Third Bridge专家称达索系统销售周期分化、云与AI需求改善但缺少离散催化

[[00:01:24]]
How has the selling environment changed in the past year, considering the macroeconomic uncertainty and inflation? Where does this leave buyers' headspace?

[[00:01:41]]
We've certainly seen headwinds with geopolitical, that's obviously played a lot of impact in certain industries like we've seen a boost in aerospace and defence, but with other, geopolitical things being what they are, we've also seen decreases in spending and investment in certain areas like automotive. That's been my experience. I think there's still demand, but it's up and down. What I'm seeing is things are more industry-centric. Customers are still investing. I still see customers interested in development. I still see customers interested in forward thinking. Obviously, AI is a big part of that. I'm sure we'll touch on that later. Geopolitically, there's certain main areas that have been impacted, and we're seeing customers just being more cautious. I'm seeing customers maybe delay some bigger spending and watching and waiting to see what happens in geopolitical and macroeconomic sense.

[[00:02:57]]
What are the impacts on Dassault Systèmes? Are you seeing things such as stretched sales cycles or a whole pause in spend? How does this translate to the company's sales motion and revenue?

[[00:03:14]]
Yes. I'm seeing customers slow down, not necessarily stop spending, but they're more careful in it. I've seen longer sales cycles is something I'm experiencing in these industries or these verticals that are really impacted. For example, in the automotive space, the decision-making process is a lot slower. The customer spend, I'm still seeing it, but albeit, I think, it's a slower, more careful, more articulate decision. It may be scaled back a little bit, but it's certainly not stopped altogether. I think customers are being a lot more careful with every dollar they spend and being very calculated in where they invest.

[[00:03:59]]
Could you quantify the sales cycles? What would you say is a typical sales cycle, and what is that stretched to today?

[[00:04:08]]
Yes, a good question. It can vary a lot. Part of the world, there's a lot of different products. There's some complexity to the answer. It's going to be a little nuanced, broadly speaking, in generalities, when I think about big-ticket spend, which would be like a, let's say, a PLM implementation, for example, a sales cycle for something of that sort from beginning of engagement to purchase order could be on the order of 6-12 months. What I'm seeing now on a product, on a spend like that, a larger, more complex investment and implementation is closer to 12-18. Again, the customers are looking at every penny, looking at every dollar. There's certainly more scrutiny on the product set and the back-and-forth, and there's just a lot of carefulness, which is sliding things out a little bit in terms of a sales cycle. On a more commodity-type product like a CAD product or somewhere in the middle, like an engineering simulation tool like a Simulia product or something like that, I would say typically, it's 3-6 months, and we're seeing it closer to 6-9 in a similar fashion.

There's ups and downs with industries. On the inverse, I'm seeing in other industries like aerospace and defence, I've not seen these sales cycles lengthened at all. In fact, I've seen some of them maybe even consolidate in some areas where projects are quick and fast. We have things over here like the Golden Dome that tend to drive a lot of energy and a lot of excited spending in that space. We're seeing customers far less scrutinising in these industries. I would say the sales cycles are the same or even condensed a little bit.

[[00:06:06]]
How do the factors that you mentioned net out? Do you think Dassault will be waiting longer to see revenue come to it? Do the different end markets net each other out, ending up with a consistent sales cycle or revenue expectations?

[[00:06:24]]
Yes. What I'm seeing is that they balance each other out. I don't think that's necessarily abnormal in my experience. I mean, you see industries go up and down and cyclical in a sense. By and large, I think they're balancing each other out. In fact, I'm seeing growth. I'm seeing healthy growth is what things look like from my seat, but I can certainly appreciate that may not be the case in every geo and every place elsewhere. I do think it's more of a balancing effect than a net negative.

[[00:07:03]]
Are customers prioritising anything in particular across the tech stack today? What top-line items are customers looking for from Dassault and other vendors?

[[00:07:21]]
Yes. Again, depending on the tech stack, it can change a lot. I am seeing strong growth in cloud-based solutions. I think it was a little bit slow to take off over the past couple of years. In the past year, 1-1.5 years plus, I have seen a big move to cloud. I think certainly in engineering simulation type applications. I see a lot of growth there. I see a lot of positivity, a lot of strong customer experiences where maybe there was some shakiness in the cloud experience two years ago. I hear a lot of positive stuff there. I see a lot of growth there, a lot of positivity in that area. Then I would say that stretches, too, for PLM implementations. When it's cloud-based, tends to be pretty nimble, pretty quick. Obviously, an on-premise implementation is just more complex by design. I'm seeing growth in cloud. I am seeing growth in 3DExperience. Its customers, I think, again, it's been a little bit of a journey if you follow the world in that sense. I do think the product set is beginning to improve. I think a lot of that goes with the cloud improvement. The hosting of it has gotten much, much better in recent years. Customers, I'm seeing an adoption, and I'm seeing more enthusiasm than I have in, say, 3-4 years ago.

