Q&A with Siemens’ CEO Tony Hemmelgarn reveals a central theme.
CEO Tony Hemmelgarn took the stage as CEO of Siemens PLM Solutions in New York City at the Siemens Media and Analyst Conference. By the time he left the stage, he was CEO of another company, Siemens Digital Industries Software. You can call us Siemens Software, says Hemmelgarn. He was still high on Mendix, acquired last October for $730 million. Mendix is to bridge Siemens PLM application, Teamcenter with MindSphere, with NX, other Siemens applications, as well as ERP solutions, databases, data lakes…whatever—a “low code” that would allow engineers to access all sorts of data, whether it be on local drives, in databases, on the cloud, streaming… whatever. He also announced Xcelerator, which is a little harder to explain.
A press conference was held on the last day of the conference. Highlights include Hemmelgarn’s favorite example of data analysts gone bad, correlating mozzarella cheese consumption with civil engineering degrees. And much more. The Q&A wanders, but from Hemmelgarn, one theme emerges: substance over hype. Yes, new technology is hyped up, but it must make business sense.
Hemmelgarn: The market is pushing us on additive manufacturing. I admit that for a few years we have been behind, but that is not the case anymore. Five or more years ago, there was a lot of hype about everyone having a printer in their house. Break a coffee cup and 3D print a new one rather than going out to buy one. I thought that was ridiculous. Everybody was not going to have a printer.
It was the height of optimism. But even today, people still do not understand what it means to manufacture for 3D printing. We see a duct (above) from a HP 3D printer. Six parts made as one, 22 percent increase in airflow. You wouldn’t have designed that kind of part in a million years. You need CFD [computational fluid dynamics] for simulation. You need those kinds of discussion. We’re still talking about printing coffee cups.
I understand new technology gets hyped up. Then reality sets in and the hype slows. We have to be ready if it takes off again. Same thing for IoT. It’s mostly all hype initially, then it adjusts, then it will take off. You want to be ready.
Same with AR/VR. We see a lot of things that are sexy and cool in AR/VR, but they can’t provide any business value and they’re not functional. AR/VR from us, our competition, is definitely cool, but is it realistic? Can you put it in production mode and get business from it? But again, you have to be ready for it when the market is there.
Q: You mentioned the technologies Siemens is offering, the cloud, in particular, was putting a lot of stress on Siemens’ customers. Is Siemens too far ahead of what its customers are asking for?
Hemmelgarn: We’re not pushing our customers. We will let them decide. It is up to them where they want to go. We have all kinds of offerings. We can stay with perpetual licensing. We can do cloud. We could do hosted cloud. All kinds of things. We’re not going to our customers and saying, You need to switch to a subscription or switch to cloud.
Three years ago, everyone was saying “I’m not going to put intellectual property in the cloud.” But now people are not as concerned. We have to be ready for what our market demands. We’re not going to force our customers into a business model they’re not comfortable with.
I don’t believe you can have a closed market. You’ve got to be open. I believe strongly that you can’t force business models on your customers that don’t make sense to them, that they don’t want. It’s not right. A year-and-a-half ago, a company was going to throw out their existing vendor and come to us. The CIO said he “wanted to make sure we’re looking at different ways of leveraging software” and he didn’t want to have big capital outlay in front. He was referring to subscription, the cloud, etc. I told him we can support all the applications on the cloud, but was he comfortable putting his data on the cloud? The head of engineering, sitting right next to him, jumped in with “Absolutely not!” Clearly, the company was not aligned.
When they are ready, we will have to be ready, too. They could speed up or slow down. And we would not push them.
Q:It’s encouraging to hear about MBD and MBDE (model-based design and model-based design and engineering), and PMI. Siemens has those tools. What percent of people actually use those tools?
Hemmelgarn: PMI has been used fairly aggressively for a long time on a number of our accounts, but there’s level of use with PMI. Using PMI on a model just to be able to document it has been around for a long time. Using PMI across the full design process and committing to manufacturing and inspection… that’s not as widely used.
I would say the OEMs are not pushing for MBD aggressively. We work with them on it. In aerospace, certain channels have been very effective with MBD. Outside of aerospace, it is less used.
Q: What are the barriers to MBD and PMI?
Hemmelgarn: Some of the automotive companies went down the MBD and PMI path. Aerospace had forgotten more about the subjects than automotive knew. Automotive was looking to learn from aerospace. Once an automobile is built, the auto company needs to make sure it was done right and it had to meet requirements. Aerospace had already been doing that and was much more advanced with their model-based approach, model-based engineering and systems modeling.
It is just process, methodology and understanding. We’re seeing across the board the managing of requirements and specifications that drive product design and provide traceability.
It’s interesting what we are seeing in the integrated circuit space with Mentor. We had no idea requirements management would be needed with ICs.
Q: As regulatory agencies globally are working on safety regulations for more automated vehicles and automated driving systems, what do you see as Siemens’ role?
Hemmelgarn: We have many tools that help on several fronts. Teamcenter can ensure regulations will be followed.
