Beeks Financial Cloud Group plc (BKS) Earnings Call Transcript & Summary

Good afternoon, and welcome to the Beeks Financial Cloud Group plc investor presentation. [Operator Instructions] The company may not be in a position to answer every question received during the meeting itself. However, the company can review all questions submitted today and publish responses where it is appropriate to do so. Before we begin, I'd like to submit the following poll. And I'd now like to hand you over to Gordon McArthur, CEO. Good afternoon to you, sir.

Thank you, and welcome, folks. I'm going to, as usual, if you've listened to this before, I'm going to do the first few slides and overview, and then I'll hand over to Fraser, who's going to go into more of the financials, and then we've got Q&A at the end. So who are Beeks? It's always difficult on this because we don't know who's attending and who's not. So Beeks Financial Cloud are cloud computing company based out of Glasgow that solely cater for capital markets, right? So we do cloud computing and we focus on the data centers that trading activities take place in. They tend to be specific financial services data centers, whether that be owned by an exchange or owned by a third party. And we deploy our infrastructure in there and connect to various endpoints within the data center and then we tie all that up. We tie all these data centers up together with a backbone that runs around the world. So just kind of going to cover some of the highlights from the first half of the year. So Proximity and Exchange Cloud is what we're going to spend most of our time talking about today. The percentage of revenue and the 2 new offerings that we see as the kind of future of the company has doubled from 10% to 20% of the first half revenue year-on-year. So kind of as expected, and we hope that will continue to grow. We've got a substantial Proximity Cloud and Exchange Cloud pipeline. And we're now starting to see material contributions from the 2 new product lines, which is good. As we announced on Monday, we have won our first Exchange Cloud contract with a partner in the crypto space. So previously, everything we've done is equities asset class specific, and we'll continue to focus there. But we've signed a deal with Kraken, one of the larger crypto exchanges, one of the more sophisticated institutional flow type exchanges for them to license Exchange Cloud in the European data center. And we've done that under our new revenue model, which we will take some time to explain today. We've been looking at this revenue model or kind of revenue share type agreements for the last 12 months, and we have decided to move forward with it. Not to say that we're ditching or halting the old model where it was a long-term contract at X amount per cabinet per month. It just means we have 2 commercial models that we will use in different situations depending on what is fit for both the client and Beeks going forward. So not walking away from the old model, but we are introducing this new revenue share model, which ultimately will be more profitable for the organization, and we'll continue to cover that as we go through, but materially shortens the sales cycle. So we are not asking customers to budget millions and millions of dollars, which usually tends to go into the next budgeting cycle. The revenue share agreement means that we are going into it with a pretty much 0 commitment from the client and we are willing to take the commercial risk on that upfront in certain situations. Operating cash flow was up 21% year-on-year. So we've delivered more profit, more cash than we have in previous years, which is obviously good. And as I say, Fraser will cover more in-depth financials as we go through. So this slide hasn't really changed that much in the last 3 years. We've got a new [indiscernible] performance for any period of time. You'll see there's a new Edge Intelligence section here, which we'll cover. But 90% of our revenue is still trading infrastructure compute network within the data center, virtual machines, and that will continue, right? That's been the heart of the business. We are an Infrastructure as a Service company, and that's what we do. We also have, as I said at the top, our low-latency wide area network. We don't run cables ourselves, obviously, under the sea. We buy routes from various cable carriers around the world and pass that together depending on our requirements. And we've got a small professional services division where we make a couple of percent of revenue. The new part of this is a new product we're building called Edge Intelligence. It's kind of related to our analytics offering where we capture as part of our analytics, and it doesn't have to run on Beeks analytics, it can run on other analytics packages out there. Banks and brokers and exchanges generate and capture huge volumes of data and it's market data, right? So it's very low-level FIX messages, text files that are [indiscernible] the biggest feed in the world. Opera can pump out 100 gigabits of data per second and the very low-level FIX messaging language. And that's -- it's got a lot of value in it, but it's very difficult to do anything with it because of the size of it. And we capture that and various other analytics offerings that capture that data. So what Edge Intelligence is going to do and the first use case for it, we've been building it for about a year is to look at anomaly detection. So this is -- it's more of a machine learning environment at the moment where we've trained some machine learning models to look at years and years' worth of data to understand what is normal. And when you understand what is normal, you can very quickly identify what is abnormal, what's an anomaly. And anomalies might last for a millisecond. They might last for a second, and they might have -- and they all have some data attached to them. But manually sifting through all that data to try and find them is impossible, right? So our Edge Intelligence platform will, in real time, alert a bank, broker, HFT and exchange that there is an anomaly somewhere in the system, and we're scoring that at the moment between 0 and 100. The higher you get towards 100, the more serious the anomaly is. So what that will do is alert in real time that there's something strange going on in your environment. At the moment, we're not telling you what. But what we will do is attach a link to the packet capture, and that specific piece of data that is suggesting there is something going wrong in the overall trading infrastructure, so trade operations people can do a deep dive into it. And then there's a lot more use cases as we go on. So we've been building that for a year. Still about 6 months from being ready, but we're working with a couple of large organizations to guide our kind of strategy there and the product road map and a lot of technical challenges still to be overcome. Can we do it at the scale that's needed? Still to be answered. And can we ensure the accuracy at scale? So that's going to be the next 6 or 7 months of work until we get into version 1 of a product. And that is how we build products. Historically, Exchange, Proximity Cloud is we have an idea, we go and talk to clients about it. If they verify it's useful, then we work with them to try and productize it, and Edge Intelligence will be doing definitely the journey that we did with Proximity and Exchange. So the 3 parts of the business. Private Cloud is still a big part of the business. Still continues to grow every month. And these are smaller, sometimes month-to-month, sometimes annual deals, sometimes 24-month deals where a client will buy a level of compute from us, some shared networking, sometimes it's completely dedicated networking. We basically build a bespoke private cloud for a client. It can be anywhere between GBP 1,000 to GBP 20,000 a month type deals. We've got some very large clients in there. And that has always been the majority of our revenue since we IPO-ed. As Prox and Exchange are built, we expect that they will take over as the main drivers of percentage revenue for the group. Proximity Cloud was our first foray into a dedicated cabinet that we deliver to clients. Single-use platform for use for a bank, broker investment manager, whoever wants it, they will use it to run their internal workloads. And we've had some success in that. We signed our biggest deal ever last year around Proximity Cloud for a large bank that is currently rolling out. And pipeline for that continues to grow. So single-use platform, rack-based end-to-end trading infrastructure. Exchange Cloud is very similar to Proximity