:03:38.280 --> 0:03:42.760
<v Speaker 2>reflected on your meeting with European Commission. Notoriously red tape

0:03:42.760 --> 0:03:44.720
<v Speaker 2>can get in the way. What are you seeing in

0:03:45.000 --> 0:03:46.360
<v Speaker 2>different markets around the world.

0:03:47.240 --> 0:03:50.000
<v Speaker 1>I mean the hardest market as we can imagine currently

0:03:50.080 --> 0:03:52.760
<v Speaker 1>is AI factories in data centers. And yeah, we are

0:03:52.800 --> 0:03:54.840
<v Speaker 1>really in the design phase together. It's in DDIA. We

0:03:55.280 --> 0:03:58.400
<v Speaker 1>create a blueprint a white paper for how an a

0:03:58.520 --> 0:04:01.040
<v Speaker 1>high factor in the future looks like guess what it's

0:04:01.080 --> 0:04:05.240
<v Speaker 1>based on a digital twin, and this digital twin is

0:04:05.240 --> 0:04:09.120
<v Speaker 1>optimized with AI obviously, But if you talk about other markets,

0:04:09.160 --> 0:04:12.800
<v Speaker 1>for example, the fast growing markets which need really need

0:04:12.840 --> 0:04:16.920
<v Speaker 1>productivity and speed. Semiconductors, a lot of investment going there,

0:04:17.040 --> 0:04:21.200
<v Speaker 1>a lot of VABs bills, pharmaceutical there's a lot of

0:04:21.200 --> 0:04:24.800
<v Speaker 1>investment also going to the United States and aerospace and defense.

0:04:24.960 --> 0:04:29.160
<v Speaker 1>We are ramping up our capabilities to produce stuff. So

0:04:29.200 --> 0:04:31.560
<v Speaker 1>there are a lot of sectors. And then last part

0:04:31.600 --> 0:04:33.599
<v Speaker 1>of lease, I mean the car industry is also under pressure.

0:04:33.640 --> 0:04:37.560
<v Speaker 1>They need to evolve faster, having faster cycles, being more

0:04:37.560 --> 0:04:40.080
<v Speaker 1>flexible in their production. So I can go on and on, and

0:04:40.760 --> 0:04:43.000
<v Speaker 1>the topic is always the same. You will be faster

0:04:43.279 --> 0:04:47.640
<v Speaker 1>in your design process using AI, coming from simulation which

0:04:47.800 --> 0:04:51.560
<v Speaker 1>verifies to simulation which creates new designs, and all the

0:04:51.560 --> 0:04:54.560
<v Speaker 1>way in the factory, way down to the point that

0:04:54.600 --> 0:04:56.800
<v Speaker 1>you were deploying more and more robots, and guess what

0:04:57.000 --> 0:05:00.919
<v Speaker 1>robots are also AI based automation devices.

0:05:01.839 --> 0:05:05.120
<v Speaker 2>Semens has an almost one hundred and eighty year history

0:05:05.800 --> 0:05:09.479
<v Speaker 2>as an industrial company, an engineering company. You know the

0:05:09.520 --> 0:05:11.560
<v Speaker 2>sense I've had from you, Roland in the times we've

0:05:11.560 --> 0:05:14.640
<v Speaker 2>spoken over the last twelve months and more is you

0:05:14.720 --> 0:05:18.880
<v Speaker 2>are positioning the Semens as a technology company. How will

0:05:18.920 --> 0:05:23.160
<v Speaker 2>your investments, your attitude towards M and A reflect that.

0:05:23.560 --> 0:05:26.560
<v Speaker 2>And when you think about the market of today, the

0:05:26.640 --> 0:05:29.840
<v Speaker 2>volatility we see in the technology sector, do you think

0:05:29.920 --> 0:05:33.279
<v Speaker 2>that Semens is being treated and valued as a technology company.

0:05:34.200 --> 0:05:37.680
<v Speaker 1>Well, the short answer to the last point not yet,

0:05:37.720 --> 0:05:40.839
<v Speaker 1>but we're working working on it. But the ons the point.

0:05:41.400 --> 0:05:44.600
<v Speaker 1>Over the last fifteen years or so, we invested thirty

0:05:44.680 --> 0:05:48.359
<v Speaker 1>billion dollars in building up our software suite. It's the

0:05:48.400 --> 0:05:53.160
<v Speaker 1>most comprehensive digital twin can be built with Semen's technology,

0:05:53.960 --> 0:05:57.720
<v Speaker 1>and it's a physics based digital twin with a fully

0:05:57.800 --> 0:06:02.360
<v Speaker 1>realistic representation using a media technology for example, and that's unique.

0:06:02.920 --> 0:06:05.080
<v Speaker 1>All in all, we have nine point four billion in

0:06:05.120 --> 0:06:09.040
<v Speaker 1>twenty twenty five, nine point four billion euros in digital

0:06:09.160 --> 0:06:12.239
<v Speaker 1>revenue including software. We are going to double it until

0:06:12.240 --> 0:06:15.400
<v Speaker 1>twenty thirty so and this is a fast growing, high

0:06:15.480 --> 0:06:19.320
<v Speaker 1>margin business and it's now even super charged with AI.

0:06:19.760 --> 0:06:22.120
<v Speaker 1>We rewrite our software so it can be used by

0:06:22.120 --> 0:06:25.280
<v Speaker 1>engineers and agents. So that means you can see higher

0:06:25.279 --> 0:06:28.799
<v Speaker 1>growth rates because you can democratize it and the accelerating

0:06:28.839 --> 0:06:33.400
<v Speaker 1>simulation and again we are supercharging our operation software on

0:06:33.480 --> 0:06:36.880
<v Speaker 1>the shop floor. We say AI technology, so yes, and

0:06:36.920 --> 0:06:39.440
<v Speaker 1>we have that advantage again to combine the real inted world,

0:06:39.640 --> 0:06:41.920
<v Speaker 1>you need to both. Hardware is more important than ever.

0:06:42.360 --> 0:06:45.200
<v Speaker 1>I think about the data centers built, think about how

0:06:45.520 --> 0:06:48.120
<v Speaker 1>how robotics school there is through the market to the

0:06:48.120 --> 0:06:51.680
<v Speaker 1>shop floor. So therefore this combination is unique and we

0:06:51.760 --> 0:06:54.600
<v Speaker 1>keep on going investing. It's a transformation. You mentioned it

0:06:54.640 --> 0:06:57.960
<v Speaker 1>almost on eady the years then the fastest transformation or history,

0:06:58.560 --> 0:07:01.720
<v Speaker 1>and we are on the right track. But it requires

0:07:01.839 --> 0:07:05.200
<v Speaker 1>again I was super fast adoption, but also for our customers,

0:07:05.200 --> 0:07:06.120
<v Speaker 1>which we are supporting.

0:07:07.240 --> 0:07:10.040
<v Speaker 2>Fresh from one of Europe's biggest keynotes, fresh from meeting

0:07:10.120 --> 0:07:13.400
<v Speaker 2>with Europe's most important leaders, Roland versch seman CEO back

0:07:13.440 --> 0:07:15.360
<v Speaker 2>on Bloomberg Tech, thank you very much. Indeed,