Good morning, ladies and gentlemen. Thank you for standing by, and welcome to PTC's Frustum acquisition webcast. During today's presentation, all parties will be in a listen-only mode, and the meeting is being recorded. Following the presentation, the webcast will be open for questions. I would now like to turn the webcast over to Tim Fox, PTC's Senior Vice President of Investor Relations. Please go ahead.
Good morning, everyone, and thank you for joining PTC's webcast to discuss the acquisition of Frustum. On the webcast today are Jim Heppelmann, Chief Executive Officer, Steve Dertien, Chief Technology Officer, and Andrew Miller, Chief Financial Officer. We've posted today's press release on our investor relations website, and we'll be putting a replay of this webcast on the website as well. During today's webcast, PTC will be making forward-looking statements, including statements regarding our expectations about the integration of Frustum into our portfolio, the value customers will be attributing to the Frustum functionality, and the effect of the acquisition on future financial results. These forward-looking statements are based on information as of today, November 20th, 2018, and involve risks and uncertainties. Actual results may differ materially from our expectations.
For a discussion of these risks and uncertainties, you should review the forward-looking statements disclosure in today's press release, as well as PTC's SEC filings. With that, I'd like to turn it over to PTC's CEO, Jim Heppelmann.
Great. Thanks, Tim. Good morning, everyone, and thank you for joining us. As you may have noticed, earlier this morning, PTC issued a press release announcing the acquisition of Frustum for $70 million. This acquisition represents an important strategic move in our Creo business, which is our largest and most profitable. I'd like to review the rationale that led us to the acquisition. Along the way, I'll shine a light on our Creo strategy and then introduce you to a series of breakthroughs in generative design and volumetric modeling that we believe will elevate Creo to be a leading position in the world of design and simulation. Let's look at today's news, plus the June news about Ansys, and put them into the context of our larger CAD strategy.
About 30 years ago, PTC unleashed a massive disruption when we launched Pro/ENGINEER, the first viable 3D CAD solid modeling product. We built a tremendous business around Pro/ENGINEER. After a decade or so of being the only game in town, a handful of companies, today's well-known competitors, essentially copied the Pro/E concept more or less verbatim. The market quickly became very crowded and competitive. The combination of product parity and high switching costs caused a stalemate to ensue. In the two decades that followed, virtually all technology innovation across the entire CAD market has been of the incremental form. All major CAD systems today are just broader and prettier imitations of the original Pro/ENGINEER software that PTC released in 1987. After a long period of incremental change, the CAD market is entering a renaissance period, thanks to a series of recent new technological breakthroughs.
I'm referring to artificial intelligence that allows design alternatives to be generated automatically, to volumetric modeling that allows super complex geometries to be modeled and printed for additive manufacturing, to IoT that enables digital CAD models to pair with real-world products and become their digital twin, to augmented and virtual reality, which allows CAD models to come to life in the real world that exists beyond the computer monitor, and to cloud and high-performance computing that allows generative and simulation capabilities to happen in real-time to keep up with the natural pace of the design engineer. These new technologies could be disruptive, only to those who can solve for the switching costs. Said differently, none of these breakthroughs will disrupt anything if they don't somehow incorporate the massive foundations of today's mainstream CAD systems.
Keep in mind that the development of these systems is measured in man-centuries rather than man-years. There simply is no clean sheet of paper that would allow established CAD users to start over with a new tool because their companies need to evolve the products that they already have in the market. They need to continue to leverage the decades of engineering investment they've made in those products while incorporating and evolving toward broader use of powerful new technologies over time. When they look at point solutions that claim to be best in class at additive or cloud or generative design, as impressive as these point solutions may be, there is no realistic option to switch. To try to use multiple separate CAD tools in parallel creates a broken tool chain filled with data loss and manual rework that destroys productivity rather than enhancing it.
The reason why one CAD startup after another has failed, even when funded with massive amounts of VC money, is because they can't possibly get customers to switch to their shiny new toy. Practically speaking, manufacturing companies need the new innovations to come to life inside the mainstream tool that they're so dependent on, not in another product parked next to it. That's why PTC has been working so hard to build this new generation of disruptive technologies directly into Creo. It's why we're acquiring Frustum, who's best in class at AI-driven generative design and volumetric modeling. Why we will incorporate Frustum into Creo as fast as we can. It's why we partnered with Ansys, who is best in class at simulation. We're about to release the Discovery Live technology built into Creo in December.