Cloud. The only difference is it's multi-home. So multiple clients can exist securely in the same rack. So we can effectively break up the infrastructure in that cabinet and assign it to multiple clients and nobody can see or be affected by the other clients' trading activities. Proximity Cloud is kind of GBP 1 million to GBP 5 million can be bigger. Exchange Cloud over the period, Exchange Cloud is very much a land and expand type of contract, and I think we're going to go to JSE again in the next slide. So it is materially bigger as we go through the contract as we usually start off with a couple of cabinets and then that can grow materially. Just a quick overview of some of the figures. So all point in the right direction. Again, historically, we've always grown between 20% and 30% kind of year-on-year. This was kind of in line from a revenue point of view. EBITDA up 25% and PBT up 37%. Positive net cash again. And 75% to 76% of our earnings are on a monthly recurring basis. So it's not a best quarter -- best half ever sorry, by no means a worse, very solid performance for the business. This shows our Exchange Cloud footprint, right, as it continues to grow. And we hope to color in more of this map over the next year to 2 years. So we started with JSE. ICE we're actually the very first ones to license it in the ICE Global Networks division. JSE were the first to put it in their core data center. The Tier 1 exchange, we have made announcements that is Nasdaq. And I'm sure there's going to be a lot of questions here, but we are still under NDA. I'm not going to be able to talk much about it. It's proceeding as planned, but we are still under NDA. So limited with what we can say. Mexico is actually the largest Exchange Cloud deal to date, signed in February. They are going for production and DR in the primary and secondary data centers. And that will -- that's kind of rolling out at the moment. We are building that in Glasgow and shipping it out over to Mexico in the next few weeks to get that live and -- before the end of the financial period. And emerging markets -- sorry, just to finish this. So Kraken first foray into crypto, very short sales cycle on the revenue shared deal model. So it's usage-based, and I think we've got more slide to cover as we go through that. So for the last 12 months, I've been talking much more about emerging markets. Emerging markets are where we are focusing our attention with Exchange Cloud. It might be kind of opposite to where people would naturally assume. But some of the emerging markets are operationally difficult for Western companies to operate in and to trade in. Some of the biggest growth rates in exchanges, in fact, all the biggest growth rates percentage-wise, volume-wise, in exchanges are in emerging markets. So there's a lot of companies want to trade in India and Saudi Arabia and Mexico and Malaysia, Philippines, but they don't have entities there. They don't have import licenses, they don't have staff there. So there's a pipeline of demand, there's operational difficulty, and that is where we actually see the biggest opportunity. I think last year, when we talked about this, I thought it was about parity between developed and emerging. The more we've gone into this year and with the pipeline that we've got, the emerging markets are probably going to be the dominant source of revenue for Beeks going forward, which, as I say, is probably opposite to what you would think, but there's growth potential there. A lot of the developed exchanges have got a lot of the market participants and we will pick up -- as we sell into the developed markets, we will pick up business and hardware refreshes and growth and new participants, but there's a backlog of demand for emerging markets. There are certain exchanges that I could probably fill 10 or 15 cabs in immediately if we can get Exchange Cloud launched in emerging markets. And that's a material source of revenue for us. So look -- as we go forward, look for us to make more announcements in and around the emerging markets as well as the developments. But that's the current map. And as I say, by this time next year, we hopefully we present it and there's a lot more blue in there. This is a JSE slide that I think we presented last year, but it's a really powerful slide, and it's worth going over again. So JSE are, I think, 28th largest exchange in the world, largest in Africa by a long way. We signed them a couple of years ago, and we thought overall, the contract will be worth GBP 2 million to GBP 5 million a push over the total contract period. We've underestimated that substantially, right? And JSE have extended a number of times, and we'll continue to do that. Pipeline is still significant with new people on capacity and they're getting towards the limits of their existing capacity. So we fully expect JSE to continue to extend. We're putting in a DR data center at the moment that is a secondary data center, banks cannot just have primary. They need a fill over if there is an issue in the primary data center. So that gives us a second site to work with. So as we get more mature with JSE and I spent a lot of time down there with the clients, with the exchange themselves, we fully expect JSE to spend between GBP 15 million and GBP 20 million once we get into the kind of fourth and fifth year of that contract. So that shows you the land and expand opportunity without being doing JSE as a service, one of the smaller exchanges globally, it shows you what the potential is for some of the developed markets and some of these emerging markets in terms of revenue over the contract period. And we fully expect after the initial contract period of JSE, they will extend. We look at these as 15- to 20-year kind of deployments. So the new revenue share model. So we've been thinking about this for the last year. And I think what finally pushed over the line for us was Mexico. So Mexican Exchange was about a year and a bit of maybe just under 2 years of sales cycle. A large part of that sales cycle was waiting for them to get budget unlocked. So they had decided they're going to go forward with the platform. We built a business case, has been approved by the project team. And then they have to make a submission into next year budget cycle to get the money to sign a contract with us, right? So their budgeting cycle was January to December. If we had taken that away and the revenue share model had been there, we would have been live in Mexico, and we probably would have been extended the contract already. So we've had a material delay because of budget constraints of the exchange. Mexico is a big deal. It's one of our biggest deals we've ever done. So it's not an insignificant spend for them. So that is done in the classic model and we charge GBP 20,000 a month or GBP 25,000 depending on what it is, GBP 2,000 to GBP 25,000 per month per cabinet over multiple years. And that's what we commit to, that's the same. The new model -- and Fraser will talk about the revenue rec on it as well. The new model, we are -- and we're not throwing away the old model. The new model is we sign a rolling 30-day contract, no long-term commitment. Don't pay us anything on day 1, but we will take a percentage split of that revenue, 50-50, whatever it is. We're not going to make it available to everyone. And in certain situations without going to stuff that's been potentially sensitive. As the exchange commit to certain activities with us, we will allow that revenue share model. So over the period of the contract, we will make more money because when a cabinet is filled, we're probably getting about 1/3 of the revenue at the moment on a full contract, whereas we will get half or there or thereabouts when a cabinet is filled under the revenue share model. So just to kind of -- it's very long-winded, very nuanced, right, but bring in sales cycles, no need for budget approval, no need for executive steering group approval, no need for Board approval, materially easier for an exchange to commit to. Materially more profitable over the term of the contract for us. There is a lag until that they are filled, the cabinets are filled. But once they're filled, that makes us more money. We have got a level of commercial risk. But again, with the kind of stuff that may be commercially sensitive, we can lift and shift that environment if we -- if it's not performing, right, if there's no demand there. So it's not like we deploy the infrastructure that we've written off, we can move it to another data center or another location if we want to. Fraser, I don't know if you want to touch about the different revenue recognition.

