Autodesk, Inc. (ADSK) Earnings Call Transcript & Summary

The exciting announcement?

Yes. So the point is I have to read you a safe harbor, so apologies for that. So we may make forward-looking statements during the course of this presentation. Please refer to our SEC filings for information on risks and other factors that may cause our actual results to differ materially from these statements. So that's one bit. The other thing I want to say, given the date, is just to express my condolences to the family of the 9/11 victims many years ago and also to thank the first responders who responded and the people who cleared and reconstructed the site. So just to say that.

Thank you so much. I appreciate it. Welcome to the Goldman Sachs Communacopia and Technology Conference 2024. Welcome to you, guys, as well.

Day 3. Are you holding up okay? Is everything good? Excellent. Yes. So welcome. And we've not heard much from Raji. Can you tell us a little bit about yourself, your background, how you got to Autodesk, and what's your role at Autodesk, if you don't mind?

Kash, it's great to be here. My first time, so thank you for having me here. With my prior companies and at Autodesk, it is really driving some similar outcomes. So it is all about how do you really build shared capabilities that drive productivity and efficiency within the company for our employees. And the second one is really about customer innovations and sort of future business growth through modernizing with cloud technologies, and that usually includes cloud and AI and services underneath that. So it feels similar across -- it's a platform transformation. And so when I got to Autodesk, it was like, wow, how can there be common shared capabilities across 3 big industries? And believe it or not, there are. And so that's our focus. A lot of that focus is also about connected data, the data plays a huge piece here, and then building out our marketplace and ecosystem. So those are my focus areas at Autodesk.

Talk a little bit more about the platform, APS. I'm curious to hear, what is the data -- the work that you've done to harmonize these 3 intractable applications that are very different end markets?

Absolutely. You said that very well.

Let me see if I still remember my Autodesk 101, manufacturing, architecture, engineering, construction, media and entertainment.

Media and entertainment, you got it. And underneath that actually, sometimes we don't talk about it, but transportation, rail and some of those are actually infrastructure underneath that. So it's pretty big and wide. When you talk about APS, it's essentially our Autodesk Platform Services. And when you look at APS, it was a key part of what it delivers. Actually, I do want to build a little bit of the context on why this is important. From a customer perspective, there are -- what they need is connected data. They want to connect their data. They want to get more intelligence out of that data. Today, there's a lot of time that's lost in moving files and creating error in this process. It's very expensive. And the second thing they need is the teams are spread everywhere. They're remote teams. They're trying to collaborate on global projects, so bringing them together is hard when you don't have cloud connected data. And third thing is Autodesk is one of many that they use in their workflow. So there's Autodesk products, and then you have -- I might be using a Tekla, I might be using something else, I might be using an ERP, and I want to connect that. So that's the fundamental thing that we are trying to solve through the platform for our customers. Now let's take this internal. When you start to talk about Autodesk, the shared capabilities that we're bringing is a key part of being able to accelerate some of the new plans that we have around our industry clouds in each of our industries. We are really trying to connect this into a workflow and getting away from product -- end-to-end products into more a workflow. And then the second part of what I just touched upon, which is the connected data, also helps us internally accelerate our AI work, which is very key. And it lays the foundation for that -- us to go faster there. So these are the 3 -- 2 context. And hence, the platform is all about that shared capability, connecting data, and then opening it up through this ecosystem that we have. I think that's how it all connects together in our strategy voyage.

Which we're going to get into in a few minutes. But does this also have a role in increasing the efficiency of research and development at Autodesk?

Absolutely. I think the key thing here is it's all the work that we're doing in terms of the shared capabilities, and we can touch on this later. It's like all of the gen AI productivity tools that we have made available to our -- all functions. And I know it was also covered in this conference before, but all functions in the company, and between those, there's a ton that we can do in being able to reshape our R&D efforts towards the future and where compute is going to be a problem or where AI skills are going to be necessary. I think it really paves that way for us to do all of that sort of future investments.

Got it. You talked about the marketplace aspect of APS. Tell us more about it. How are you enabling, with this platform, new companies that can build products in your marketplace, sell to your customers?