It's why we've already delivered our first round of ThingWorx IoT capabilities built into Creo and the first tranche of Vuforia AR/VR capabilities built into Creo as well. PTC is on a path to deliver future generations of Creo that truly are disruptive because they combine the most powerful new technologies with a proven foundation that's deep and wide so that the new capabilities can readily be adopted. Our existing customers know that with Creo, you can open a design you made with Pro/E decades ago and now incorporate technologies like generative, real-time simulation, IoT, AR, additive, and even direct modeling into the next version of the design that you save. Adoption of new technologies in Creo is fast and easy. No switching costs, no broken tool chain, just enable the right license code. Companies beyond the Creo install base can benefit from these powerful capabilities, too.
In a consolidation scenario where companies want to standardize on a single platform, the advanced technologies of Creo are very attractive, and it's generally accepted that Creo has the breadth and depth of power to replace any mainstream CAD system, and it represents a big upgrade when replacing low-end systems. Plus, in any scenario where the company or the customer wants to retain their incumbent systems, yet explore these powerful new technologies, Creo has a unique capability to coexist with the incumbent system, thanks to our popular Unite technology that allows Creo to work well in a mixed CAD environment. Let's switch gears and take a deeper look at what Frustum adds to the Creo picture, how it complements Ansys, and how it will come to life inside of Creo.
Keep in mind that a primary goal of any CAD user is to get efficient designs that are validated to work to market as quickly as possible. The traditional approach has been to design in a CAD system like Creo, and when the design is finished, to validate it in a simulation system like Ansys. That approach is powerful, but because it pushes simulation to the end of the process, a significant amount of rework is necessary if problems are discovered in the validation phase, as they frequently are. When Ansys leveraged GPUs and other high-performance computing or HPC techniques to develop their new real-time simulation capability called Discovery Live, PTC jumped on the opportunity to build it into Creo because the ability to pull simulation upstream and apply it to each and every design change is a huge breakthrough.
It means that the user is guided at every step along the way as they iterate, and every part will have been simulated hundreds or even thousands of times before we even get to the validation phase. The odds of finding major problems late in the game drop precipitously. Of course, it would be better yet if the software didn't just follow the engineer through iterations but could engage even earlier to proactively generate a design that could be used as a starting point or perhaps generate a series of different approaches to pick from. That breakthrough is called generative design, and it's precisely what Frustum does. When using Frustum technology, the engineer provides a description of the design requirement in the forms of loads and constraints, the space available, one or more material options, and one or more manufacturing processes such as casted, forged, or additive.
The user can even factor in non-engineering parameters related to supply chain or cost. The Frustum AI engine fires up and gets to work using a high-performance computing technique to quickly generate a range of design alternatives for the engineer to select from. When the user selects their preferred approach, we've completed the generate phase, and we move on to the iterate phase. In the iterate phase, the engineer would combine the generated part or system with other parts and systems, whether generated or manually authored, and iterate through changes, collaborating with numerous other stakeholders along the way. At each step in the iteration, the Ansys Discovery Live technology will help the Creo users understand the implication of each and every change, and in so doing, guide them quickly toward a robust solution.
When the design is complete, the Ansys Discovery AIM capabilities can be used to simulate and validate the entire design with high-accuracy simulations across a range of different physics domains and techniques. That I've outlined the big picture, let's double-click on Frustum and take a closer look at the magic it represents. I'll start by showing a few simple examples, then move on to a more sophisticated example that incorporates Ansys as well. Let's start by looking at a typical part designed today by a typical engineer. This is a bracket, holds a shaft, bolts onto a base. You can see we've come up with a part that we probably created with some kind of a machining process, and it's a pretty typical design. Now, if we turn to Frustum and say, "Frustum, tell us how you would design this part. Here's what I want you to do.
I'm going to specify these four blue circles as anchor points because I want to put bolts through those four holes, and then these areas can't move." These two larger blue circles are effectively like bearings, and I'm going to put a shaft through there, and I'm going to apply pressure on that shaft, forces. One last piece is I took the outside wireframe from the previous design, and I said, "Try to keep your design inside that envelope because I know I can replace the previous part if you can work within that space clamp." Frustum takes over, and it generates very quickly iterating here with AI, generates a very efficient design that would use a fraction, frankly, of the material but could cover and carry the necessary loads and anchor the shaft in place. That's a quick example.
