Denis, share the details of financial results that we issued yesterday. We'll definitely spend some time to discuss the rest of the year. You've seen that we announced a couple of important information, including updating our objectives, so we definitely will explain exactly what it means, the rationale, and we will have definitely some time at the end of the presentation to answer any question you may have, especially on this topic. I would like to conclude the presentation about next, next means first, something we announced a couple of days ago, which is the availability of our new generation processor called Coolidge™2. More globally, I would like to explain, or re-explain for those who follow us regularly, why Kalray is at the heart of this AI revolution, which is ongoing, and to also further explain in detail, okay, how does this market work, especially with a semiconductor perspective, and where does Kalray sit in this overall market, which is, of course, a certain number of segments, as any large market. Let's start with just a summary, in a nutshell, of who we are and what we do. As you certainly know, Kalray, we're a semiconductor company. We have developed a new type of processor. We started developing this new type of processor about 15 years ago. Nowadays, this type of processor has a name, this is called DPU processor, for Data Processor Unit. And the DPUs are considered as one very, very important type of processor that will be deployed in the modern infrastructure and modern data-intensive, you know, type of application to process data faster and more efficient. We are a deep tech company. We've been designing, developing, maturing our technology and our solution for the last 15 years. Now we are really in this state, as you know, where the technology is mature, where the product is available, where the product positioning is, I think, pretty clear, hopefully, and where it's about scaling up our business. That will be, of course, I think, most of the discussion and question we may have about our capability to go further and growth and scale. In terms of a couple of extra information about the company, we are about 200 people, headquartered in Grenoble, with about 100 people there, with about 50 people in U.K., and then the rest worldwide, especially in U.S. Very talented people, a lot of brain and technology assets and understanding. We've got great shareholders, industrial shareholders, like NXP, Safran, Renault, and Mitsubishi, who've been supporting the company for many years now, as well as very, very important partnership for us, one of them being Dell. We will come back on this one. We've got licensing with IBM, with Amazon, so... And these are among our strong partner we managed to put in place in the last few years. In terms of, once again, just to clarify for those who join us, what is the vision of the company and our main mission? The vision is very simple. It was the same 15 years ago, but now I think it's very obvious and mainstream for everybody. The vision is that data is more and more important in our modern societies, and we strongly believe that the faster and better a company can get insights about what is this data, the winner the company will be. And so our mission is really to be a global leading provider of hardware and software technologies and solution to enable all the companies worldwide to extract meaningful information from data. And that involve solutions for making data-intensive application much more faster, much more efficient, much more smarter, as well as the next generation of data centers. Last slide about Kalray in general. So just to re-explain what's our portfolio today. Of course, our core product is our processor, this DPU. It is the result of 15 years of hard work, a lot of investment. And now, as you know, we are shipping the 3rd Generation, and we are now what we call the 3.5 Generation. I will come back on that later on. We have, of course, invested a lot to make this unique processor technology available, easy to use for our customers. That implied first to have what we call acceleration cards, which are cards that can be plugged into servers and that offload demanding workloads. And then on top of this card, we have also developed what we call reference solution, which are reference hardware architecture that our customer can use as such as a product or copy to make their own product. And then software, because software is, of course, very, very important when you build a solution. And we invest a lot in software, essentially, including the acquisition of a company based in U.K., early last year. So today we've got a very comprehensive portfolio for any data-intensive application type of requirement for our customer. Goal being, once again, to process data smarter, more efficiently, and more securely, wherever this data resides. Okay, so let's now focus on what happened in H1. First thing is that H1 has been a pretty good semester actually. We first had a very significant growth in terms of revenue. We already announced that in July. So as you know, our revenue for H1 was EUR 50.3 million, which is, if you compare with H1 2022, last year, it's about more than three times bigger, and very close actually to the overall revenue we generate last year. What is very interesting, what we announce today, is also that we made very, very significant progress in terms of EBITDA, since we, we reached a breakeven at EBITDA level, which means that we are about EUR 2 million better than last year. And of course, this is a major achievement, that's also a major stage and milestone for