Yes. There is always confusion when people use the word marketplace. I do want to take a minute to say our marketplace at Autodesk is about 3 components. So you have cloud APIs that people use. There's an app marketplace, where you can build an application and actually sell it to our customer base. And then the third thing is we have 30,000 partners and third-party developers and all of that stuff, who are part of this what we call ecosystem or community. So these 3 together make what we call as a marketplace offering. And it's actually a key thing because right now, people use it in 2 different ways. They use it to extend or customize their workflows, so customers use it that way. And then partners build new products, so we have ArcGIS Pro from our Esri or Vcad. These are companies that are building new products on top of these APIs. But the future is going to be something different, which I think is people are going to want more of these APIs in their real-time workloads because they are moving towards automation. They're moving towards connected workflows. And for that, some of our APIs just become part and parcel of their workloads. So it's going to create more stickiness, more user retention for us. I think that the key thing here is customers get to grow in scale, partners get to grow their business, which is really important for us as well. That's an ecosystem. And then the third thing for Autodesk is because of that stickiness and sort of growth and retention, it will be a huge enabler for our future from a growth perspective. Anything you want to add, Simon?

Yes, just to sort of add to that. So conceptually, think about sort of how people integrate into Salesforce. So conceptually, it's how to think about it. And for that marketplace, there's sort of 2 key power cords that power that marketplace. One is data and one is transactions. And so the data side, Raji's already talked about it, having granular data, so data that isn't locked up in a file. And then you have granular data unlocking that data so that it can flow around the ecosystem, not just internally within Autodesk but within the ecosystem of our customer ecosystem and our channel partner ecosystem. And then the second one is around the transaction. So the new transaction model is a key component of that, which enables the financial transactions to flow around the ecosystem as well. So when we talk to you around Autodesk Platform Services, when we talk to you about the new transaction models, those are the 2 key power cords of the marketplace ecosystem, and that's where the company is heading.

So Raji, if I reflect back on AU 2022, I think Platform Services was a big component of that discussion as well as the industry clouds, right? So can you talk to us a little bit about how the industry clouds fit into Autodesk's longer-term vision?

I think you probably heard all of it within this Box core platform. So when you talk about Autodesk Platform, it's the APS layer, which is the Platform Services. And then you heard about industry clouds, which are almost like a single offering within each of our industries. So it's very similar to the concept of marketing clouds and other stuff that you've seen from other companies, right? So we believe that we want to be able to get to a point where because we are filling up our portfolio, we want to get to a part where people just naturally go from one place to the other. And not -- today, what they do is they have different products. You have to open up different products for design and make and operate and all of that sort of their life cycle. It will be an integrated and interoperable sort of end-to-end offering that we can actually -- that's what the industry cloud is all about. It's really finding everything in one place. And you've seen that with Fusion. We are essentially bringing in a lot of our offerings within the manufacturing portfolio into one. And that's a great sort of a little bit of a teaser into where we are going with the other industries as well. It's bringing that together so it's a natural workflow. And yes...

Raji, you guys -- Autodesk announced Project Bernini in May of 2024. So can you talk to us a little bit about kind of the opportunity you see there and how this can be kind of infused across the broader sort of portfolio?