What if I said to Frustum, "Do the same thing, but give me a part this time that I know I can make efficiently in a casting process." Think that I'm going to be pouring a molten metal into a casting, and I need to be able to remove it and so forth. Here's a design that's castable. Much more efficient than the original part, not quite as efficient as the last one you saw, but I can make it in a casting process that allows me to work with high volume. I might say, what about a forging process?
Now Frustum thinks through and says, "Okay, here's what the design would look like if you tell me you want to forge it." For example, I need a parting line, so when the dies come together and slam the blank and I pull the dies apart, the part falls out. It doesn't get built into one of the dies. That little line that you see in the geometry is the parting line, which is necessary for an efficient forging process. What about an additive process? By the way, feel free to use this concept called computational geometry, meaning compute a geometry that's far too complex for a human to model, but very efficient. If we look at this geometry, you can see air pockets through it and whatnot when we turn at different angles.
Extremely complicated geometry that you could never make by hand in a parametric CAD system. Now, I might say, that's a good-looking design, I suppose, but it seems to me it would get all gummed up with dirt and grease. Why don't you try again, but this time wrap a shell around it so that it looks solid from the outside, yet I have all that very sophisticated and very efficient computational geometry inside the part. I would end up with a part like this. Hopefully that gives you a sense for how Frustum would approach this design technique and the concept here of generative design being able to give you lots of different options, depending upon the range of inputs that you give it to consider.
For a more powerful example, let's turn the clock back to my LiveWorx keynote in June of 2018, where Ajei Gopal joined me on stage and we announced our partnership to incorporate the Ansys technology into Creo. You may remember that we were showing how Ansys real-time simulation inside Creo allowed us to quickly see and address a design flaw in a Polaris snowmobile part. You remember, we were looking at this part, we put on the deformation plot then later switched to a stress plot. We saw that there's a stress concentration in the upper inside part of this part. We're going to make a manual change to a manually designed part. We added a sort of X-shape stiffener into this extrusion.
As we align the front edge, align the back edge, we then commit the change, within just a second or two of committing the change, the Ansys live technology has resimulated the part and already showed us what the results were. This demonstrated how Creo users now have the ability to iterate the design while seeing in real time the implication of each change. I know it's a powerful concept because many LiveWorx attendees said that seeing this breakthrough was one of the major highlights of the show. The part in question was an aluminum extrusion made in a process somewhat similar to the way a pasta maker works. You force the material through a die, it generates the right cross-section shape, then cut slices from it. How do we know that the extrusion was the right approach to start with?
Had we already incorporated Frustum technology into Creo back in June, we could have asked the AI technology to suggest some options regarding the best design approach or starting point. Take a look at some of the concepts Frustum came up with when we did this subsequently. All of them meet the design requirements while optimizing around weight or manufacturing process or whatever we deem important. I trust you can see how this represents a tremendous productivity advantage. Whether you're an engineer or an investor, you don't want to be limited by your own imagination or experience, and that's why we all brainstorm and collaborate with others. For many problems, you'd get a better answer much faster if you were collaborating with a digital assistant like Frustum, who knows how to produce products that are optimized for manufacturability and can meet customer needs most efficiently.
When you combine the real-time generative design from Frustum and real-time simulation of Ansys, you enable a single engineer to do the work of many. Artificial intelligence teamed up with simulation has the potential to create the first major disruption in CAD in 30 years. With Creo incorporating breakthrough new capabilities from Frustum in the generate phase, coupled with breakthrough new capabilities from Ansys in the iterate phase, proven best-in-class capabilities from Ansys in the validate phase, Creo is being propelled into a clear leadership position in simulation. Note that in addition to Frustum and Ansys, Microsoft will play an increasing role here too, as all three phases now require intense computing power, and ultimately that will come from Azure.
In the near term, a big step that Ansys and Frustum both took was to move the execution of the real-time codes from CPUs to GPUs, which can execute real-time codes up to 400 times faster. As companies move to ever larger and more complex problems, there will be a point at which the economics favor moving the computing load from hardware on the desktop to GPU farms in the Azure cloud that provide elastic compute on demand. I've discussed these opportunities with executives at Microsoft, and you can imagine how excited they are to provide a high-performance computing service to PTC and our customers. We're excited about the progress we've made in our Creo strategy, and the customer reaction has been tremendous. These latest capabilities are important because first, they'll enable a major new upsell and cross-sell opportunity for Creo.