the company in our way to be a profitable company. And you, you will see that I will explain why we managed and how we managed to be breakeven this semester. And also you will see that our, our commitment is to be actually breakeven for the whole year. In terms of a little bit more details about this revenue, so as you know, this revenue actually got an interesting split, which was the same last year, of one third hardware, two third software. Which really I think is a good indication that today Kalray is a sales solution. And I believe that's very, very important to be able to sell a solution. There's a lot of advantage, and that's really the way the industry, the semiconductor goes. So if that's something of interest, I will definitely ask any question you may have on the split and how you see this split moving forward, but I believe that's very an interesting topic. Something which is also very important as a takeaway is that our margin is 62%. This is basically the margin we were expecting, let's say medium term, because this is the margin of our industry in general, for the player doing well. The good news is that we reached this margin already this semester. Whereas, I mean, the revenue has been growing like hell. I think being able to combine growth in terms of revenue, good margin, and also improving EBITDA, I think, shows that this H1 has been a pretty good H1, with a lot of progress, and also shows that the company is moving into another phase, closer and closer to profitability. In terms of location of the business, very similar. You can see that U.S. is and remains a very significant part of our business, which makes a lot of sense because of the US market is big, because US market is also a little bit advanced, usually compared to Europe. Europe is not far away, and we see a great momentum in these both countries and with both location with market, which is demanding a lot about technologies that we're providing. Okay. Asia is not there. It's very anecdotal. The reason is that we really decide to focus on U.S. and Europe. Also, we know that if we go to Asia, it will be an indirect route. And as you will see in the next slide, we are building our indirect route. We are building the way to work with partners for indirect sales, and we want to make sure that it's working fine in the U.S. and Europe before growing and scaling in other locations. Asia will be more something for next year. In addition to these numbers in H1, I think what was also very important to visibility is about what's going on with large contracts that we sign, because of course, these large contracts are the very high potential for the next years. One of them being what we call the jumbo contract. If you may remember, we signed a very, very large contract with a large American-based company a few months ago. This contract has a couple of phases. A- and, so it's important for us to give you a status where we stand in terms of these phases. We are in the 2023 phase, let's call it, and the very important information here is that we are completely in line with this, earning. With this schedule, the availability of Coolidge™2 was very important, of course. A- and, as you will see later in the presentation, we will provide the first samples of Coolidge™2, the first board, to this large customer in the coming days. Completely on track with the goal to have them fully validated, the new board they are expecting before the end of the year. Okay. Just as a reminder, as you know, the next phase will be to go to pre-series, pre-series pre-production, okay? Whereas the volume production, what we call volume production, is supposed to start by 2025. We also share last time when we had a couple of large contracts, especially in the media space. Media space is very interesting for us. One of the reason is that the media, of course, is processing a lot of data. Required to process this data in a very, very efficient way, usually in multi-location, sometimes actually with implementations, which are merging both what we call hybrid implementation, with some part in the cloud, some part on-premise. And we have very, very relevant solution for them. So, we have a sister company called Pixitmedia, which is using the Kalray portfolio, but focusing on the media industry. And we definitely see the continuous traction here. We signed a couple of contracts, one of them being a multi-million euro contract for this year. It could be even bigger next year. The information I want to share is that, the implementation deployment is moving smoothly, and we are very, very happy, and of course, we believe that by having very successful product deployment, we are entering into repeat business, which will be very interesting for the rest of the year and especially for the coming years. Okay, now let me let Denis comment to you the financial details of H1. Denis? So we will present you the financial statement, which have been reviewed with a limited review by our auditors. And so the first slide is the income statement. So Eric already mentioned the EUR 15.3 million euros revenues, as well as a break-even EBITDA of - EUR 34,000. So the important item here is maybe the growth in the operating expenses. This growth of 34% is the growth of the company as well, and the team, especially the team. But it's this growth is a little bit minor by the fact that the perimeter for last semester was only including one quarter of the subsidiary acquired last year. So in terms of net result, we end with EUR 3 million versus EUR 7.5 million. So you can go next. The balance sheet shows actually EUR 80 