Q: And as they evolve, as new standards or requirements appear, you’ll be incorporating that into the system?
Hemmelgarn: Generally, we do. We may get into trouble is when there’s not a clear standard. The great thing about standards is there’s so many of them. [draws laughter].
Q: Can you tell more about the announcement with Bentley Systems and Capital Asset Lifecycle Management (CALM)?
Hemmelgarn: We all know Siemens has an investment in Bentley. The process automation group owns the Bentley relationship. Customers were asking for an enterprise data management for process. We didn’t have a solution. Our people were suggesting Teamcenter, but I thought we were missing a solution. How to manage a complex set of documents, data and processes? You have to create and understand a plan and then do configuration control with it.
But many of these sites don’t use Bentley. They’re using other systems. To make everything work together, we came to this approach.
Interestingly, we were probably one of the Bentleys’ first customers back in the mid-80s. I knew them very well when they were at Intergraph and then after that, when Bentley was spun off from Intergraph.
We used a product called PseudoStation. It was an emulator that ran on a 286 PC, so you didn’t have to buy a $75,000 Intergraph workstation. Bentley started with PseudoStation. It was used for drawings mostly. They went after the market of people using Intergraph but couldn’t afford to buy the expensive workstations.
When I say that there’s a big difference between doing work for data management for your application and doing enterprise data management with something like Teamcenter. Take Tecnomatix, for example. We bought Tecnomatix 11 years ago. A very good application. But they hit a wall because customers were still on the board and didn’t need the database; they had Teamcenter, an enterprise tool. They didn’t need a worker tool that manages individual data.
We were hearing similar comments from customers in the process industry. They needed an enterprise tool to be able to manage it all. It wasn’t about just managing files, checking them in or out. It was about all the data associated with building a plan, managing a plan, and so on. This is an example of where customers probably pushed us as much as anything else saying, “Look, we know what Teamcenter can do for the process industry.”
Q: We’ve heard a lot about digital simulation and digital twins. How does Siemens view that advancement in technology in relationship with the skills gap in the engineering workforce?
Hemmelgarn: I think about digital processing and what we do and there’s between us and the workforce. There are ways for the workforce to catch up. There are companies in emerging countries, and with their manufacturing processes, they are thinking about leapfrogging current leaders. The Germans, for example, are extremely good at manufacturing. But additive manufacturing is a new playing field. We’re all starting from scratch. Suddenly, those 100 years of manufacturing expertise don’t mean a hill of beans in additive. In fact, that experience can be a detriment.
Our customers say, “I’ve already got the parts. I just want to print them.”But what’s the point of printing a part that was designed the old-fashioned way? You saw the shapes we’re coming out with like this 3D-printed part (above). If you’re not designing your parts specifically for additive, you’re kidding yourself. Additive is a great equalizer. And everybody’s learning at the same time.
Now, about skills in general. There’s this belief [that] with enough data you can make decisions, but I argue that lots of data doesn’t necessarily mean the right decision. If you have the right tools, you can make the decisions a lot faster and a lot easier.
There’s going to be more about companies willing to change their processes than people changing their tools.
You can get into trouble by going faster—while keeping everything the same.
Let me give you a CAE example. CAE people are trained using one tool in school, and they will never switch. A customer was analyzing undersea piping equipment. They were designing a valve, and the design tradeoffs were flow rate versus erosion. Too fast of a flow rate and you will have erosion in the valve. We used a design optimization tool, coupled with our CAE tool, coupled with NX, and found 300 different solutions of the valve in the time it would have taken them to do one valve the traditional way.
Still, they wouldn’t switch. We had to go to their management and show them the example and tell them [how] disruptive it would be if the competition was able to do 300 design iterations in the time it took for their guys to do one. How long before their business is disrupted?
Back to the skills gap. There’s a skills gap in the workforce between new engineers and the older ones, and between those who adopt new tools and methods and those who are used to a certain way of doing things. For example, what happens to you if your simulation tool cuts down your simulation time exponentially? Would you encourage people to go into a field where AI with machine learning is taking over? Do you need an engineer? I know the answer is yes, but what will be the future engineer’s role?
CAE is just a tool to be applied by an engineer. Same with AI. Same with data. Like I’ve said before, with enough data, I can show a correlation between the number of people getting civil engineering degrees and the consumption of mozzarella cheese. You and I know that makes no sense whatsoever.
If you are doing design iterations and the program comes back with something ridiculous, the average person may not know it’s ridiculous, but an engineer would know. Engineers have those skills. Why not leverage those skills with other tools to help you decide?
Digital Twin
A digital model with its digital twin is a mechanism. You can validate if you are in the right place because you’ve got the visual model. You can play it back and you can see what’s going on. A digital twin is being underestimated for IoT right now. It’s a very valuable. We show a digital twin and light bulbs start going off. They realize it’s more than just maintain a machine because anybody can tell you you’ve got a maintenance problem. You don’t need IoT to tell you when a machine is overheating. But what if it could take the data and feed it back and start running simulations and say why is it overheating?