Second, they'll create very compelling differentiators when competing for new business in the SMB space. Third, they'll place Creo in a strong position for CAD consolidation opportunities in larger accounts. Fourth, they'll allow Creo to infiltrate competitor accounts by delivering complementary breakthrough capabilities while coexisting with incumbent CAD systems. The Creo business performance has been strong for years now, and the Ansys partnership, coupled with the Frustum acquisition, gives us a lot of fuel to keep it going for years to come. We're thrilled with the Frustum technology, which has been validated at global manufacturing companies like Honeywell, Tesla, Stanley Black & Decker, and Boeing. Frustum has a strong and talented team that we look forward to welcoming into the PTC family. You can expect to see us deliver the first tranche of Frustum AI technology in Creo 7, with more capabilities coming out in Creo 8 and beyond.
In parallel, we'll incorporate the TrueSOLID volumetric modeling kernel and continue advancing the research and development of generative and volumetric capabilities while we continue to work with strategic customers on those efforts. I believe the Frustum acquisition will move the needle for Creo and PTC in the coming years. In terms of near-term financial impact, given the acquisition integration strategy I've outlined, we will absorb Frustum into our R&D run rate and deliver the first product capabilities in FY 2020. We do not expect Frustum will impact the guidance for FY 2019 that PTC shared on October 24th, and we believe that it represents another important tailwind to our FY 2023 long-range plan. I'm very excited about the momentum we've regained with Creo, the opportunity that this acquisition presents for our Creo business, and I continue to be incredibly bullish about the future that lies ahead of us.
With that, I'll turn the call over to the operator and begin the Q&A.
Thank you. At this time to ask a question, first, please ensure your phone is unmuted, then press star one and record your name when prompted. If you need to remove your request, you can press star two. Again, press star one to enter the question queue, and do please limit yourself to one question. If you have any additional questions, you may re-enter the queue. Our first question is coming from Ken Talanian with Evercore ISI. Your line is open.
Hey, Ken.
Hi. Thanks for taking the question. Historically, there's been little share shift among CAD vendors. Do you expect a new CAD war to ignite as features such as generative design are integrated?
A point I tried to make, Ken, is that you need a reason to shift, and you need a way to overcome the switching costs and technical obstacles. You need a impressive new capability, but you need to be able to solve for how you can either coexist with an existing system or actually have the functionality to fully replace it. I think that's a very interesting position for PTC because we do have the functionality to replace anything, and we actually have a unique capability to also coexist with practically anything. I think PTC is in a strong position with cutting-edge new technologies on a robust platform.
The people I think will have a problem are people who are either creating a standalone, cool, new, shiny toy but lack all that capability, or who are developing a impressive new capability on a low-end CAD system that couldn't replace a mainstream CAD system even if you wanted to. I think we're in a unique position, and I do think that there's going to be potentially more share shifts happening here, thanks to technologies that are more disruptive than incremental entering the picture.
Great. Thanks very much.
Thank you.
Please stand by for the next question.
Go ahead if you have a question.
Yes, the next question comes from Jay Vleeschhouwer with Griffin Securities. Your line is open.
Hey, Jay.
Thanks. Hey, good morning. Pretty interesting technology news. A couple of things. You're expecting to integrate this with Creo 7.0, which is about 16 months away in your current annual release cycle. Are you anticipating replacing any of the topology optimization or lattice structures capability you've had in Creo with this newly acquired technology, or will this be simply complementary integration? Also, in the Creo roadmap session at LiveWorx back in June, there was an interesting reference in the slides at the time to Creo as a service. You guys haven't really said very much about that. Is that something implicit in your remarks about moving more and more of your capabilities onto Azure? If you could just talk about the business or technology implications of that Creo as a service reference.
Yeah. On the first question of will this replace our previously delivered topology optimization, yes and no. The no part is we'll retain that for compatibility. The yes part is that this is a much more advanced approach and will effectively, over time, obsolete the current capability in favor of something that is far more powerful. One way to think about generative design, it is topology optimization on steroids. It's about generating many different designs that contemplate many different variables and sources of input and so forth. It's topology optimization and so much more. I do think this will be our go-forward strategy, but for compatibility purposes, we'll retain the capabilities we've delivered. Then on the second question, which was Creo as a service, what I would say is already Creo is having capabilities built into it, like augmented reality, that require our cloud service.