So I just want to put that in sort of broader context. And the way I sort of simplistically is the sort of the sexy AI stuff and then the -- which everyone is focusing on, but there's also a bunch of unsexy AI stuff, which people need to focus on more. So on the sort of sexy AI stuff and the sort of sexy end, the sexy is the model, which is Project Bernini. That's our multimodal model, which we've built. We haven't used Anthropic or OpenAI. We built our own model. And the important thing to understand about it is that we care about 3D AI, not just AI. And the reason I emphasize the 3D, using a simple example, is that if you type in or scan an image of a water jug into any model, all of them will produce wonderful different jugs for you to look at. But if you rotate them and look through the top of them, few of them will have a hole in it to hold water. That's a problem for our customers because they're trying to build stuff in 3D. They don't care what the surface looks like. They care that it is in 3D. And at the end of the day, [ the bit ] can stand up in the real world and obey the laws of physics. So drawing inference in 3D is a big deal in being able to do that. And by the way, we've been thinking about this problem for 8 years. We didn't suddenly have a [ hard tag ] last year. That's why we've been thinking about it a long time. The other thing you have to sort of think about is data intensity, which is if you're inferring 3D, that's potentially a lot more data, which means slow and expensive, unless you can solve the data intensity problem. So to make inference using less data, again, that's a problem Bernini is solving as well. There's a bunch of other things, which we talked about on our Q1 earnings call. The point is, though, that in terms of kind of the sexy end of sexy, we are leading the industry by miles, and none of our peers have announced a model. That's the good news. The bad news is that for it to be adopted widely, it's going to require our customers to fundamentally transform the way that they do business. So today, the design process is more of an additive process, literally brick by brick up from the bottom. Whereas with AI, it's going to be more of a subtractive process, where you want to put design parameters in and it will come up with 500 different versions of it, and then you subtract down to what the final version is. And that just means it's going to take time to do, first thing. And secondly, there's a data intensity issue because it's going to -- a lot more data points, and I'll come back to that because that then is a good segue into the unsexy stuff, which I want to talk about. On the other end of the sexy spectrum, which we're doing today, which is kind of automate and correct, which customers don't want to do. And so we gave you a good example of that on the Q1 earnings call. which is the drawing automation in our Fusion, which is it takes our customers today 3 days to produce the final 2D version of a 3D object they've designed. And they have to create the dimensions and the drilling instructions so that it can be manufactured. And it takes them 3 days -- a qualified engineer 3 days to do, which they really don't want to be doing because it's incredibly dull work and it's error prone. And with the drawing automation, depending on the complexity of the object, we can do that in seconds or minutes, something that takes -- a massive time saving. And so the way that gets monetized today is by delivering more value to the Fusion product, which we can then monetize through price. And we can do that because it's a relatively low-compute form of AI. So going back to Bernini, the problem with Bernini and with high-compute AI is that if you try and put high-compute AI through a subscription business model, you're going to blow a hole in your margins. So you have to come up with other ways of charging for it. And that's where the unsexy stuff becomes important. So first unsexy thing is the business model. So you can't use subscription, so you have to have some form of consumption or value-based pricing. So the fact that all of our EBA customers are already on consumption pricing, we have a mass market version of that, Flex is already available. So we've already productized the consumption bit. But sitting underneath that, and this is why I emphasize this, is a bunch of plumbing work we've been doing over the last few years in our platform to enable that work. And again, few of our peers have begun to do that work. That takes time. The second thing Raji has already talked about, so I'm not going to spend a lot of time, is around data accessibility. It's if you have all of the vast majority of data is locked up in files and extracting that data takes time and effort, which means you're going to have slow and expensive products if you're trying to scale it. So you have to blow up the files, have accessible data in data lakes or common data environments if you're going to scale it out. So that's work again we've been doing for years, and our peers will need to do it as well. And then the sort of final thing is around trust. And there's sort of multiple dimensions of this, but what it boils down to is that our customers have a bunch of data but don't have sufficient data to build their own models. We don't have data, but we have the capabilities for AI and the ability to act as Switzerland, as partner. But we have to trust each other to do that, and trust has sort of multiple dimensions around data security, around IP protection, and just the way we interact it. So it's -- even though we have access to a bunch of data, it's how we do them, how we interact with our customers is really important as well.

And I think building on what Simon said, he's painted a great picture of both the opportunity and the challenges in getting there. From an opportunity perspective, one of the key things that Bernini, going back to our Bernini topic, is having this as a research experimental open out there is actually going to teach us a ton that, from a learning perspective, that we can take back to commercial data. And also, real use cases. As you pointed out, we are in the business of not just digital but digital to physical and physical back to digital. That means precision and accuracy. The minute we don't sort of provide that as a value, it's going to be a trust breaker. We got to build trust by providing that sort of accuracy and precision. So that's where a lot of our time is spent is ensuring that this 3D -- you talked about the hole in the middle of that weave and how do we bring that to buildings and manufacturing and parts and assemblies, that's going to be the crux of where we are spending time. And we plan to do this iteratively with the customers. So there's a handful of customers who are very interested in actually participating in this journey with us. And we -- if everybody is fortunate enough to have that sort of engaged customer base, and a lot of the AI stuff is you want to test your way into it. And this is definitely where Bernini is going to give us learnings that we can take to smaller models that we can actually roll out through this process. But immediate focus is on the examples that Simon gave, which was around automation and conceptual augmentation in design and even timely insights, things like that, that is going to be multiplicative in terms of the productivity and efficiency we can give to our customers. They see that as value before we get to generative options for 3D. And that's the cool stuff and the paradigm-shifting stuff, but there's a ton we are doing in terms of productivity enhancement there.