[[00:08:55]]
What is your revenue growth outlook for Dassault in the industrial and manufacturing end markets – excluding healthcare – over the next 2-3 years?

[[00:09:11]]
Yes, good question. It certainly seems like double digits from where I'm sitting, at least in my space. Again, geopolitically, things can change pretty quickly. I think that's a big risk factor, but things holding stable where they are, I'm looking at low-double digits, is something that is what I'm forecasting and what I'm seeing in the customer space.

[[00:09:39]]
Consensus estimates are expecting 5% growth for Dassault in 2028. What do you think drives the difference? What are you experiencing that perhaps hasn't come through to the analyst side yet?

[[00:09:57]]
It could be that discrepancy, it could be just in the customer base that I'm dealing with. I'll clarify that I tend to interact a lot with defence and aerospace customers. That may skew some of my perspective a little bit from the broader sense just because my personal interactions are going to be heavily weighted there. I think even from a broader sense in our neck of the woods, we're still seeing decent economy growth. I'm still seeing customers looking at what they're doing and looking to innovate. I think with the advent of artificial intelligence and process compression and a lot of opportunities that brings us, I think as long as there's no paramount shifts in everything that's happening, I still see pretty steady growth. Again, it's going to be choppy in maybe some industries, but I think the net positive will remain pretty solid.

[[00:11:14]]
You mentioned factors such as cloud and AI. What do you think will be the main growth driver in the next 2-3 years?

[[00:11:25]]
I think it's definitely going to be AI. I think Dassault with the latest release, there's some technology that is out/coming out that they've begun to announce to customers here in just recent weeks. I think that technology is going to be pretty revolutionary in our space, and I think it's going to drive a considerable amount of business in cloud since it's rooted in cloud, it's rooted in a SaaS model. I think those two things are going to be symbiotic and continue to be a significant growth driver in the next leg.

[[00:12:26]]
Do you think AI solutions will be sought after? Would AI hasten movement towards 3DX [3DExperience], for example? How does the AI driver translate to Dassault's product portfolio?

[[00:12:43]]
Yes, good question. The way I see AI working is more a significant process compression. Whereas I'm using big round numbers here, if a company had 100 engineers, I think you could get the work of 100 engineers done with 20. I don't necessarily think that's going to result in mass lay-offs, but I do think the scale of what people are going to be able to do is going to be pretty tremendous. I think Dassault is in a very unique situation in a good way because Dassault has a data platform that's pretty unique in the sense. For example, if someone is designing some type of product and they have data in the platform, they have data in engineering simulation form, the artificial intelligence is able to take all this information and make incredible design suggestions, requirements and all these things and it's a scaler effect. I think because it's rooted in the cloud and it's rooted in the platform, I think it's going to be a significant growth driver in those two areas in particular, right? It's a scale at a pace that I think we've not yet seen.

Personally, I'm very excited about it. I think it's going to be pretty revolutionary in our space. I do think that Dassault relative to all the competitors has a unique situation here because of and certainly, it's going to be different, look a little different for customers that have mixed environments. For the customer that is in the Dassault environment and has a considerable amount of data in that Dassault environment, I think the scaling effect, the productivity, the process compression, all these things are going to be tremendous and drive a significant amount of usage for cloud-based token and of course, the platform because all this is rooted in the platform and that SaaS model.

[[00:14:55]]
What is the typical pricing model for the platform? When we say platform, I'm assuming this is 3DX.

[[00:15:05]]
Correct. Yes. I mean it's a little different because there's not a one-size-fits-all. It's a hard thing to answer because either someone could buy Abaqus and there's a legacy Abaqus version and then there's certainly which has just been around for a long time. Then there's also a version of Abaqus in an instance that exists in and on the platform. I guess it depends on the scope of what somebody is doing. If someone is interested in, okay, quote, unquote, 3DX, but in terms of kind of a classic PLM implementation, that pricing model will look very different than someone who's using products like an Abaqus on the platform. I don't know, it can vary quite a bit. I don't know, let me pause there and ask and get clarity on the question. Are you thinking more from a PLM perspective, from an individual licence perspective? Just to understand the scope of the question.