The Ansys capability, in the near term, runs on GPUs, but soon enough, we're going to have to move that up to cloud-based high-performance computing, and now Frustum the same story. I think you're going to have Creo with a service. Now, whether the Creo executable itself should run in the cloud like Onshape does or not, I don't know. I think the jury's a bit out on that. Frankly, I think Autodesk tried that and then retrenched back to an executable that really runs on your desktop, even in what they call their cloud-based CAD. That's simply because you can't get enough performance for each and every mouse click when it's going up to the cloud and back today.
I say that in part because, as we prepared our notes for this call, we did it using Microsoft Word in the cloud, what's called Word Online, and it was not a pleasant experience. It sort of reiterated to me why it would be best if you ran the executable on the local machine and then farmed off certain necessary services to the cloud, where the cloud could dramatically over-perform the local machine. I think of it as Creo with a service or with a set of services is likely the near-term strategy.
Thank you.
Okay. Next caller.
The next question comes from Jeff Hojlo with IDC. Your line is open, Jeff.
Hello, Jeff.
Thank you. Hi, Jim. Good morning, and congratulations. Question is around the future of generative design and kind of where you're taking the acquisition. I think part of the promise of generative design is that you can incorporate a big data approach to design and development, where you can incorporate supply chain and cost information, different manufacturing process options, as you alluded to a little bit in your presentation. I'm wondering specific to PTC, how this ties into your IoT and manufacturing strategy, specifically around ThingWorx and your additive manufacturing aspirations, as well as your partnership with Rockwell.
Jeff, well, first of all, thank you for doing the work and the press release quote and so forth. I think it's a very interesting question because we already have ThingWorx, which is a huge data pipelining tool built into Creo. We can have Creo reach out into any system or any physical product and gather data and feed that into the generative design capabilities of Frustum. We're taking input from anywhere and making it available inside the CAD system to Frustum. For example, you could say, "Hey, Frustum, look at how this actual physical product is working and optimize me a better version of it, given consideration of how it actually is being used right now." That would be a digital twin feeding generative design. I think there's some real synergy here between ThingWorx and Frustum.
What I would say is, on the second part of your question, how does it fit into our overall strategy with Rockwell and so forth? I think at the higher level, our strategy is about digital transformation of industrial companies. We talk about how you can transform the product, the industrial company's product. You can transform the industrial company's processes, and frankly, you can transform the industrial company's people. Ansys and Frustum give us great ways to transform the look and feel and shape and design of the product, including the manufacturing process to go to additive or whatever. Our partnership with Rockwell is really about the process, the production process, the supply chain process, and that's, again, ThingWorx acting as a data and analytics tool to build applications that optimize production and supply chain and frankly, other processes. Moving on to people.
That's really our augmented reality and virtual reality technology, where we give people a superpower to see and use digital data as if it were part of the natural scenery. We overlay instructions or provide human-machine interfaces that are virtual or what have you. There's a whole bunch of examples. I think that this capability is part of our three-part strategy of product, process, people, and it's going to be a big needle mover in the product part of that strategy.
Okay. Thanks, Jim.
Thank you.
Next question comes from Steve Koenig with Wedbush Securities. Your line is open.
Hello, Steve.
Thanks. Good morning, Jim. Let's see. I wanted to ask about You had a customer quote, Jim, that suggested that Frustum's generative capabilities are already in the market. Maybe just clarify or give us color. Does Frustum have actual product in the market now? If so, will you continue to sell it? Kind of a related question, whether or not there's capability in the market now, what do you see as being the initial applications in terms of industries or particular aspects of the design workflow, where you might see initial traction with the Frustum technology?
Yeah. Okay. Those are great questions, Steve. Frustum does have a product in the market. It's kind of an early generation of the tool set capability. Our strategy, though, is that this functionality needs to be in Creo. It's not interesting to have it in a standalone product parked next to Creo. We will take that product and all the IP it contains and rebuild it into Creo. Now, we'll use that product in the meantime for marketing purposes and to engage customers around the concept and how to get feedback and so forth. We're not going to sell that product. In fact, we'll stop selling it, to be frank, while we build a solution that's being built for the ages as opposed to a temporary solution. I think, where will we use it? I believe we're going to see it show up all over the place.