I'm a mechanical engineer that I used AutoCAD LT before I used Windows, but I'm going to ask you a financial question. I'm going to ask you a financial question. This whole transaction model, walk us through the nuts and bolts of it. How should we interpret the moving parts and the complexity? Maybe break it apart and put it back together for us, if you don't mind.

Yes. So there's a sort of operational element to it and a financial element to it. So split them into the -- and it is complex. And by the way, we added quite a bit into our Q2 opening commentary to explain the moving parts in the financial bit. So if your heads explode and it's bleeding, then more of the information there. So on the operational side, we have had a multiyear transition from -- because we've been around for 40 years as a company, back in the day, we were shipping CDs around the world, physical CDs, monetizing the value of our product upfront, entirely upfront, and distributing through a wholesaler and a retailer. When we went online, we kept most of the dynamics of the offline model. We were just distributing physically, but we kept monetization mostly upfront with perpetual license and maintenance, and we kept the wholesaler and the retailer. And then since Andrew became Chief Exec, we've essentially been accelerating the transition to more of a SaaS model. So first of all, moving from monetization upfront to monetizing over the lifetime of the product. Also, there's the subscription transition. We then did the same thing to the cash flow, so moving from multiyear build upfront, so monetizing upfront, to a multiyear build annually, monetizing over the lifetime of the contract. And then the sort of final phase is removing in most of the business, in our indirect business, the wholesaler from the equation and going to a direct billing model. So the new transaction model, to your question, is that final bit, just to put it in context for those of you who haven't been paying attention, is that final bit of the equation. And what it really is designed to enable us to do is a few things if you sort of think revenue and the unit cost of servicing that revenue. On the volume side, today, in 2/3 of our business, which is serviced by the channel, we don't really know what's going on with the customers. So if there's a customer where we're selling into 2 separate departments, we don't know. If those 2 departments have different attach rates of our products, we don't know. If they have a different mix of the products, we don't know. If they have different usage rates, we don't know. If there's a third department which isn't being serviced, we don't know. We will with the new transaction. So we can then go in and actively manage in 2/3 of our business in a way that we haven't in terms of attach rates, in terms of usage rates, in terms of selling more products to them. So we think, firstly, it will show up new opportunity for us to grow the business and expand our relationship with our customers, first thing. The second thing is that with the old model, we didn't set the price to the customers. We basically sold to our resellers with a discount, and then they sold it at a smaller discount to our customers. And so we didn't really control the pricing in our end market. In the new transaction model, we will control pricing in the end markets. And so that will have some consequences, which is if you look at our resellers, and I'm simplifying, there's 2 types of resellers. You've got sort of value-added resellers who are taking our products and building services on top of them and doing the hard work of helping our customers undergo their digital transformations. Let's, for the sake of argument, call them value-added resellers. And then there are other resellers who are just taking our products and selling it on actually more like a shop window. That business will become harder because their ability to flex on price will become harder. So that business will likely shift either to the value-added resellers who -- if they want service, or it will shift directly to us through the eStore for those customers who are just much more sensitive on price. But the bottom line is it will give us more price consistency in the market and more ability to test and learn on pricing in our end markets. And then on the sort of cost side, obvious one is that we will be paying our distributors less in -- across the big bit of the business. That will be -- net effect of that will be less because we have to build and stand up and run our own billing platform. But it will also allow us to lower and manage the unit cost of servicing the business. And we've given you 2 examples of that. The first one is the self-service functionality is that with the new transaction model, customers -- and this may not sound like rocket science because it isn't, but we haven't been able to do this until now, will be able to go in and purchase -- do stuff themselves, basically. And that means there will be fewer human touch points. They need just fewer human touch points for any given transaction, on average. What they will also be able to do in addition to that is they'll be able to do what we call co-terming. So today, because we have that fragmented reseller environment, you have lots of different contracts with customers, which means, I don't know whether it's every month of the year, but many times during the year, they're going to be talking and renewing and spending time renewing contracts. And that's time for them, and it's time for us in terms of effort. And so what co-terming allows them to do is to unify all of the expiration dates to the same date, so we're only having that conversation once a year. And by the way, when we have that conversation, we're typically having a more enterprise conversation so we can then start selling them more in. We're only doing it once a year, so there's less effort per dollar of transactions of doing that as well. So it's really around allowing us to serve our -- integrate more closely with the customers and serve them better and serve them with more stuff and also to reduce the unit cost to service them. And that, over time, allows also to drive sales and marketing efficiency and margin over time. So that's the sort of the operational side. And then the financial side, and this is where your ears will start bleeding a little bit, is you have to think about sort of the gross and net effect. Because at the same time that this is happening, there's an accounting change going on, so a geography change. So we paid our resellers about $600 million for their services in fiscal '24, so in calendar '23. And that was done -- that cost is recognized pre-revenue, it's contra-revenue. So we have gross revenue, contra, and then we report net revenue. With the new transaction model, because it's what's called an agency relationship, that $600 million is shifting into sales and marketing. Now if you just lifted it and shifted it, your revenue would be $600 million higher, your cost will be $600 million higher, your dollar profit before you optimize would be the same. But your margin percent would be lower, okay? So bearing that in mind and also do understand that, that $600 million will shift ratably over time, it doesn't just go big bang at one moment, you need to have 2 concepts settle in your head. The first one is, excluding that geography effect, what is happening to underlying margins? And you can see what's happened to underlying margins. If you look at Slide 6 in our Q2 earnings deck, they've gone up by about 300 basis points over the last -- on an underlying basis, over the last 3 years. And essentially, the work we're doing on the new transaction model and other stuff will help us continue to drive that underlying margin higher -- the percentage higher on an underlying basis. But then you have to overlay over the top of that, when you're modeling it, the reported margin percent, which is that -- the rate of shift, which all else being equal, is it's not a drag to profit dollars, but it is a drag to the margin percentage. And I know it's complicated. It makes my ears bleed and I am sure it makes yours bleed, but just to make sure everyone sort of has an idea of all the moving parts.