Yes, sure. Good afternoon, everybody. Yes, I think -- so the revenue -- the accounting is very different under the 2 different models. As Gordon says, we're doing this for commercial reasons, but it does have a difference in the accounting. So on the traditional model, we take about half of the contract value upfront on day 1, okay? So if you looked at our results over the last couple of years, you see we've got now point in time, we've got point over time right now going back before Proximity and Exchange Cloud, we always had this nice recurring revenue profile, which is smooth and always kind of go upwards. So the revenue share model, what that does is it pushes it back into a monthly recurring, right? So -- and it's really because of 2 things. One is there's no contract value on day 1. So on day 1, we will get potentially 0, right, whereas before there's an actual contract value. So the accounting is very different. So we've had to reflect that in some of the forecast because we've always assumed historically, there's this upfront element. However, we're doing it for the right reasons and commercially it's better, as Gordon said. So the cost is exactly the same. So this is purely accounting. And if you take the fact that we're -- we've assumed in this slide here that it will potentially take a year to get to full capacity. If it does that, we get about 1/3 more revenue overall, but straight to profit because the cost is exactly the same. So as we get -- as we deliver more of these, we'll become a bit more okay with how we forecast it. Will it take 3 months, 6 months, 9 months, 12 months to get to full capacity, we don't know yet, but we will get better at predicting that as we deliver more.