I think you're going to see many different industrial applications. I think you're going to see it in medical devices for sure. I think anywhere where performance is paramount, I mean overall design performance, and maybe cosmetics not so much. Now, I'll come back to that point. One of the things you see is that you get designs that look like Mother Nature built them rather than the engineer you've been sitting next to for the last 20 years built them. I think it'll take people, in some cases, some time to adjust to that this actually is a better design, but it's a different-looking design. Now, one question that our CTO, Steve, and I have been talking about is could we teach Frustum, for example, to optimize around cosmetic look and feel?
There are actually AI technologies that are capable of doing that, and maybe we could make that a variable as well to consider in the analysis. Nonetheless, I think we'll see adoption across many different industries. Probably slow at first as people try to understand the power of this technology. Again, the fact that it, on one hand, supports additive in a brilliant way, but on the other hand, isn't restricted to additive, really encourages me that this could be adopted a lot faster. Because adopting additive is its own set of opportunities and constraints. For example, can produce amazing designs, but not in high production volumes. We can produce, as you saw in my little demo set here, amazing designs that are made in a traditional subtractive process as well. If you need a high volume part, that's a better approach in many situations.
You can go either way with it. All right. That's a long answer. I'll stop there and move on to the next question.
Next question comes from Monika Garg with KeyBanc. Your line is open.
Hi, Monika.
Hi, Jim. Thanks for the question. Just as a follow-up to the last question, maybe could you talk about the cases where generative design technology is used right now today? The idea seems to make a lot of sense, but why it hasn't taken off yet? Is it the compute power, very expensive before? Or as you said, this has to be integrated with the current solutions? Thank you.
I think there's really two limiting factors. One is the new technology, and people are just learning about it. The main problem today is that it sits next to a CAD system, and frankly offers poor interoperability with it, typically generating blobs of geometry that the CAD system really doesn't understand. I think that the strategy of a standalone startup company type solution for generative design, frankly, is not a workable strategy. I can tell you the Frustum guys knew that, and they were so excited because they said, "We've built this amazing technology, but frankly, we can never replace a mainstream CAD system with it.
We need to be inside a mainstream CAD system, and therefore, we're looking for a home where our technology can flourish and actually end up in productive use." That, to me, is the key thing, is that all of these cool things are standalone. Imagine you have a product with 700 parts, and you want to do two of them generative, to get started and try it out. The two have to live with the other 698 in the same assembly model, right? You have to be able to simulate it and so forth. I think that the real breakthrough here is to mainstream it into a mainstream CAD tool.
I just want to add something. Hi, Monika. This is Andy. In general in technology adoption happens when the perceived benefit's greater than the perceived pain. The perceived pain is often when you have to give up something you've been doing or using for a long time. If you had to give up the ability to design in a parametric tool, to be able to simulate in tools like Ansys, if you have to give up having all the data in one place and have to transport it from one system to another, all that perceived pain is going to prevent adoption. That's why we believe it's so critical that it actually be in Creo. That reduces the perceived pain. In fact, takes it down to almost nothing so people can go after the benefit.
Thanks.
Maybe to try to explain this data lossy broken tool chain problem, rather than explain it in CAD, I'm going to switch to a more familiar medium and talk about Excel. Let's say you made an Excel spreadsheet of a P&L statement, and you have revenue and cost and profit. I'm trying to make it simple. We won't get into 606 versus 605 and so forth. You have revenue, cost, and profit. You convert that to a PDF file and give it to somebody, and they say, "Oh, great, I have a PDF editor, I wish our company had more profit." They just change the profit line.
In your spreadsheet, you computed profit, but when you converted it to PDF, the computational underpinning was lost, and it just became a picture on the screen, and somebody edited the picture and created more profit without creating more revenue or less cost. You'd say, "Don't even give that back to me. I can't update my spreadsheet with that. First of all, I don't even know what you did, and second of all, it's nonsensical." That's the problem when you take CAD data. Similar thing, like I'll try to make a simple CAD example. Let's say you're talking about the wheel on your car. You got a flat tire, you need to take the wheel off and put the spare on. There's five holes.
The CAD system looks at that wheel and says, "No, actually, there's one hole that occurs five times in a radial pattern." Please don't take that wheel into some other CAD tool and move or change the diameter of just one of the holes, because that's not going to work. Actually it does work. It's like the PDF example. If you have the right tool, it'll do anything, but they'll give it back to you, and you'll say, "That's nonsense." You violated the fundamental concept of the design. That's why if you have it inside the same tool, you do things in a way that don't corrupt the data. If you do it in a separate tool, you're constantly corrupting data and then doing human negotiation around, what did you do? Why did you do it? You can't do it that way, et cetera, back and forth.