Certainly. High single digits, [ a point shift ] in the sales and marketing.

Yes. It's $600 million. So it's a meaningful number shifting from -- and that provides a tailwind to revenue growth. Well, a tailwind to cost growth or a headwind, depending on which way you're a optimist or a pessimist. But before you optimize, it doesn't affect profit dollars. But then as you optimize, you start getting the benefit on the profit dollars.

And how do you go about the process of optimizing that? Because there's so much more you're going to learn in that expense number that you can be more efficient with, maybe with channel partners. The way of doing business has been a 40-year thing, to your point. So how do you uncover efficiency?

So we'll give you more details on that over time. But just to sort of give you some understanding of the process, which is we built -- this is not something that sort of magically appeared in November, which is when we really started first. Actually, we were talking about it beforehand. But the -- this is something that's been in the process because we've been building the billing platform for the last 3 years and testing it. And then we launched Flex on top of it about 1.5 years ago. We then went live in Australia, which is when we started talking to the market more fully about it. And then we had the U.S. launch in June, and then we've got the Europe launch next week in about a week's time. And then Japan, which is the last big one in November.

Did you build this platform?

I mean that's -- the back end is all multiyear work that we have gone through, so it's actually pretty significant. The gravity of what Simon is talking about here is, I don't know if people actually understood like the multidimensional change your partners going through from becoming resellers to VARs, that's changing their world, customers buying in a different way. And then you have us, from Autodesk's perspective, actually taking off -- training wheels off and directly interacting with customers who have been sort of -- partners have taken care of it, so there's a multidimensional transformation here. And I think the key thing was this was in the works, so the financial systems had to be rebuilt, some of the back end. I mean it's a massive effort.

That's so massive. It's huge. I don't know how you guys decided.

It was hard.

Why did you do this? It's like nuts.

Yes. But I think that is the point that I think Simon is making is it is essential for that future growth, for us to actually have that interaction with the customer, our partners to become more VARs versus resellers, and customers to be able to interact with us. I think it is essential for that. So it's sort of a necessity that we had to go rebuild this. I think that is a key part of this.

That's my tough question for the end. Why would you do this? It's self-inflicting.

Just to sort of finish that thought is that the -- and to your question, your previous question, and then I'll answer that one, which is that this is a complicated process. And -- but at the core of it, what we're doing is we're transferring the ownership of the relationship from our channel partners directly to us. That is the most important thing that's happening. And we're very focused on that happening and not messing it up.

What could be the risks that are cognizant -- that you're cognizant of that -- with this?