million total asset as well as total liability. This is the same amount as the end of last year. The main item here is the intangible assets, which includes the EUR 13.3 million goodwill of the last year acquisition, as well as EUR 19 million R&D capitalized. The most important balance asset, sorry, is the accounts receivable. Accounts receivable, which doubled from EUR 4 million- EUR 8.5 million. This is mostly the increase of the revenue, obviously EUR 15 million versus EUR 4 million, as well as a concentration of the billing in the last quarter of the first semester. In term of debt, we have reduced our total bank loans with the end of the convertible bonds plan issued in February 2023. In addition, the other debt decreased by EUR 6 million, and this is due as well to the exercise of the price complement related to last year acquisition. In term of cash, as a result of what we just saw in term of gross margin, so the 62% mentioned by Eric earlier, we tremendously improved the free cash flow by EUR 10 million, and we end the period with a strong EUR 23.3 million in cash. Good. Thank you, Denis. Okay, so now, let's look about H2, and let's give you the update and especially explanation and clarification about what we announced this morning. Just as a start, so definitely for us, 2023 was, and is, a very important year. And one of the main actually objective we had, and that's something that we shared with you earlier this year in January, is really to—it should be a new major step in building a highly profitable business. Okay. So that's really what we've been focusing on, especially, as you can see, the result that we showed you in H1. So how about H2? H2, we, as you see, we've announced actually that the objective revenue for this year, we believe that we'll have difficulty to reach EUR 40 million, and we reviewed our objective to EUR 30 million. What is the main reason here? The main reason is, despite a very strong commercial, pipe dynamic, what we see is that, the one part of the growth of this year was expected to be via indirect sales. Especially when we say indirect sales, you may remember that we put in place a partnership at the end of last year with Dell. That is the largest, actually, OEM worldwide sector. What we see is that this partnership is moving in the right direction, is generating already a significant part of our revenue. However, it is not as fast as we were expecting, and there are two reasons for that. The first one is that we realized that setting up and having this type of relationship, really operational on the field, takes more time than what we expected. Basically, very concretely, what happened is that we got very strong support from the exec management team at Dell, we're working with. Of course, this support, and the communication and so need to be translated on the field, and that takes time. We need to have new materials, new portals. We need to train the sales force. We need to change the habits and so on, and that takes time. It's going in the right direction, but yes, it takes more time than we were expecting. The second also thing that we realized, and that's, of course, something also that we realized, little by little, based on concrete data that we get, is that the sales cycles going indirectly, so through indirect sales, especially with a partner like Dell, which is a big company, are longer than expected. And when I say when expected, I mean longer than when we sell directly to our customer, and significantly. Okay, so just to give you some data, so typically, direct sales, once the customer is well qualified, we've got a sales cycle between three and six months, which is pretty good. Okay. What we see through indirect sales, and this can be significantly actually extended more from 9-12 months. So it may be better even by a year. But you see that basically close to the double, and that has an impact because we definitely were expecting this revenue to go faster in H2, because the funnel was big, and H1 and we are still very big funnel is growing, but the time to closure it happens. So we definitely see business, but it's much longer than what we expected. Okay? And so this is the reason of this change in the operations and what we believe we will reach at the end of the year. Having said that, so I underline the disadvantage of indirect sales. The point is that indirect sales are—have also a lot of advantage. That's a way to move fast, that's a way to scale up, and also a way to leverage existing structure, existing teams, from our partner, and not to be able to invest in large organization in terms of, sales, in terms of marketing, in terms of operation. So that's a model we need to follow, to scale up. And as we are a very ambitious company, this model need actually to be deployed and actually extended. So then to say that first, we, we receive, the... One first benefit of this model is that, actually, thanks to that, we have, let's say, a point of profitability, which is lower, and that has been proved, since you can see that our EBITDA H1 is already actually at break even. Our commitment here is that, and that the update we also have in terms of objectives, the update is that we commit to reach break even for this year, which we believe is really a strong evidence that the company is moving to the next stage, and getting closer and closer to profitability. If we step back and look at the high-level financials since the last five years, of course, even if we are a little disappointed to have to review our target revenue, what we see is that there is still definitely a strong traction, a