Yes, from a broader PLM [product life cycle management] perspective – essentially the place I wanted to end up at.

[[00:16:14]]
You mentioned the engineers going from 100 to 20 – I appreciate those were example numbers. If we do see a level of decrease in engineers, could that lead to seat-based contract value pressure on the platform?

[[00:16:46]]
Potentially, but I actually don't think it will be because I thought quite a bit about this. I don't think it will be revenue-impacting even if there is seat reduction because the licensing model is going to be a credit consumption. If someone is using so much AI such that the need for a certain number of engineers is not needed. Maybe those licences go away, I think the investment will still be there because they will need to be investing to experience that productivity gain in credit usage, cloud, SaaS, things of that sort. There may be a named-user licence that drops, I think the overall spend will be net positive considerably because it's just the money is going in a different area that makes sense. I think the rate of consumption will be pretty tremendous.

[[00:17:44]]
In the Q1 2026 earnings call, Dassault mentioned making changes to how it is looking at resellers, favouring a shift towards an ecosystem-style model. What impact do you see that having on the company's go-to-market strategy?

[[00:18:12]]
Yes, good question. I guess some of that's yet to be seen. I haven't seen many changes yet or much commentary from other than what obviously was mentioned in the Q1 call. I do think when they're saying that, and this is my opinion, I think there's some interest in consolidating some of the different channels per se, like the SolidWorks and then the CRE and CP channel, SolidWorks and then, say the CATIA and everything, strategic channel. I think there's some interest because there's some product overlap. I think there's some interest in blending these things, recognising that a lot of partners operate in both and have business units set up to accommodate both. That can make things a little inefficient at times. I think where products overlap, where a customer may use SolidWorks on the front end and connect it into the platform, I think our customers are driving a more unified environment and solution-based environment than where, say, 10 years ago, it was bifurcated pretty notably. I think there's a grey area in the middle that's being customer-driven. I think the effort to consolidate some of that and simplify the ecosystem from a more customer-centric perspective is less than a product perspective.

[[00:19:47]]
Do you think what we mentioned will require quite large changes at the Dassault partners?

[[00:19:56]]
I don't think so because while there is some structural implications internally in the way partners manage the business because a partner will have, as an example, they'll have a CRE SolidWorks quota and then a CPE CATIA ENOVIA quota. They may have two different quotas depending on it. Just structurally, there may be a little bit of reorganisation, but I don't think it's going to fundamentally change the business because customer-facing, I'm already doing that. It's more of a back-end, back-office-type operations change. When I'm in front of a customer, when I'm talking about our solutions with a customer, I'm not thinking about Dassault channels. I'm thinking about putting the best solution in the hands of the customer to accomplish what they want to do. I think that's what's driving this. I think Dassault is recognising that and recognising that let's adjust some of the ecosystem, some of the operations, just the way we interact with our partners to accommodate what they're already doing in front of the customer. I don't think from a business perspective, it's going to impact much, but I think if anything, it will streamline and make life a little easier in more back-office operations, etc.

[[00:21:20]]
You mentioned that you haven't seen much change from the Dassault side. When do you think we might see more deliberate guidelines or clear changes from the company being communicated to partners?

[[00:21:35]]
Those things tend to happen a calendar year, is my experience. I've heard there's some whispers and some general commentary. I expect more guidelines to come out on this as we… Let's say H2, we're in H2, it's quite honest. Second half of the year, I would expect some of this as we move into Q3, Q4. Whatever structural changes or formal changes happen will happen January 1, January 2. I think that's when we'll begin to see it. That tends to be the timeline by which Dassault operates. I've never seen them make organisational structural changes to how they operate and what they do until they get to January, February when they start what they call the operation shift. They tend to be pretty firm in that, but there's not big changes until the turn of the calendar year. To sum it up, a long answer short, I'm expecting some communication guidelines to be communicated, more messaging in the second half of this year and then the structural implementation of it to likely happen in 2027. That's my personal perspective here. I don't have any other data otherwise, but that's my experience that tends to be how they function.

[[00:22:53]]
What is the CX today when customers are looking for a Dassault solution or partner?