Well, so to manage those risks, we need to keep our channel partners focused on that process. We need to keep our employees focused on the process. And then that affects how we communicate the process as well. And so this is a long preamble to say, to your question, which is I'm not going to precisely answer your question because it -- so we were -- are communicating to you as part of managing that process, and it is complicated. So -- but the point is that we've been building the billing platform. We have an implementation process, which we are in the process of doing. And then we've said there's an optimization phase. And in the same way that we have very clear and detailed plans on the build process, we have very clear and detailed plans on the implementation process, and we have very clear and detailed plans on the optimization process as well. And we are sequencing those and implementing them in a structured way so that we don't mess everything up before [ channel partners join ].

And Kash, maybe one more addition to add to Simon is we have a choice because of the fact that customers are going to want to buy in a new way. And this is actually an important aspect of not just growing our customer base -- actually, future customers. And they may not be the big accounts, but for -- as we go down and look at mid-tier, they don't want self-serve. They're going to want to be able to directly interact with a company and buy. And it is a necessity. We need to go through this process and with all its sort of challenges to be able to get there.

We have time for one or 2 more questions.

Simon, I want to touch on IIJA funding and kind of where we are in kind of that allocation process. How many of these projects are shovel-ready? How levered is Autodesk's AEC business to this relationship?

So the way to think about the IIJA is it's a cherry on a very large cake. It's that there are billions and billions of dollars spent on infrastructure every year. And the IIJA adds a bit of extra, a cherry on that very large cake. So it's interesting, but it's not game-changing in terms of the overall market. But that's not to say that it's not important. But the bit that's important is not the bit that everyone's focusing on. The bit that's important in the IIJA is the $100 million that is being spent to incentivize the federal and state departments of transportation to digitize their workflows, connect them into cloud. And as part of doing that, they're also adopting BIM. So whenever -- sorry, building information modeling, for those of you who aren't paying attention. So whenever somebody says connected workflows in the cloud and building information modeling, the person at the front of that -- the company at the front of that team will be Autodesk. And so what it's doing is it is opening up the cake as an opportunity to us, so rather than just the cherry. Now because it's government, it will take a decade or more to happen, but the cake is very, very big. It's probably one of the biggest untapped or under-penetrated opportunities for Autodesk globally in construction is the U.S. infrastructure market. There's an incumbent in that market. And what we're slowly doing, as they adopt digital workflows, is being able to crack into that market. So if you look at sort of Pennsylvania Department of Transportation, it's a good recent example of that. Historically, they have single-sourced on a competitor operations. And what they did in their most recent is they've now dual-sourced in specifically for their next-generation data platform on Autodesk Construction Cloud. And what will happen over time is that the proportion of the business within Autodesk Construction Cloud will grow. And so you're going to see that. We hope and expect over time happening in other departments of transportation slowly over time, and that will give us access to more and more of that very large infrastructure cake. So yes, focus on the cake, not the cherry.

And Raji, to bring you back in...

This will be the final question.

What was that?

It will be the final question.

On Autodesk Construction Cloud, you guys have taken kind of a buy-versus-build approach. I mean I imagine that you've been very involved on the back end integrating these products. So how complete is kind of the Construction Cloud platform today? Any product gaps you need to fill?

So...

Yes. Go ahead, Simon.

I'll start. So we basically -- if you think about construction, we've got pre-con and construction. And then within construction, you've got the key workloads of project management, site management and cost management. All of those are in the integrated Build product, so we basically made a bunch of acquisitions and then integrated them. We're in the process of doing the same thing with our preconstruction offering. And then do you want to sort of add to that?

Yes, I think a lot of the work that we're doing is not just integrating but actually figuring out what is the next suite of products. So we're going through that process. So it's not just about pulling it all together. So ACC is our sort of leading one in that Build portfolio. The second thing is just actually making it -- a lot of what you talked about, the opportunities and the cherry and the cake is about data regionalization. FedRAMP Moderate, that we just got authorized for, these are the key elements that are required for the cloud products, so that's what we are building in the -- into that platform. So there's -- so that you can expand beyond what it is doing today.

Fantastic. On that note, thank you so much for participating in this conference. I hope you have a productive set of meetings and wish you good fortune with your new journey.

Thanks for having us.

Thanks so much.

Thank you.

Let's give them a round of applause, please.

Thank you.

Thanks, everyone.