strong improvement, and that we're going the right direction. We will not share detail about 2024. You may have some questions, but what I can already assure that definitely we see a strong growth, and we will continue growing very fast. And you can see here that the curve is actually very explicit about the direction we go. Now, before answering any question you may have, I would like to spend maybe five minutes to talk a bit about next. What is next? And talk about Coolidge™2, which, of course, is a very important topic. We announced a couple of days ago, but also talk about the overall markets, the AI market. Why Coolidge™2, especially the AI capability of Coolidge™2, can really leverage the revolution that we experience due to AI adoption? And also, maybe I'm going to give you a little bit more details about what does it mean, what means AI for us, where do we sit versus something that you may hear about this company selling this. I think it's important to, in a way, sorry to say that, educate, and make sure that everybody understand that we're part of this explosion. We will benefit from this adoption of AI massive, and I will explain the next slide, where and why. Starting with Coolidge™2. Coolidge™2, as you may have seen, is what we call the 3.5, Generation 3.5. Why 3.5? Because Generation 3, in the sense that it uses the current node technology that was already for Coolidge™, but we actually improved massively the performance through architectural optimization, so that we've got a product which is actually, to give you an indication, for some use case, 10x more performance than what was Coolidge™1. Okay. Let me give you a couple of additional details. Coolidge™2 first. It's been available since the end of August. So you may remember that we announced the beginning of the manufacturing at end of July, or by July. I don't remember actually the date. And so we work with TSMC, and you can see actually that the timing and the speed of manufacturing actually is very good. We're back to what we've seen pre-COVID period, so less than three months. And so once we receive the first samples, then you need to port, of course, the software. And we are now in a position where we checked most of the features of Coolidge™2. We can already run some algorithm, and we are already happy to share that we reach the performance objective that we had for Coolidge™2, and we could check it, which is, of course, a very important milestone. In terms of just a high-level information, so it's the first DPU with AI capability. If you look at, I will come back on the competition for the processor with AI, but we are the first DPU with AI capabilities and with the strong processing capabilities. More than 50 TOPS in integer, also in terms of FLOPS, which is the way you operate, that's the way you use massively in the industry to run inference for AI. It was 25 teraFLOPS, which is about 10x what we have in Coolidge™1.... Also, a big effort has been done on the power consumption, because power consumption, of course, is very, very important. Just to give you an indication, we managed to reduce power consumption by between 20%-50%, depending on the use cases, which is massive. Okay? Of course, Coolidge™2 is provided with what we call an SDK. An SDK is a set of software environment so that our customer can either develop a product or use the product and the complete stack as such. And we, of course, so this SDK is compatible withCoolidge™1, is actually compatible with standard environment that you have in the industry. Because this has been our strategy for the last 10 years. It will be deployed with the chip to our first customers in the coming days. Talking about first customers, I'll give you an update on what we call the jumbo contracts. As you can imagine, this is, of course, forever top priority to provide Coolidge™2 to our customer, so they can start testing and checking that we reach the performance that they were expecting. So this is ongoing, and I'll tell you when we receive the first samples in the coming days. Now, if we step back and let me try to give you a little more about, okay, what is happening in AI? Why do we see a very strong potential here for Kalray as a company, and where do we put our focus, and where do we see our potential growth? First thing I want to share is that if you look at why AI is booming, I think there are really two triggers, two drivers. The first one is that our modern society generates more and more data, so there is a huge amount of what we call unstructured data available, okay? It can be on the web, it can be through IoT, so a huge amount of data available. The point is that you need to analyze this data in order to make this data meaningful. The traditional algorithm that have been used for years, and for centuries, actually, are the traditional algorithm. The way they've been built is by understanding how this data work and understanding, if you have this data as an input, what you would expect as an output. Okay? The big, big strength of AI approach, it's a new way to build this algorithm. It's a new way to manipulate processed data. It's a way which is extremely interesting because actually, even if you don't really understand this data, even if it's a bunch of data, very unstructured from many sources, actually, you can use this data to develop an algorithm and to have very significant and, and good result with a lot, lot of accuracy. So AI, in a way, is the only way to manage this huge amount of data. That's the only approach to do that. So having this data is necessary, and vice versa. I mean, if you don't