[[00:23:09]]
I think it's evolving like the rest of the world. I think our customers are spending a lot of time online researching solutions. I think a lot of it is… There's website, there's SEO driven. I think if a customer wants to engage, I think it's very proactive on the customer's part. We do see a lot of that. We see a lot of customer-driven interaction, if that's where your question is headed. I think I feel we're seeing quite a bit of growth in that area of inquiries from customers trying to get more data, more understanding, more insight on various solutions via website and stuff. I think as with the rest of the world, I think it continues to digitalise and you see SEO and things of the sort play a pretty prominent role. I'm not sure if that's the direction your question was headed. That's a lot of what I'm seeing of that. On our side, we certainly try to be as proactive as we can in outbound engagement, installed base growth and whatnot. For new customers and such, I do see a lot of it where the customer comes to us and the cake is half-baked, so to speak, because they've been doing a lot of reading and studying and consuming of content that's available online to understand the platform and the benefits and so on and so forth.

[[00:24:45]]
If customers are being more proactive and educated, could they end up going direct? Could we end up seeing such a behavioural shift?

[[00:25:00]]
I don't think so.They recognise to scale, it's going to have to come through partners. There is an emphasis on the partner network. There's an emphasis on scaling. From everything I see, has put a lot of… Even when I was at Dassault, and I continue to see some of this now, they continue to invest and add headcount in the area of partner support and the building up the education, training, the execution, the partnering with partners. I'm seeing even more emphasis than when I was there now on that effort. I don't see that happening. I think Dassault recognises that, that's where the big… The next leg of growth from a product sense may come from AI. It's going to come from the platform. It's going to come from cloud. The field work is going to have to come through partners. I see a lot of emphasis on that.

[[00:26:05]]
You mentioned expectations around consolidation. Would customers be typically utilising multiple partners?

[[00:26:20]]
In some sense, again, you see the partner network in general is… There tends to be a general mode of consolidation. You see a lot of partners acquiring other partners and such. I think that problem is less of a problem today than it was many years ago. Many years ago, it was complex because part of the nature of those business is a broad portfolio, and it was difficult for partners to be complete experts and say, Delmia as well as Simulia, the technical investment of that could be a lot. What's happened in recent years is because they're different people, if that makes sense. The manufacturing expert is a completely different person than the mechanical engineer with a ton of expert and finite element analysis, they're different people. It was a very expensive per headcount and technical expense to be experts in all these broad areas. What's happened in the past five years-plus is a lot of partner consolidation. The Simulia expert partner acquires the Delmia partner. You see these merging of businesses. I'm seeing a lot of consolidation on the customer side, whereas certainly, customers can work with whoever they want.

The Dassault is pretty adamant about that. They want customers to be able to choose the best partner regardless of the solution. I do see a lot of consolidation where customers are working with less partners and really product consolidation in that sense or vendor consolidation, excuse me, related to the products. A lot of that's been organic because of this acquisition mode. It's certainly less prevalent than it was. I think that's actually… It's not a key driver of growth, but I think it's an element of growth. It's easier… Personally, it's easier to grow a customer when you have the entire product set, and you can pitch the entire vision. You can help the customer see the entire vision from a single voice and it's not piecemeal. That continues to improve, and I see that trend continuing, actually. I think that's going to be better for the partner network and better for the customer, too. It's just the world we're in that consolidation effort among VARs, but I think it's good for the customer.

[[00:28:53]]
What do you see a change in the partnership model enabling for Dassault?

[[00:29:05]]
I think nimbleness to the market, speed to solution, I want to hit that earlier. I think it's building a network that is shaped around the way the customer is already doing business. Which I think makes us more nimble, and I think it makes our solutions easier to implement and better for the customer altogether. It's arranging the network to accommodate the way the customer is already using our solutions.

[[00:29:38]]
How do you see that translating to revenue?

[[00:29:45]]
For reasons I mentioned, I think it will be very good. Like I said, it's easier. If we have partner consolidation and we have channel consolidation, so to speak, where these two things are coming together, it gets a lot easier to sell, it's a lot easier to manage a customer to own the account to implement the entire 3DX vision because it's coming from a single source. It's a clear messaging. You don't have too many cooks in the kitchen, so to speak. It's better. I think it will be better for the customer, and I think it will be really good for revenue because products are consolidating, customers are consolidating, vendors are consolidating and the ecosystem is consolidating. I think the net revenue will actually be quite good. I think it will make speed to sale and everything… It will make the network nimbler and better overall.

