So if the macro situation is healthier, if a macro is recovered, which we expect on the computer, consumer and the communication segment, the demand will redirect to us, improve the loading situation. So there’s so many different variables right now and giving the insufficient visibility, I can’t give you anything specific, but we definitely feel optimistic about it because the inventory situation is improved in many of the market segment already. We just have to continue monitoring the progress and we will keep you updated quarter by quarter.
Unidentified Analyst: Great, thank you. So my last question is in terms of 28 nanometer or 22, because on the demand side, actually we found some of the clients are migrating from 22, 28 to maybe FinFET process nodes, including SSD controller or Wi-Fi 7. And on the supply side, we found that your competitor in China is planning to international mass reductions for 28 high voltage process node. So demand is moving into the FinFET process node. And on the supply side, we know that UMC had kept good positions in 28 nanometers. But on the supply side we found that there’s new players are planning to enter the high voltage process market. How do you look at or how do you plan to keep your market shares or keep your technology leadership in high voltage, or in those kind of niche markets or your products, or to maintain your market shares?
Jason Wang: I think fundamentally, the answer to that is we need to stay competitive. We, in the past and even currently, right now at this moment, we work closely with customer, strategic customers to build specialty technologies, differentiation and lasting value for them to enhance their, our product portfolio, as well as our customers product portfolio and with a diversified capacity. And so that we want to do that to offer our customer with a competitive solution. And so I think that’s the answer to that. The competition is everywhere, and the key answer to that is we have to stay competitive, and we definitely strive to do that.
Unidentified Analyst: Okay. Thank you. I have no more questions. Back to queue.
Operator: Thank you. Next one. Laura Chen, Citi. Go ahead, please.
Laura Chen: Hi. Good afternoon. Thank you, gentlemen, for taking my question. I think my question is also kind of a follow up on the CapEx outlook and also the depreciation. Since Jason, you mentioned that the CapEx will pick out in this year. I’m just wondering that looking forward, how should we look at the depreciation cost trend into next year, how would that impact our gross margin?
Jason Wang: The depreciation impact will be kind of delayed factor. So if the CapEx is peaked in 2024, I think the depreciation expense probably will be peaked two or three years later. So the trend will still be on a minor upward trend all the way to 2026.
Laura Chen: Yes. So the magnitude of the depreciation cost increase for the following few years, you expect that will be more like a steady increase where there will be gradually scale up. Because still in the past two years we see the depreciation cost, CapEx increase.
Jason Wang: Given our ROI driven base CapEx, and we are managing our overall depreciation expenses as a percentage of overall revenue. So I do expect the increased magnitude should be under control and also acceptable by the investor community.
Laura Chen: Okay, thank you. And also for the AI opportunities. I understand that Jason, you mentioned that will be 10%, 20% addressable market. And where are we now and do we expect there will be a secular growth for the following two, three years?
Jason Wang: Well, I mean Laura, like I said, we still in the early stage of the, of the edge AI. Most of the focus today is running the computation chips for the AI modules today. So we have a very small exposure there. But we do expect the edge AI will start coming in. And there are some projects that we have in discussion with customers. They all have a very high expectation the market will taking out soon. But I don’t have any specific to say. We do project those application or those features will be used in the edge AI in the next few years and that will probably representing at 10% to 20%. It could be even higher. But at this point we project about 10% to 20% of the AI.
Laura Chen: Okay, thank you. Very clear.
Jason Wang: Thank you, Laura.
Operator: Ladies and gentlemen, we are taking the last one, the last question. Timm Schulze-Melander, Redburn Atlantic. Go ahead, please.
Timm Schulze-Melander: Hi, thank you very much for taking my questions. I just had two, if I could sneak them in, please. The first one was on the advanced packaging and chip integration. You talked about hybrid bonding moving into grace of volume in 2024. Just wondering if you could share a few more details about the end application and if that’s the partnership that you had press released with ASE, Faraday and Cadence. Is that the application? And then I had a quick follow up, please.
Jason Wang: Well, first for the wafer-to-wafer hyper bounding solution today, the applications for the RF front end module is not completely correlated with the ecosystem press release. But to address your part of the question about the ecosystem conversation for the ecosystem, we are not talking about one specific solution, we are talking about the UMC overall 3D IC solution is leveraging the entire ecosystems to support that. We do not do turnkey solution means the end-to-end solution. We’re only providing wafer-to-wafer hybrid bonding or the interposer end solution, while we have all the ecosystem partners to support the customer to provide the total solution. So that’s not only limited with the high hybrid bonding.
Timm Schulze-Melander: Okay, but for that hybrid bonding, are you actually doing that wafer-to-wafer bond yourself in a UMC facility or is that being outsourced or shared with, with some partners?
Jason Wang: It’s in house, yes.