have AI to manipulate this actual data, you have a hard time to extract meaning information from the data. The second very important driver, which is, of course, very, very important for us, is that the reason why AI is booming right now is not because the technology is being developed now, because the technology exists for many, many decades, like even 40 years ago. But the point is that now we've got chip semiconductor processor with the performance that you need to make AI relevant, okay? That's why we experience by this combination of a huge amount of data available and this huge performance available also on the chip side. That's why you can see that AI is booming and why you can realize the power of AI right now. You can really experience this power by yourself. In terms of where AI run, let me clarify it a bit. We are listening and hearing a lot about ChatGPT type of AI. So this is what runs in the cloud. So when we say in the cloud, that means it could be public, can be private. So that's one location where AI runs. And I will share with you in the next slide, okay, what does it mean? How does it work? And then there is a second location where AI also is extremely relevant, where you will see a huge potential, especially in the future. It's the edge versus the cloud, which is the edge. Edge, and I come back on that, means at the edge of the network, so not in the cloud, but at the edge. If you look at the way these two locations run AI, it's very, very different. Let me explain how it works. Of course, a very simplified view and where Kalray is. Let's start with AI in the cloud. Once again, when I say cloud, it may be public cloud, okay, Amazon, Google, and so on. Or it can be private cloud, which means data center in large enterprise, typically having their own servers and running their own algorithm and their own AI network. So the way it works, and it's always the same, you've got one part that we call the compute, the compute farm. Compute farm is the part of data center which is in charge of computing. What does it mean? Computing means that you take this input, and then you analyze this input and calculate parameters. You calculate the weight to build your network and to train this network. So the compute part is really about crunching and running mathematical function. Okay. It happens when you have to run very large amount of one type of calculation. One approach is very relevant, and one type of process is very relevant. These are what we call GPU. Okay. You certainly know GPU, now it's getting more and more mainstream, and one company is very, very good GPU, the leader in GPU is NVIDIA. Okay. And so what this guy will provide, they provide very expensive card, very powerful card that you can put into farm, farm of GPUs to run this compute. Now, having said that, a compute, a GPU needs data. It needs data to make the computing, okay? So if you don't have a very, very efficient way to provide data to feed the GPU, it doesn't make any sense to have very expensive GPU, okay? Because they will be in an idle mode for most of the time. So you need very, very fast storage to feed the GPU as fast as possible and make sure that they don't starve. How do you do that? You need actually what we call primary storage, very, very fast storage. When we say fast storage, it means millions of files per second, millions of IOPS which are input, output per second, and also very, very low latency. And not only low average latency, but also low tail latency, which is the, you know, the maximum latency that you would expect when you ask for information. And this is where Kalray sits, okay? You've seen us talking about fast storage, and just I want to make sure that having this slide in front of you, understand that first, yes, Kalray is part of these needs, where you need to have a very fast and efficient infrastructure to—especially to run the training part of AI. So we are part of that. And we start actually generating revenue, and we see a lot of traction here by selling our solution with storage as part of Gen AI opportunities. But we are not definitely not trying in a way to play in the ground of the GPU, which has been especially ground in a company like NVIDIA, which has been around and investing massively for the last 10 years in this domain, okay? So we focused on a part which is also very, very important, also growing very fast, and which is much less exposed to big player. And here the competition is definitely not easier, as we mentioned many, many times. And here, for example, there are a couple of products from different competitor. I will come back on that in the next slides. But definitely less exposed. Now, as I mentioned to you, we've got the cloud, where you really run AI, but there's the edge. So edge is a bit more segmented. You have what we call the far edge, which are typically the sensors, your phone, camera, and then something we call the near edge. And near edge are mainly a couple of servers with some cards. So just to make it clear, on our side, again, our business to sell very powerful cards, and we serve this near edge market. So what do we mean by near edge market? What kind of application? So it's about industry, it's about logistics, healthcare, you know, it's about also media, telecom. So these are the main verticals having a strong demand in terms of edge processing capabilities, and that we'll use massively AI in the future to basically run all the processing we need to be efficient. If you look at the way it works, as you can see in the next slide, it's very, very different. You don't have a huge