Thank you, Jason. Welcome everyone to Parade Technologies 2025 Q4 webcast investor conference.

Parade Technologies fourth quarter 2025 consolidated revenue was 125.22 million U.S. dollars, and the net income was 17.53 million U.S.

dollars. Both basic and fully diluted after-tax earnings per share were 22 cents U.S.

dollars and 22 cents U.S. dollars respectively.

This results compared to 127.41 million U.S. dollars and a net income of 21.30 million U.S.

dollars, or 27 cents U.S. dollars and 27 cents U.S.

dollars per basic and fully diluted share in the year-ago quarter. In U.S.

dollars, the fourth quarter revenue decreased 14.32% sequentially and was down 1.72% year over year. The gross profit in the fourth quarter of 2025 was 22.63 million U.S.

dollars, a decrease of 15.27% from the previous quarter and a decrease of 2.11% compared to the same quarter of last year. For the fiscal year 2025, the consolidated revenue was 531.09 million U.S.

dollars, representing an increase of 5% from 506.81 million U.S. dollars in the prior year.

Gross profit was 226.01 million U.S. dollars and the operating income was 87.2 million U.S.

dollars. Net income was 87.75 million U.S.

dollars, or 1.11 U.S. dollars per basic share and 1.10 U.S.

dollars per fully diluted share, compared with net income of 80.68 million U.S. dollars, or 1.01 U.S.

dollars per basic share and 1.01 U.S. dollars per fully diluted share in the prior year.

On November 11, 2025, Parade introduced the TC1316B, the world's first fully integrated AECQ100 qualified embedded DisplayPort touch with TCON embedded driver. This new device targets next-generation automotive cockpit LCD displays requiring higher resolution and higher refresh rates.

The TC1316B combines Parade’s broad portfolio of patterned touch screen technology and the market-proven EDP TCON, in-cell and high-speed signal technology to provide excellent picture quality and responsive touch performance. The TC1316B leverages Parade’s true-touch pattern technology and design expertise, built on our experience of shipping over one billion touch devices and the proven in-cell performance from the consumer notebook market.

The device delivers superior EMI suppression, low power consumption and advanced signal processing algorithms to ensure reliable and responsive touch for the automotive cabin. Based on the current business outlook, Parade is providing the following guidance for the first quarter of 2026: revenue is between 119.75 to 132.25 million U.S.

dollars, gross margin is between 40% to 44%, operating expense is between 33 to 36 million U.S. dollars.

大家好，欢迎大家来参加普瑞公司2025年第四季度线上发布生会议。首先由我跟各位报告普瑞公司2025年第四季度财务数字。本公司2025年第四季度合并营收为美金1亿2502万元，税后净利润为美金1753万元，除权后基本与稀释每股净利润为美金0.22元，约当新台币6.97元，及美金0.22元，约当新台币6.94元。而去年同期合并营收为美金1亿2741万元，税后净利润为美金2130万元，除权后基本与稀释每股税后净利润为美金0.27元，约当新台币8.69元，及美金0.27元，约当新台币8.66元。2025年第四季度合并营收较前一季减少14.32%，较去年同期则减少1.72%。在营业毛利方面，本公司第四季度营业毛利为美金5263万元，较上季减少15.27%，较去年同期则减少2.1%。令本公司2025年自结全年度合并营收为美金5亿3109万元，较去年美金5亿0581万元增加5%。营业毛利为美金2亿2601万元，营业利益则为美金8702万元，税后净利润为美金8775万元，除权后基本稀释每股税后净利润为美金1.11元，约当新台币34.51元，及美金1.10元，约当新台币34.3元。相较于去年，税后净利润为美金8068万元，除权后基本与稀释每股税后净利润为美金1.01元，约当新台币32.55元，及美金1.01元，约当新台币32.39元。7月2025年1月11日，普瑞公司发表TC1316B，该产品为全球首款通过AECQ100认证EDP界面的触控与实时控制器整合驱动器全整合式晶片TTED。这款全新晶片专为 以美金1.01亿元约当新台币32.39亿元。7元2025年1月11日，普瑞公司发表 GC1316B，该产品为全球首款通过 AECQ100认证 EDP 接面的触控与持续控制器整合驱动器全整合式芯片 TTED。这款全新芯片专为需高解析度与高刷新率的次世代汽车座舱 LCD 显示器设计，整合普瑞广泛的触控屏幕专利技术以及经市场验证的 EDP持续控制器，内嵌式显示屏与高速讯号技术可提供卓越的画质与灵敏的触控效能。GC1316B应用普瑞的 TrueTouch专利技术与设计专长，这些技术与经验源自超过十亿颗触控装置的出货实际以及在消费性笔记型电脑中经过验证的硬件效能。该芯片具备优异的电磁干扰抑制能力、低功耗特性与先进的讯号处理算法可确保汽车座舱内触控操作的稳定性与敏感度。基于目前对2026年第一季营运的展望，我们预期合并营收为美金119.75到132.25百万元，合并毛利率为40%到44%之间，合并营业费用为美金33到36百万元。 This is my presentation for the 2025 Q4 and the Year 2025 financial results.