[[00:30:50]]
Does this lead to quicker or accelerated growth?

[[00:30:57]]
Yes, I think so. Geopolitical situations being what they are, I do think it has a strong potential to shorten sales cycles, accelerate growth, accelerate… You compound that with what AI has the potential to do. I think both of those things are big indicators that growth acceleration is possible, if not likely.

[[00:31:24]]
I want to understand more where the growth is coming from. Where do you think the impact is felt – existing customers or quickly acquiring new customers?

[[00:31:51]]
It's a good question. We'll have to wait and see. I think it will catch on quickest with existing customers because if we're thinking the accelerator is AI cloud SaaS model, the more data a customer has in connected to the 3DX platform, I think the better these solutions are (inaudible 32.14). I think the customer installed base that already is using our solutions, we'll see an immediate kick. I think we'll obviously be able to attract new customers, and we'll see the customers move to the platform. I think it will be a driver to get customers on the platform. They'll see the vision and recognise that this is the direction we want to go. I need to get on the 3DX platform because of this accelerator. I think Wave 1 will likely be a nice increase in spend and revenue from existing customer base. I think it will be a wonderful attraction piece to new customers as they see the solution in the marketplace and what it's doing.

[[00:33:02]]
Dassault set out three priorities – transforming customers, expanding to new customers and then Universes, its industrial AI solution. How would you break down the expected growth across those three areas?

[[00:33:35]]
Are you thinking in terms of just broad percentages? How would you like the data communicated?

[[00:33:42]]
Yes, just broad percentages.

[[00:33:48]]
It's a good question. It's projection and guessing. I think I would not be surprised to see, again, low-double digits to 10% to 20%, somewhere in that range on existing customer base. Out of the gate with again, I think that accelerating in time. As the AI model picks up, as these 3D Universes of industrial AI continues to go, I think the consumption rate of AI will increase and be a very healthy revenue driver. I think we'll see something probably… Again, this is just me thinking. It may be high single digits for new logos initially. As that catches on, as the success is demonstrated as the cloud continues to get robust, as data is gathered, I think that the single digits on new logos could trickle into double digits in a year or more. I don't know. We'll have to wait and see. I think the installed base, the existing customer, that transforming of the initial customer will be quick. I think it will be a pretty immediate revenue impact as it's experienced. The new logo customer, we continue to see customers adopt the platform. That's still happening. Obviously, I think it will be a lag. I don't know, 10% to 20% on new customer, maybe single digits for new logos, but transitioning into double digits, lower double digits over time.

[[00:35:34]]
When you mentioned your growth outlook, you said double digits. Do you mind putting a rough range on that?

[[00:35:50]]
Yes, 10% to 20% growth, maybe leaning towards the 20% growth. I think that's very possible.

[[00:36:00]]
Potentially 3-4x where consensus sees. What was the main driver of that?

[[00:36:13]]
I think the application and the usage of the way that this industrial AI is going to impact the business. Because customers will have to buy credits to use this stuff and consume them. I think that's going to be a tremendous growth driver for cloud, SaaS and the platform. Again, because existing customers already have a database set for the AI to leverage, I think that's going to be a huge growth driver for them. I think customers will we'll experience that quickly and be excited to invest.

[[00:36:53]]
In terms of transforming customers, cloud grew 8% in Q1 2026. What do you see driving customer cloud migrations in 2026-27?

[[00:37:10]]
Well, obviously, AI can't get away from that. Next year, this year, I think that is going to be a driver and certainly more next year than this year as it continues to become more known. Currently, the growth driver as it is, I think just the cloud has gotten better. I think ease of cloud, the rising cost of hardware, the complexity of managing an on-premise 3DX implementation. I think all these things are drivers for customers to go to the cloud. I think it's just an easier implementation. It's an easier thing to manage. Every year, it gets better, it gets more reliable and the customer experience continues to improve. Again, I think all the growth drivers are pointing towards cloud and more of that SaaS model. I think AI is only going to accelerate that.

[[00:38:11]]
What runway is there? I believe cloud migration has been a theme for several years now. How would you split the industrial innovation and mainstream innovation customer bases between on-prem and cloud?