So again, we're kind of showing we're investing in the sales team with a significant investment in the sales team, and we'll continue to hire to tackle these emerging market areas. So Adam Bradley who joined as Head of APAC invested in some London Resorts. Swad, who's kind of head of presales over in New York to support Jerry. So the sales team is -- the direct sales team that are targeting exchanges and banks and brokers has grown substantially. We're very -- it's a very nuanced and involved sale. So we're very picky about who we employ in that sales team because you have to have domain knowledge, you have to have the right contact, you have to be able to understand and articulate various parts of exchanges and banks or value proposition. So we hire what we believe are the right people in the right areas. I kind of jumped jump Edge Intelligence slide. So that kind of goes -- as I said, the first thing we're looking at is anomaly detection, but we've got a bunch of use cases as we add on AI modeling and AI forecasting into the core middleware that we've already built. We've got a number of customers working with us as we did in Proximity and Exchange Cloud to shape the requirements, right? And we kind of tend to give -- if they want to buy, they kind of get a bit of a sweetheart deal as a development partner. But we try and target the bigger ones, right, that we can capture -- you can do the monster accounts then especially because a lot of the problems here is scale. If you can do capture and model the bigger accounts than the small ones become easy. So kind of launching next year, we have invested in it. So we have started to articulate out to the market. And hopefully, by the end of the calendar year, we are starting to see some signs of revenue flowing in the Edge Intelligence world, which gives us another revenue stream to go. I'll hand over to Fraser to talk more in detail about financials.