farm, with, you know, a very, very well-identified function in a large data center. Here, what you have, you have servers, a small number of servers, one, two, maybe 10 maximum, located in many locations. Here the implementation is very different. What you need is to use data server and then to have a technology which allow to do a lot of different type of processing, okay? Not specialize in one processing, but being able to, to have what we call intelligent computing. This is also where our technology shine, because our chip is strong in AI, but it's also strong in computing, it's strong in communication. This is actually the second market that we focus. Once again, nothing new. Mentioned several times that we're focusing on storage, we're focusing on edge computing, especially for smart vision and 5G. Just to repeat and to re-explain where we sit and how we differentiate versus other companies in this AI market. AI market, for sure, I think there is no ambiguity here, will grow like in the next decade. The reason for that is that AI needs chip. You have no way to be efficient in AI if you don't have chip. The more powerful your chip farm, the more powerful power you have, the better your network will run and the more value you will bring versus competition, okay? This is very, very, very important. Anyway, the industry needs chip, okay? This is a market which will grow very fast. You can see a lot of different report in the industry. A 30% growth per year for the next decade sounds very reasonable. The second topic, which is important here, is that what we say is that it's a market which is what I call reasonably segmented, okay? As you can see, it's not one type of product that will actually win the whole business, and that's what I wanted to show with you and to explain to you. At the same time, it's that you don't have one product per customer, okay? It's reasonably segmented, which is good for a company like us, because it's a way to differentiate, it's a way to have market, but also any way to remain in large category with large niche of a couple of billion. And then, of course, as you know, high barrier of entry, which is tough to develop semiconductor. It's really haute couture, it requires a certain amount of iteration. Usually, if you don't have free iteration, you don't have a real product which is mature enough to win production with good yield. And it's also a huge investment. It's hundreds of millions of dollars. So that means that, it's not like you can have a new competitor, in the next day. It's long, and you can see what's happening. And as part of this market, so DPU, as mentioned, will be along GPU. These are two different approaches with different usage stages. Once again, GPU, very good for one type of calculation, very high performance, expensive, but very good, especially in the cloud. For packet manipulation, GPU, they cannot do it. When you need to run intelligent computing, they cannot do it. So it's very, very complementary, and the market of DPU will be, I mean, as growing like very similar to GPU, and with a growth a character, I mean, of growth of also very similar, about 30%, for the next decade. We resume now about Kalray, about the DPU, for some of you, you should be described, but I think it's important to repeat where we stand and how we differentiate. So as you can see, I mean, the number of DPU provider is not so huge. Basically, you've got 6-7 companies worldwide. And actually only four are really leading this domain. You can see, for example, that Intel is trying to push a solution, which is an IPU for a specific usage in the infrastructure. Not so much successful, because that's not easy. But you can see that the four leaders right now in the marketplace are NVIDIA and [Inaudible] NVIDIA, okay? Of course, these are very nice companies growing and having an investment capability. Usually, they don't get capabilities. So the key question is: okay, how do we differentiate? How can we be successful? You know that our architecture is very, very different, with large benefits, benefits in terms of pure power, power consumption, but also because for some use cases, we're much better. And so that if we reposition of these use cases, we've got a great advantage compared to the others. The second important point is, and I think you understood that through the previous slide, is we focus, focus on two segments: the storage and edge computing, where the competition less fierce. We see that there is really an opportunity to become a leader on these segments, and which are less exposed from the competition. We're focusing on other segments. Then also a third aspect, which is also very important actually, is a differentiated go-to-market. So we've got basically two tracks, where again, that's something that we have realized if you follow us for many months. So we've got two tracks. The first one is by building a full system, easy to use, easy to adopt for our customer, and that's the reason why we bid, we acquire this company in U.K., and we build a complete solution with software, hardware, and we are able to provide a complete solution today for fast storage. And because that's the way to really penetrate the market easily. And then the second approach is really to focus on large customers and to really leverage the fact that being a smaller company than the large other companies which compete with, we have the capability to have a closer relationship, a more flexible relationship with these big players, and in a way, to work in a co-development mode, okay? And that's what we've done and what we are doing