[[00:38:37]]
I would say, as is, I'm going to guess, 60/40 on-prem to cloud, but transitioning to the inverse of that over the next one or two. I think we'll see 40/60 on-prem to cloud to transition to 60/40 cloud to on-prem is what I expect to happen. There's always going to be a certain set of customer base, and it's a big set of the customer base for Dassault that needs to be on-prem because of ITAR and certain cybersecurity requirements related to Department of Defense, government, things of that sort that will just never get on cloud or I don't know, I'd say never, but it will be a long time, I think, before that happens. Those tend to be big Dassault customers. Just because they're heavy use of CATIA, heavy use of the platform, etc. I think there will always be a heavy set of the customer base that sticks on-prem despite what the cloud can offer despite AI, so on and so forth. I think as a partner, as someone who services some of these customers, I think we need to figure out how we can navigate AI existing as a SaaS model cloud with customers who are going to have trouble accessing and benefiting it.

That's a problem we have to solve, and I think it's solvable, but that's a challenge we're going to have to figure out in the next year or so. Aside from that, the push seems pretty strong. I think the cloud is much more stable than it has been five years ago. The hardware for the cloud continues to get better. The disruptions, the downtimes are almost non-existent. The models run fast, the data pulls fast. Again, you add industrial AI, and I think that's going to be a significant driver to move the customers that can away from on-prem to more cloud instances of the platform and product usage.

[[00:40:47]]
Where does this leave the pace of migration? Is it going to remain stable, accelerate or decelerate?

[[00:40:54]]
I think it will accelerate. I really do. Again, I think we're already seeing an uptick, at least the way it appears to me. We're seeing more customers willing to entertain cloud than we did, say, five years ago. I do think as the cloud hardware gets better, as AI continues to offer incentives for customers to move because you need to be on cloud to use these AI tools. I think the temptation of that, the lure of that, the benefit of that will accelerate this, and we'll see a faster transition from on-prem to cloud for those that can.

[[00:41:32]]
For the customers who can move, what do you think is causing the hold-out? Why would they have not done so already?

[[00:41:45]]
Good question. I think initially, say, 3-5 years ago, there was some hesitancy on the cloud performance. I think that has largely been taken care of. I think that was an issue. I think customers… We talked about macroeconomic things. I think there are certain customers who are just slow to do it because they're… It's easier to maintain status quo to invest. They're just more careful. They're exploring every avenue. I think the value prop hasn't been as good as it is now and it is continuing to be essentially. The speed to cloud, the ease of cloud and then with AI, just the value proposition of ease of use and the managing of complexity, it's just getting better and better. I don't think that's going to change. The hardware continues to get better. The way the cloud is going to continue to get better. Maybe you could say it this way, almost tongue in cheek, but as bad or as good as it is today, is this worse that this bad as it will ever be. I think we have every reason to think it's just going to continue. The performance will continue to get better, and that will continue to move customers. I do think performance issues many years ago were a big hesitancy for customers.

I think they saw downtime and there was some… We need to wait until this improves, but it's there. The cloud is very reliable. I use it every single day, quite literally, every single day. I have no performance issues, and it really is. It's a great business model. I think the timing is good as hardware and everything continues to improve. Sorry, I feel like I'm repeating myself. I think the performance was a big issue, and it's really there at this point.

[[00:43:51]]
You mentioned that we will likely reach a ceiling because of security concerns. Where do you think the cloud-to-on-prem split might end up stalling out? Is it 80/20?

[[00:44:12]]
Yes. 80/20 is probably a good nice round number way to think about it at some point. I do think there will be a solution for these customers who today cannot go to cloud. I do believe that there will be a way to make that happen. I don't know what that looks like today, and I think we just have to wait and see. There's also a comfort level. It may just be that these customers, even if the technology is there, they just may not want to. I think we just have to acknowledge the fact that there is going to be a segment of the customer base. Again, it's a notable customer base because they tend to be large customers, but it's by far the minority. It's these Department of Defense, these government-centric entities that do not want their information anywhere on a cloud. However they're using the Dassault tools, sometimes they exist in instances that are not on public networks, things of that sort. There's always just going to be no matter what… How good the performance is, there's going to be a segment of the customer base, albeit a small one, notable one, but small one that may never get on cloud.

I think that will erode over time. I think you can continue to see that trend, maybe even get to 90/10, but 80/20 is probably a decent ratio of what we can expect for upside in the near term over the next several years. Until some of that cybersecurity stuff is accomplished and the customer sentiment is willing to say, we're okay with that because it may technically be savvy, but you may just have a customer set who's like, even if it's technically there or certified, whatever, we're just not comfortable with that. That's the nature of the world we're in.