Okay. thanks, Gordon. So as Gordon says, we've got a decent set of results in terms of sales growth, right? So looking back since IPO and over the last 5 years, we're typically around 25% to 30%. So slightly at the lower end of that in this half, but still relatively consistent. And a big driver of that, as Gordon mentioned at the start, was Proximity Cloud growth in the period, right? So that was around Proximity and Exchange about 22% of the overall revenue. And as we look forward, we expect that trend to continue upwards. The deal sizes are, as Gordon says, the large U.K. bank that was delivered in the period influenced that growth. But the size of the deals in those products are a lot higher and they are driving the growth there. Slightly lower in terms of the ACMRR stands for annualized committed monthly recurring revenue. So not quite as -- a bit more modest in terms of the growth there, 7%, which is historically quite low. We did flag in the results that we had some higher churn in our legacy business, historic part of the business, it's about less than 10% of the overall business. Again -- and the reason for that, we changed software provider for very good commercial reasons. And as part of that project, a lot of these legacy clients took upon that situation to clean out some of their historic estate. So a bit of a one-off and not expected to recur and didn't impact overall profitability because of that better software license charge that we've now got going forward. So that's in terms of recurring revenue. So again, pleasing to see that the profit is going in the right direction. So as Gordon says, we've had 25% growth in EBITDA, 37% growth in profit and improved profit margins again. That's in spite of some of that churn and also the Market Edge Intelligence that we're investing in during the period we're expensing. So we're able to grow margins 1% or 2% while still investing. So the profit is going in the right direction. Staff is largely unchanged. We've changed the headcount slightly. We've invested in sales, as Gordon says, but we also looked at some of the software development team. And that's got some natural attrition as we've had a new set of software streamline some of the processes in there. So we get a bit more efficiency out of that, which is good. And we're also capitalizing a bit less that as well because we've now established products. So again, profit margin is still improving with those going on as well. So -- and as we look forward, particularly around Exchange Cloud and in terms of the different model, we believe that there's quite a bit in terms of leverage that we still get. And if you look at the forecast, we are forecasting higher profitability and we're comfortable based on what we know that's going to transpire. Again, profit is one thing, cash another. So over the last 18 months, we have started generating cash. So cash is going in the right direction as well. Operating cash flow is up over 20% to GBP 5.8 million. And the sort of the reduction in the net receivable/payables. There's kind of 2 main drivers for that. Part of that is the upfront element of Proximity Cloud, where we take the revenue upfront, but we collect the cash over time. So there's a big accrued income increase, but that will start unwinding over the next number of years. So looking back the other way. And then I think we mentioned in the results, we got a delayed post year-end payment of just over GBP 1.2 million early January. So not ideal, but nothing untoward on that other than internal process and it was paid early January. So in reality, our cash should have been GBP 1.2 million higher at the end of the period. We continue to invest. So PPE/infrastructure, that's the cost of our putting the infrastructure into the data centers. We utilize a bit of our existing inventory. So GBP 1.2 million excluding stock, which reduced by about GBP 1 million. So we'll always have that cost in our cash to put that infrastructure out there. And then the intangible assets is the cost of our development team. We've repaid some debt in the period. So the balance sheet is in as good a position as it's ever been since IPO. We are generating cash. We have money in the bank, and we have low levels of debt, and we expect that trend to continue.