with the customer we announced. We've got a couple of very large project. Same thing, it's about working closely with customers to make sure that this customer, they have a unique in a way, not only access to unique technology, but also a unique partnership with a company like us to basically have the best technology for their own product. To help this, I mean, we covered the... what happened in H1, the detail H2, and also the outlook and how we see moving forward. I hope it gives you a good indication of, of where we are. Now more welcome with Denis to answer any question you may have. Ladies and gentlemen, if you would like to ask question, please click the Raise Your Hand button on your computer or press star six, six, sorry, your phone keypad, and wait for us to speak to you. Mr. Ben-Amran, you can put on your microphone. The floor is yours. Hello? Hello, hello. Yes, we can hear you. Do you hear me? Yes. Okay, very good. I will ask my question in French, because I see that most of the participants in the conference are French. No problem. I will, I will translate you. I will translate I will translate your question, don't worry. If you want to answer in English, you answer in English. My question is [Foriegn language] quite simple. It's just to come back to the sales warning that was announced this morning, which is still quite important, since in fact there is a 25% gap compared to the initial guidance. I'm going to try to understand. You published your turnover in July, this scale, which was rather good, and you confirmed the guidance. [Foriegn language] Why, in fact, didn't you perhaps try to adjust this guidance in July, since it's a little less than two months gap? And if Intel was quite slow in orders, maybe you already suspected it since July? Yeah. Okay. So I will repeat your question in English, and answer in English, okay? So thank you. Thanks for the question. So the question, let me reword the question now. Of course, do not hesitate to let me know if I did not reword properly. But I think your question is about, can we announce this morning that we updated our objective from EUR 40 million- EUR 30 million, which is a significant decrease of about 25%. And so your question is, okay, why do we announce that today, and we did not announce that in July when we had our last press release and conference? So this is definitely a very fair question. Let me step back and explain a little bit more how we can today provide guidance and how we can forecast what's moving on. I think I just want to first clarify, and I completely understand that's not other views, is that we are a company which is still in a different mode or different phase than most of the companies which are listed today, especially on Euronext. The reason is that, as you can see, we are in this phase of scaling of our business since about 18 months. We, and the way we grow is not about 10% per year, right? As you can see, I mean, anyway, 2023, it's about doubling our business, right? We don't have a lot of metrics. The metrics we had, the main metrics we had, were about the business we did. The business we did so far was mainly about the direct sales. Of course, you've got much more control when you have a direct sale, because you have direct contact with customers, and you've got data. It's back to data. You can average, you know the sales cycle, and so on. There are seven parameters that we know, and that actually we learn from an indirect sale is the conversion rate. Conversion rate is amount of business you close versus the amount of business you put in the pipe. Okay, so all of this data you certainly know very well, to calibrate, basically, and to extrapolate what you have in your commercial pipe and be able to provide a number to say, "Okay, in six months, I should trust that in nine months and 12 months." Okay. So my point is, first, because we don't have a lot of history, it's very difficult to have precise data here. First point. Second point is, the fact is that the main difference between what we announced earlier this year and what we announced today really comes from the indirect sales. Indirect sales is even more tricky. More tricky for us, because why? Because it's the first time we put that in place. As you know, it's very recent. What happened is that we've seen the funnel growing very, very fast. The funnel is growing very fast, and it's very aligned with, let's say, the feedback we had when we discussed with Dell. They told us, "Guys, I mean, you will see a lot of opportunities, a large funnel," and so on. Okay. But however, we did not have a lot of history about how this funnel is converted, both in terms of conversion rate, because, of course, the conversion rate also depends on the type of customer that our partner deal with, and also the conversion rate of our partner, and in terms of length of sales cycle. It happens that the conversion rate today and the sales cycle from the indirect route is very different. This is what happened. So we realized that basically, whereas, whereas our sales cycle for direct route is between 3-6 months, so now what we see—and once again, it's always difficult when you have a few deals. Okay, what we see is in average for the indirect route is more between nine and 12 months. Okay. I think, actually, I mean, what is even more surprising, more surprising for us, I think, you know, deals which were very, very advanced in the sales cycle, you know, in the purchasing order, you know, when, when you... Once again, I don't know your background, but if you know how a sales part works, when you are in a 