[[00:46:20]]
Could you quantify the change in customer value upon cloud migration, if any? Does the annualised spend change as a customer goes from on-prem to cloud?

[[00:46:34]]
I think it varies a little bit. There's never one size fits all when you're dealing with PLM. Every customer is unique in a sense. I do think the SaaS model proves better for customer value. Year one spend from a customer going on-prem to cloud, it may be high single digits, 5-10%, something like that. The long-term value is really where it is. Year one, maybe 5-10%, but year two, we do tend to see an increased spend in software licences and instances. Again, I think AI will accelerate that. 5-10%, maybe year one and then 10-20% years two, three, four, five, etc.

[[00:47:23]]
Is that from additional software licences and modules or additional people using the platform?

[[00:47:30]]
Both actually. There's some interesting artifacts I've seen. I've noticed some instances where I tend to see… There's an inflection point. It varies size, let's say, if it's a… I won't even give… When you get to a small, medium business, when they tend to get to the 20-30 users, when you have 20-30 users of the platform, what you take, this is cloud. You tend to see an inflection point in there where it's steady growth to 10-20 users. When you get to that 20-30 somewhere in there, there tends to be a hockey-stick effect where the adoption rate increases pretty rapidly across a customer base. Again, it's more of a steady slope, 5-10% year one, maybe in year two. As customers add users, as they bring in more products, as the platform begins to scale to an organisation, when you get to that 20-30 range of users, I tend to see a pretty big adoption rate increase, the rate of increase goes up pretty dramatically. Again, that associates with the spend. The nice thing about cloud is it's easier to add these users in when it's on cloud because it really is just a licence. There's not an implementation aspect that you have to worry about. The data is on the cloud. It's easy, it's quick, and that accommodates that rapid scale.

[[00:49:03]]
When do you think we might see industrial AI uptake and token usage driving revenue generation?

[[00:49:24]]
Personally, I'm hoping to see it the second half of this year as we see an initial adoption or initial wave of adoption in Q3, Q4. Likelihood, just markets being what they are, the big rate of increase will come as we finish out the year and maybe transition to 2027. These things are just now coming to market. We'll have to wait and see. I think we'll see a nice wave. I'm excited to see what happens in the second half of this year, but 2027 is, I think, just having a full year, 3-6 months for customers to understand it, begin to adopt it. As word spreads and the demonstrated value is seen, I think 2027 makes a lot of sense to see a huge influx of revenue.

[[00:50:12]]
How does Dassault's approach to AI differ from those of competitors?

[[00:50:20]]
Great question. I want to make sure I'm careful here. I'm not saying anything that's not confidential. I think the fact that it's embedded in the platform, the fact that it's a model that was built. It's not just a bolt-on cloud or bolt-on ChatGPT, it's not anything like this. It's something that Dassault has developed, and it's specifically targeted to work for engineering purposes. There's three agents that they've got, but thinking of one called Leo. Leo in particular, I think, has potential to be tremendous because it's engineering centric and it's designed to guide users to design products based on known data and known requirements. Because it's using customer data, because it's not pulling from outside sources, it's tailored, it's careful. It's a speciality tool. I think the robustness will be pretty tremendous. I'll just say that the demos that I've seen, the content that I've seen is game-changing in a lot of ways. I think a lot of our competitors, the stuff that I've seen, I'm certainly not an expert on our competitors, but the stuff that I've seen feels more like marketing and rebranding of optimisation tools than actual AI-driven design and engineering. If that makes sense.

It's taking a tool that was really geared towards design of experiments and optimisation, rebranding it as artificial intelligence and bolting it into a solution. Again, speaking in generalities, that tends to be what I've seen. Dassault handled the whole thing very differently. Here in this AI boom, there wasn't just some frustration on my part at times. I kept waiting to hear more on what they were doing on the AI piece a year or two ago. They were pretty quiet, but it's because they were developing their own thing in-house. I do think that, that's because it's an internal code, it's specifically developed. It's integrated into the platform. I think it's uniquely different in that sense. I think the performance and the scale of it is going to be pretty tremendous in the way that it operates in the Dassault ecosystem. It will be unlike anything else our competitors to do because of all the data that exists there.

[[00:53:12]]
Dassault's management mentioned an AI-at-the-core strategy and that it has rebuilt its architecture from the ground up, which is now complete. What do you think this looks like for users on the ground?