All right. Back to me. So as I said at the top of the call, we've kind of maintained our 20% to 30% top line growth. Business is 13 years old. It has grown every single half year since it's existed. So I think we've got a good track record of delivering. Now we've been in public domain for 7 years, I think. So again, we fell within our kind of expected area of growth. The business, as Fraser said, the big thing was getting to cash generative, cash positive. We are there now. We expect that to continue to increase, gives us the ability to invest where we want. We don't have to -- unlike some of our old still MSP competitors, we're not using a lot of free cash to pay our private equity owners interest payments. So we are available. We can invest and innovate within the business. And I think we're building up a track record of being company known for innovation in the space. There isn't a lot of it, but that's where we'll focus and continue to innovate going forward. We're not a big M&A player. People have followed us for a while. We do look at stuff occasionally. We may or not do something small. I don't think we're looking for big material M&A opportunities. There's more bolt-on feature function where we would struggle to build ourselves because there's only so much we can do. So being cash generative helps that as well. We got a lot -- with the exchanges, when we sign an exchange, their sales team are effectively selling our services. We've got very strong relationships with IPC, our biggest partner. We have a lot with them in terms of joint bids. We are a customer of theirs, they're a customer of ours, so a good partnership. And we really are -- the market is coming to us, right? I think proving that we have deployed this infrastructure now it works, adds value to exchanges and banks and brokers, whether it be Proximity or Exchange. We have -- we build a product, we innovate it, we come up with the idea, but the market is certainly coming to us. So we're in a good position. The business is financially strong, and we've got a record of -- a track record of delivery, which is what has been called. So going forward, what are we focusing on? Apart from revenue recognition discussions, which is not really something we want to continue having ever again. The current year is convert the Prox and Exchange pipeline is substantial. We've got more equity exchanges in the final rounds of negotiations. Continue to invest in Edge Intelligence to get that ready and continue to innovate, right? As I say, we have a substantial software team. We have ideas within the business of other features, products we want to develop and continue to automate. Rather than throwing people our problems we need to automate systems more. And we've been doing that. I was telling another institutional investor this morning, before I come down, some of the senior tech team are showing me a diagram of how long does it take for the network team to build out a Proximity Cloud because as we -- on Exchange Cloud it doesn't matter. As we have to deploy more, we need to get better and faster doing them. And the first year we did took us 250 man hours to build that network. The last one we did that took 25. So that is us automating and innovating around the technology map that we've got. So by doing that and not having to continually employ people, that will drive that operational leverage and drive the company. Last slide, and then we've got a Q&A. We're in the right place, right? The market is coming to us in terms of cloud computing and exchanges and banks. As we've been saying for the last 2 years, Exchange Cloud is our main driver. As Edge Intelligence comes on, does it become as big an opportunity? We don't know yet. We genuinely don't know. But the initial conversations we're having are very, very promising. But like any new product, as you build it, you don't know until you get customers signing it. Kraken is a great win for us, first revenue share. And it will really be interesting to see -- it's really unknown for us because it is crypto and it is the first revenue share deal we've done. So we'll try and keep the market updated. I know there's a lot of nervousness after the change in revenue recognition model and what's happened to the numbers looking out into the next 12 months. That will wash out as the revenue share deals start to become more profitable with uptake. So we will try and keep the market updated of how these revenue share deals are progressing. The pipeline, as I say, is strong for Exchange Cloud, and we hope to make more announcements. When I say short term, short term in exchange land is not days or weeks, right? So -- but we are various stages of pipeline from final MSA clauses to doing the initial business case to submit to the steering groups and what have you of various exchanges, so very strong. And yes, we're excited, right? Our job as a director of the company is to look at the long-term health of the business, like what's the long term -- what's the right thing to do in the long term? Forget the noise that surrounds any short-term change and the market does not like change. However, we're in a better position than we've ever been, right? The opportunity is bigger than it's ever been. We continue to grow our total addressable market. I can't quantify that before anyone asks. I never can because no one covers our niche market. So this is a great position. The team is stronger than it's ever been, a couple of really exciting senior technical hires that we'll announce in the next few weeks. Continued investment in the sales team to drive the new products that we develop. So yes, it's been -- it was an okay period in the first half, and we need to head down in the second half and continue to win more business. So I think that is us. So thanks for listening, folks. I will look at Q&A. And apologies, there might be some questions in here that I cannot answer, so bear with me.

All right. First question, has there been any changes in the amount of exchanges around the world making inquiries since you announced new contracts? Yes. We are now getting incoming, right, which we never got before. I'm not going to go into it because it's commercially sensitive, right? And I'm sure our competitors always listen to these calls anyway, so I'm careful what I say. But we are getting inbound from some of the most interesting exchanges in the world. So yes, the more and more we exchange -- the more and more we sign, the more pervasive and easy it comes to sign other ones. And then to the second one, is there any sign of due diligence speeding up? Yes, one of the exchanges that we hope to announce next told us they were not doing light touch due diligence, but because of some of the previous announcements that we made, they are making certain assumptions that some of the bits and pieces, they might look at and have been done. So yes.