Pretty soon people say, "This tool's not creating any productivity. It's just creating angst and frustration." All right. Sorry for that.
Thank you.
Sorry for that long answer.
That's great.
I'm trying to make it a little bit simple.
Helpful explanation, thanks.
Thank you.
Next question from Gal Munda with Berenberg Capital Markets.
Hey, thanks for taking my questions. I've got a few. The first one is just, you guys have had this open platform approach for a while through Unite technology. Given how closely Frustum was to Siemens, one of your competitors, what does it mean for NX integration right now, considering the fact GD said kind of has to act within the CAD system? Would you ever open it to other customers, other CAD systems, like a Dassault SOLIDWORKS?
Yeah. Okay. There's a couple of different things I have to unpack there. PTC does have a very open approach. We have this technology called Unite that allows you to bring data from popular CAD systems into Creo and work around them, make reference to them, or even modify them and write them back out in the native format. It's designed to allow for a peaceful coexistence of multiple CAD systems. We came out with this technology a couple of years ago, and I'm trying to remember precisely when. It's been a good thing because it allows people to try out Creo, next to other systems. For example, we entertained a customer here at our corporate headquarters maybe two or three weeks ago. I don't remember precisely when. I was in a meeting with them. This is a customer who has Inventor and SOLIDWORKS.
Frankly, neither one of them, in the customer's eyes, are very good at sheet metal parts. They bring Creo in next to it and say, "Wow, this product can do sheet metal like crazy." First of all, parking it next to allowed them to test it, and now they see how they could actually change and transform over to a Creo system that can ingest all their historical data from these two systems. That's an example of we have two systems. We wish we had one, and the one we're going to pick is neither of the two starting points. It's Creo. We're going to migrate everything to Creo, and that will be our one consolidated system going forward. Switching to NX, again, there's several pieces to this. First of all, Creo can do what I just described with NX.
We can bring NX parts in, we could build things around them, write it back out as NX formats. There is a separate thing, though, that I want to point out, and that is that some amount of Frustum technology, part of it, is actually built into NX. Siemens licensed this technology, and you'll see that if you do any investigation. During the acquisition process, of course, we weren't really in a position to talk to Siemens about how to disposition that situation, but we'll do that now, and we'll try to determine should we continue licensing some technology to NX or not. That's a situation where their opinion, our opinion, have to be integrated together and see if we can come up with a mutual opinion, or not. We'll figure that out.
Definitely, this technology put into Creo makes it much more compatible with any other CAD system than it ever could be standalone.
That makes sense. Just as a follow-up, how do you guys, it might be early days, but how do you guys plan to monetize it? Is it like an additional feature within Creo or an upgrade? I know Autodesk and Fusion 360 are trying to use kind of credit-based approach in terms of the compute power used and the way the many times you generate it, you basically pay credits, that correspond to that. Is it similar thinking behind that, or would it be a separate product? I think at the moment it is a standalone product, right?
Yeah.
That basically costs almost as much as Creo.
Yeah. It will show up as a module or perhaps a series of modules within Creo that you have to license on a subscription basis. Now, we'll have to figure out the details, but one of the places that the subject of credits becomes interesting is around high-performance compute. Because, sooner or later, if the compute is being farmed out to the cloud, like let's say Azure cloud, then somebody has to pay Azure according to how much compute they actually use. It might end up being that there's a payment for the right to use the software, there's a separate payment for the high-performance computing, or maybe we'll find some way to integrate those two together. I think we need to figure that out, but certainly, this is a technology that we're going to upsell and cross-sell and so forth.
Okay. Thank you so much.
We are aware of how Autodesk is attempting to monetize this, and it's a pretty clever approach for the high-performance computing piece, but we'll cross that bridge when we get there.
Okay, perfect. Thanks.
All right. As a reminder, all parties press star one to ask a question. Next question comes from Gabriela Borges with Goldman Sachs. Your line is open.
Gabriela.
Hi, good morning. Thank you. Question on what you found to be most differentiated on the IP side for Frustum. It's a way of asking if you are successful with the strategy, what's the barrier to entry to stop other CAD providers trying to build something similar? As part of that, I'm hoping you can touch on how the actual algorithm has been trained. What is it that enables Frustum to come up with a really smart set of ideas and generative design that actually makes sense for the engineer? Thanks.