Having a digital model that’s fully representative is very helpful. Less than fully represented—you’re limited. You can’t represent electronics assemblies. Like the Polar watch. Imagine trying to design that watch and you could only do the mechanical and the industrial design.
I don’t think you guys are giving us enough credit for the electronics applications.
Q: Siemens filled the gap in electronics with the Mentor acquisition. Earlier, you announced the acquisition of Mendix, which filled a gap in your software portfolio. Do you see other gaps in the Siemens Software portfolio?
Hemmelgarn: The most limiting factor in the adoption of IoT is not enough apps. There’s lots of IoT platforms, but where are all the apps? How could we help our customers develop apps faster so they could value from IoT? That’s how we got interested in Mendix. Then we saw what they were doing with other applications. Why not use Mendix to augment Teamcenter, for example.
When we did our due diligence, we saw that Mendix was used by SAP. So, SAP had similar problems to us with IoT but with enterprise applications. Similar to Teamcenter, you don’t want to tweak Teamcenter and break something that’s running the operations of your business.
So, SAP had contacted Mendix to help them make apps for their users. Companies using SAP did not want a third party coming in and messing with the enterprise software. They’d rather have their customers build their own solutions using Mendix to extract data from SAP.
We thought we could do that for a lot of our enterprise applications.
Q: Thanks for saving engineers from having to code. I know you can’t reveal what acquisitions are next, but do you see other gaps in the Siemens product portfolio?
Hemmelgarn: There’s lots of areas we can go. There’s service. You heard us partnering with IBM. We considered building our own service. Or we could acquire a company. But there’s two major players in this space: IBM and SAP. Here we filled in a gap a different way: with a partner that is strong in that space.
There’s other areas. We’re having many conversations. We want to keep building out our additive capabilities because we see the market going there. We want to continue to fill in electrical and electronic design. Product design is happening in a complex environment. We’re thinking about IC design, too.
Q: But is there a major acquisition, a gigantic one like a Mentor in front of us right now?
[Tony indicates there is not].
Q: Everybody’s talking AR/VR and AI. What are you seeing and hearing?
Hemmelgarn: We will continue to grow in machine learning and AI and AR/VR,but, again, there are realistic applications versus cool demos. I challenge our guys all the time to not show cool demo where we know the technology is still shaky. Let’s wait until the technology is doing everything we want it to do and you can really apply it.
We’ve been talking about PMI for a long time. We’re leveraging PMI and planning to go a lot further. This is how I push my guys. How do we keep leveraging the technology and bring it across the board in everything we do? The crazy thing about our industry is that even when you think something’s mature, like CAD, which we’ve been doing CAD for 35 years, out comes synchronous technology. We came out with convergent modeling, modeling with faceted geometry, additive manufacturing… many things that change the whole thought process of how you might design and do things.
We had a very good presentation about a year ago by Ford about what a digital twin means. A digital twin is not new. We’ve been saying that for many, many years.
We used to create a book with all the issues we’d run into when we made a car. Today, the car goes together without any issue. That’s because now every nut, bolt, screw of that vehicle has been planned through a process and built into a model to be able to validate it’s going to go together exactly the way Ford wants. But now, let’s also get the vehicle characteristics. Will the Ford accelerate, brake and handle like you expect? How’s the ride and handling? How do I simulate more of the functional characteristics? That’s an area that we’ll continue to grow out with products. Not just automobiles, but any system because we want to know how it’s going to behave—not just is it going to snap together.
Q: I’m having a hard time getting my mind around Xcelerator. Is that a new product?
Hemmelgarn: I need to bring the marketing folks in here. I told them, I have 22,000 people in this organization and every one of them asked that question?
Let me try to answer. These days every product is claiming to be a platform, but not everything can be a platform. The NX guys call NX a platform. Teamcenter is a platform. Mendix is a platform. Not everything can be a platform.
Q: Then what is a platform? What have you defined it as?
Hemmelgarn: We did long studies on this to come up with the Xcelerator name, and it’s starting to resonate. I think we are recognized as having the most comprehensive digital platform.
But what does it mean? We said we exist to Xcelerator product development for our customers.Xcelerator is the name for our full suite of products and services—everything we have and do is part of Xcelerator. Mendix is a key part of Xcelerator. Using Mendix, extending Mendix as an application platform development tool that allows us to use the domain knowledge to develop app services for people to create solutions to drive and to augment IoT like you saw with some of the demos of existing products. Cloud applications will be created using this development platform.
Xcelerator is our portfolio. It’s not a product. You don’t go buy Xcelerator.
We couldn’t wait for you guys to come up with a better name than PLM. PLM doesn’t work for us anymore.
Q: PLM was practically your middle name, though.
Hemmelgarn: Somebody might get three cents every time we say PLM.
Q: So, Xcelerator is the portfolio and Mendix is a product?
Correct, says Hemmelgarn, happy to have us finally understand what he announced on the first day.
With that, the press conference concludes.