70%-90% stage, which means, okay, you've been selected and so on, we've been very surprised to see that even if we've been selected, the customer said, "Okay, we go, we're good. We've got feedback from Dell. Yes, it's done." We spent three, four, or five months before the deal is closed. The reason for that is also because we are part of a large deal from our partner. We are lowering influence, actually, in the... because the deal actually covers much more things than what we provide. That's the beauty of working with a company like Dell. We know the advantage, but that means that having all the data and having this basically this history is not trivial. So we may have been able to say, "Okay, based on the first indication, we may not be EUR 40 million." It's always difficult, you know, when you just start having some signs, you need to convert, but we've seen this summer has confirmed what we what we see today, okay? And that's why we need that to review our objective. Once again, I want also to emphasize that review objective, the review objective is very important one, definitely, but we also make the most to make sure that to take into consideration that and keep one of our objective, which was also very, very important for us, which is the EBITDA, and to, to make sure that even if there is this, reduction of, of revenue, we also reach a very important milestone, which is to be, to be break-even this year. And which is better actually than what we, committed a couple of months ago. Okay, thanks. Mr. Le Bihan, you can put on your microphone. The floor is yours. Yes. Thank you, Eric. Thank you, Denis, for the presentation. I have a question about the EBITDA break-even for 2023. I was wondering if you take into account the inflation context, I mean, by the inflation in the global economic context, and how you would manage the impact of the inflation on the cost part. Okay. So, maybe I can give a first answer, and then, and then- Okay Denis, go into the details. So of course, I mean, the—what you see in H1, and what we forecasted for H2 includes the inflation impact. As you can imagine, the inflation for us is mainly on one side, the payroll. Okay. For information, we have an annual round of salary increase every year. It was in July, and of course, we had to increase significantly the payroll to be relevant and competitive. Okay? That, of course, includes the inflation impact, and that's in our numbers. Then it's about also our suppliers in terms of tooling, in terms of... Of course, and I will let Denis give the details, we review all of the price. Some of them were negotiated actually for a few years, so did not change. Some of them has been reviewed and are taken into consideration. Denis, you want to give more details? Well, yes, one of our biggest item is the software tools, the electronic design automation software tools, and this one is three years contract already taken into account and accounted for the period, obviously. The rest of the expense has been planned and forecasted for the second semester, according to the current last negotiated price.... So, in terms of foundry costs, those costs are set already and are in line with our forecasting as well. Thank you. Ladies and gentlemen, if you would like to ask a question, please click the Raise Your Hand button on your computer or press star six on your phone keypad and wait for us to speak to you. Just a comment, if I may. I mean, don't hesitate to ask the question in French, okay? I will translate the question so that we can keep the international audience. Okay, we've got a question on the chat. Yes, please, if you can type your question. Okay. My understanding is that the question is about what will be the impact of this delay in 2024, and do you expect to catch up? Back to also a question, which is a more global question about what is your outlook for 2024, right? Because I think that's definitely the question. What we see is, as you understood, we start having a better understanding of key parameters of the cycle when we go in direct. Of course, now we've got more data to basically be able to forecast for the next 12- 18 months. Anyway, what we see, the funnel is growing very, very importantly. Very large funnel, so we definitely expect a very significant growth. We will focus in the coming months on two things around indirect sales, okay? Because indirect sales is the way to really grow the business and to scale up very rapidly. The first one will be, of course, about continuing our very strong and good relationship with Dell Technologies, to make sure that we provide all the information they need, all the training, and we make sure that we work with them on day to day, excuse me, to grow this partnership. This partnership is approved to generate several tens of million of revenue in the coming years. There is no ambiguity about that. The potential is here, and the motivation of both parties is here. At the same time, what we see also, especially on a certain number of verticals, it makes sense also to leverage partnership of distributor on specific vertical, and once again, to leverage their sales force. There are a number of action in place that we may announce short term about putting in place other partnership to deploy our solution massively. The goal is that each of these partnership generate several tens of million of euros in the coming years. What we do with Dell Technologies right now is very important. Once again, is setting up our model, and this model is definitely, for us, a guarantee to be able to grow very fast in the coming years and of course, in 2024.