Yeah. Steve, being our CTO, this might be a good opportunity for you to jump in. You want to take that question?
Yeah, I can. Two pieces.
Great.
I would say the most exciting part is the path that the leadership at Frustum, their CTO and CEO, have taken in terms of how do you define complex geometries that we wouldn't want to define, I'll use the term parametrically, thinking in terms of established CAD. They've figured out ways, and Jim showed the complex lattice structure, to kind of describe a geometry as an algorithmic fill as opposed to a solid geometry, meaning it's a more mathematical equation. They put a lot of intelligent thought into how volumes are defined and how volumes blend into other volumes. That automated blending that you saw in terms of creating that casting part, they really respected the manufacturability aspect of things.
The level of automation that they actually put into blending all of that together was quite profound in terms of where they're at on the IP side of how their technology works. There are some other underlying capabilities in the technology, how they're doing multi-objective, multi-physics incorporation, as well as bringing in together, I'll use the term work steps. A plus B, and then do step C afterwards. They're not all done in parallel. They can be done in sequential logic. There was a question earlier about how do you verify for, let's say, supply chain material costs and so forth. They have devised a very nice platform inside their software architecture to allow that to exist. On the algorithm side, I'm going to be a little less specific on that, just to not be open this early in the conversation.
They do have a model by which technology, especially that which is generated on the cloud side, can understand the geometries that it's creating and start to train and infer kind of the behaviors and characteristics of the geometry that gets created. That is an area, quite frankly, where a lot more development needs to be done by all of the vendors in the space, I'll even say Autodesk included, where there's certainly a lot more creativity required on that end of it. Great first start from them so far in all of those categories.
Maybe I could add just a couple points. One is that the founder of Frustum actually was a generative design professor at Columbia. His name is Jesse Coors-Blankenship, by the way. Then the second thing I would say is there's really only two companies right now that have technology like this or anywhere like this. They're not identical, but it's PTC and Autodesk. I think the challenge Autodesk is going to have is it's in a very low-end CAD platform. Frankly, it's parked next to a very low-end CAD platform, and I think they're going to have trouble displacing or coexisting with people. I think that, in our view, Siemens and Dassault are far behind the state of the art where the Frustum technology is. We think we're in a very good place with this.
Ultimately, the winning solution is going to have to have the capabilities of mainstream CAD system and generative design at the same time, I think we're going to get there first.
That's very helpful, Colin. The follow-up, if I may, is actually on this idea of compute power that's come up a couple of times over the course of the call. Do you have a sense for, with your current installed base on CAD, how much of a constraint is the access to compute power, and how do you think that's going to change when you go to adding generative design, Discovery Live, all of these new and advanced things? Is there going to need to be more investment on HPC, either on-premise or in the cloud, or is there still a bit of room to run before we need a big investment cycle?
There is some room to run, certainly some room to get started, because every Creo user today has a high-performance graphics card that was previously only used for graphics. It contains a graphics processing unit, AKA GPU, which it turns out, NVIDIA's proved this, that certain analysis codes can scream when running on the GPU instead of the CPU. Every customer right now has a GPU sitting there being barely used. You take this by extension into bigger and bigger and more complex designs, sooner or later, you either have to get an even better GPU card, graphics card, or you say, "Maybe it's silly to be buying all this hardware when we could just get elastic compute from Microsoft." I think different companies will make that determination at different points.
I think everybody, nearly everybody, has enough to get started with and run for quite a while before they really realize, I need to scale up either with on-premise hardware or cloud-based scalable compute.
That's really interesting. Thank you.
Okay.
Is there time?
We got time for one more, if there's one more queued up.
All right. We actually have no further questions at this time. I'll turn the meeting back over to Jim for closing remarks.
Okay. All right. Well, thank you, operator. I'd like to thank everybody for joining us on the call and spending time this afternoon, or morning, in the case may be, on such short notice. We're tremendously excited, as you can probably tell, about the addition of Frustum to the PTC family. Their patented AI-driven generative technology, when coupled with Ansys, really set PTC apart and positioned this to be the first major disruption in the CAD industry in about 20 to 30 years. All right. Thank you very much, and have a good day.
Once again, with that, we'll conclude today's conference. Thank you for participating, and you may disconnect at this time.