[[00:53:33]]
You mean the user logging in and using AI?

[[00:53:37]]
Exactly. How does that AI-at-the-core with the rebuilt architecture change their customer journey or usage of the AI solution vs a different vendor who doesn't have that same approach?

[[00:53:54]]
I think that's the inherent advantage, because the AI is able to learn and benefit from everything that user is done, from CAD to PLM to simulation data to manufacturing data, etc. All of this is pulled together to help guide the user to make the right design, decisions from systems engineering, the whole thing. That's what's incredible about it. It's an ability to pull insight and data from all these different aspects and inform the user immediately of excellent design decisions that they may arrive at, but maybe it takes them weeks to get to. The AI is able to guide the user and coach the user to provide options to the user of design and engineering decisions that may have taken, again, weeks or months for them to arrive at based on data that they would have to go collate in the search and run through. That to me is what I think is going to be game changing because whereas it could be many departments funnelling in and checking and let's check our design decision with the systems engineering team. Let's check our design decision with the electrical engineering team.

Let's check our design decision with the manufacturing team, etc. All of this now has the potential to be calculated and considered at once when design decisions are and engineering decisions are being made. That's what I think it means that the whole architecture is built around that. The user has the ability to be informed from all things at all angles at all times. I think that is extremely unique and what will be tremendous for the user and for the business.

[[00:55:40]]
Are there any barriers to uptake here? For instance, when I think of those virtual companions, would they only be suitable for customers on the full 3DX platform, or anything else that could prevent uptake or delay it to a further year?

[[00:55:55]]
My understanding is they're only on cloud. That would be the only restriction I'm aware of. If you're on cloud, you're always going to be on a current version. To my knowledge, we'll have to wait and see. To my knowledge, I think that's the only restriction is someone just needs to be on cloud to access this technology.

[[00:56:18]]
Dassault's management mentioned that AI is easy to price without having to destabilise the current pricing system. Can you walk me through this?

[[00:56:35]]
I think it goes back to what I had at earlier. I think it's consumption. I think it's an accelerator for growth and that it will compound on itself. As these tools get used, as customers are reaching out to AI for guidance engineering decisions. The customer is not doing anything different than they would normally do. They're still using and referencing Simulia. They're still using and referencing Delmia, I know all these different things. What's happening is the customer is able to get to these decisions and make better decisions faster. I don't think in time, again, in time, there may… I can imagine a scenario where maybe there's some roles reduced. I think the benefit of what AI is going to bring from consumption rates and that is going to far outweigh what some users losing. Because the productivity gain, the scale of what's being done is going to increase. I don't think it's going to erode the business. I think it's going to accelerate it. There's a very real scenario where it even adds rolls because the scale and the productivity gain is so great that companies are able to do that much more that they see it's worth adding more people in to scale R&D and get work done, R&D projects done faster.

As it is with AI, there's so much just in the world in general, none of us know what the world is going to look like as AI changes everywhere. We'll just have to wait and see. I think it will be good for the customer. It will be good for the partner. I think it will be good for DS. Personally, I'm very excited about it if you can't tell.

[[00:58:30]]
To make sure I'm following, given management mentioned not destabilising the current pricing system, is 3DX already consumption-based?

[[00:58:45]]
It's more user-based. It would be consumption-based in a sense of solving jobs for an engineering simulation. That would be the exception. Generally speaking, it's Triram role-based. I think what they're saying is they don't expect to lose all these users. These roles are going to be there. The pricing model is going to be the same, and they're just going to be able to add on top of that consumption rate. I do think that's generally true. Over time, we'll have to see what happens because, again, AI is changing everything constantly, but I think that's what they're saying.

[[00:59:21]]
What ACV [annual contract value] impact would you expect AI to drive in the next 2-3 years?

[[00:59:33]]
I'm really hoping for that 15-20% overall. I really am, and I think that's very possible.

[[00:59:43]]
Fantastic. Thank you so much for giving me some of your time today. We have to wrap up there. It's been an absolute pleasure, and I hope you enjoyed it as well.

[[00:59:54]]
I hope this was useful for you. It was a pleasure. I always enjoy talking about these things. I hope it was useful and beneficial to yourself and anyone else who listens.

[[01:00:06]]
Super useful. It would be great to stay in touch. Otherwise, enjoy the rest of your day.

[[01:00:11]]
Thank you. Have a great weekend.

Transcription ends at 01:00:14 of the recorded material