I'll pick the next one up. So once a customer buys the Beeks Exchange Cloud system, is there any follow-up revenue in following years, not including expanding the system, i.e., is there an ongoing rental payment or service fee? So I think on that one, so ultimately, you have -- revenue recognition doesn't always equal cash, right? So on the traditional model, you have this half upfront and then the remaining contract value spread over the contract term, right? But that's purely revenue. That doesn't necessarily mean cash. So most Exchange Clouds and Proximity Clouds are on a monthly recurring basis, right? So this change is really only revenue as opposed to cash. Again, someone, please explain the financial impact of the change to revenue sharing model. I think that was a pre-submitted I think we probably answered that.

There are suggestions that your offering might be -- not be unique. Can you give the name of direct competitor in the niche market? I mean where are the suggestions. I don't -- I'm sure competitors are looking at building similar stuff, but track record means everything, right? And so yes. I think where you give an explanation for the late payments event so it's [indiscernible]. It's a big existing client who's been with us for many years. They just missed the payment.

Yes. So regarding the share revenue, retaining hardware ownership model over what sort of period the hardware cost be depreciated. So we depreciate our hardware over 5 years, which is appropriate. Proximity and Exchange Cloud and they would get replaced, but there will be no change to that. And then what are the risks associated with owning that hardware in terms of obsolescence? Again, historically, we've always done that and infrastructure changes, but we're not seeing a risk of obsolescence on that type of hardware.

Will the new revenue model apply to existing customers? No. And there is no swapping between the 2. When you pick -- when a customer picks which model they want, and it might not -- both might not be offered then you're on it forever, right? You can't have a client who says, we'll do revenue share until we prove it and fills up and then you take less. So absolutely not. I mean, I wish we had done JSE and revenue share, they're delighted we didn't.

Does new revenue model mean higher cash outflow for Beeks initially? No, not necessarily. So just now -- so the cost for us is the same, so the cost of that hardware. It then becomes dependent on how quickly that rack fills up in terms of what the payback period is just now. So we are going from a fixed fee to a variable fee. So at this point, the honest answer is we don't know, right? If it takes 6 months to fill the rack, it will pay back quicker. If it takes 15 months to fill the rack, it will pay back that longer. So we don't quite know. It's a bit of an unknown just now.

Well, Beeks have to fund CapEx when the deal is on the revenue share model and you need to raise additional funds? No, there is no difference, right? The old model is GBP 20,000 a month, we buy all upfront. We still buy all upfront. We only get -- we just charge 50% or whatever the percentage is as and when that revenue comes in. We are not -- we will not be raising money for any operational things going forward. If we had to do M&A and we didn't have the money in the bank, then we would look at it, but we are not in the market for raising more funds.

Does...

Do you want to take it?

Yes. Does the new revenue model reported revenue growth will be materially lower in FY '26 before recovering in '27?

Me and Fraser have been fighting about that for weeks. So we're in a very difficult position where looking out in the next 12 months, it is difficult for us to forecast Canaccord or Progressive or whatever because there's a seismic change in our commercial model, right, for the right reasons, right? It means we make more money. But over the long term, it means we make money quicker. So we've really struggled with how do we forecast accurately. It's hard enough, right? Crystal ball gazing 12 or 24 months in advance is very difficult, right? And I know we want -- we thought could we have a more material range. We expect to go back to normal in 2027 in terms of MRC growth. We've always been a monthly recurring revenue company, right? That's much more valuable rather than one-off stuff. So as we get back to MRC, as we've got experience, and what happens with these revenue shares and the reality, then we'll be able to forecast. But we fully expect when we go to 2027, we'll be back into the normality of where we've always been in terms of percentage growth on ACMRR. It obviously becomes harder as we get bigger to maintain the percentages, right? But this new revenue share model should take a year to wash out and then we become more profitable. For Kraken Exchange, 30 million users. Based on your financial modeling, how long do you take to cover your costs? That's a good question, Andrew. Don't have a clue. Like 30 million users, but you've got to understand -- and I don't know, I'm guessing here, but probably 12.5 million of them are trading crypto on the phones, right? And where we come in is market making, buy-side, sell-side, algorithmic trading. A mom sitting at home buying 0.01 of a Bitcoin on a phone is not interested in our services, right? So Kraken is Exchange Cloud. Kraken is an Exchange. As we put in the RNS, we have licensed Exchange Cloud, it's Exchange Cloud.