Now I transfer to CEO, Dr. Jack Zhao to answer your questions.

Jie, may begin.

Thank you, Jack and Yoming for taking my question. For my first question is kind of housekeeping, could you give us the updated breakdown of the product in fourth quarter in the four segments?

Thank you.

Thank you. In the Q4, the TS product line was below 50%.

Our DP product line was around 40%. Our TC, which is source driver, and the TT with touch, include our TTED device, both were above 5%.

Thank you, Jack. That is very clear.

My second question is regarding to the market outlook on the PC, especially the notebook. Could you give us some update or highlight regarding to the visibility, especially into second half this year?

Also how should we think about the gross margin outlook for Parade for this year? Thank you.

Yes, I think there are many information around in our industry. Really, we heard the DRAM and flash memory are in the shortage side.

I think the question came from is whether both one will impact the notebook shipment. During the CES show, we had an opportunity, or myself had an opportunity to talk to the major OEMs.

I think the tier one OEMs, in terms of their forecast, seems to remain similar or not much change. They do recognize that the DRAM or NAND device might price increase, but they don't see they will have a shortage.

That's based on what we discussed with the major OEMs, and that's what we got information. Recently we did see the major OEMs would like to sourcing the device early to have inventory.

Once they have a DRAM or memory available to them, they just build the system. That's pretty much what we heard or observed from the current market situation.

We certainly hope as we move to the second half, the situation may change or may improve. After all, the notebook or the PC are essential for productivity improvement.

That's what we currently see as we discuss with our customers.

Got it. Thank you, Jack.

I'll go back to the queue.

Thanks for taking my question. First question is still regarding the outlook for this year.

I think Jack has touched a little bit about notebook outlook. Can you give us more detail about the gross potential for both DP and also PS product line in 2026?

That's the first question. The second question, I want to check with you, how are you thinking about gross margin?

Because I think you have revised down the gross margin bracket by about two percentage points. Can you give us some more details about this?

Is this going to be the gross margin range for the next couple quarters, or what are the reasons behind it? Thank you.

Thank you for the first question. Overall, the market, and we still remain think the market this year, total number of notebook shipment should remain similar as 2025.

That's what we base on we think. For our device, we have continued growth, our high-speed PS product line, and we see the momentum from this year, from the start from Q1 to each quarter, our high-speed device will increase.

That's driven by the more adoption of high-speed USB 4, the timer, our USB 4 hub, and as well as other high-speed device. In addition, we had last year introduced our TT device for panel.

For notebook, adoption rate or the momentum is pretty high, going to the Q1 and each quarter for volume shipment. Those will become another driving force for our revenue growth each quarter in 2026.

Even though the overall notebook number, or PC number, may remain similar as 2025, we think we will continue growth really based on our high-speed device, and as well as our TTD and the TD adoption rate increase in 2026 and pretty significant. In terms of gross margin, on your second question, we are taking into consideration industry-wide the OPEX and assembly cost and testing cost increase as we observed potentially.

We take into consideration of this. I think we will remain or probably improve quarter to quarter to see whether the gross margin can continue improve.

We think that we will not have any more significant changes at this moment. That basically reflects taking into consideration industry-wide the OPEX and assembly and testing cost increase.

Thank you. Then my follow-up question is on the ASIC-like business.

Can you give us some more update about the progress of ASIC-like business? When should we expect to see some revenue contribution?

Then your confidence level on these customers will roll out these ASIC-like products. Thank you.

Yeah,

as we shared with previously, our ASIC-like strategy has been progressing well, and our team are busy on the execution. Since most ASIC-like chips are pretty big, and we have both on the display side and on the high-speed side.

Those are using FinFET and using the Deepsemicron, using our high-speed IP. We expect by the late of 2027, or maybe 2027, I think, depend on the schedule of customer deliver onto the market.

We expect we will start to see the revenue contributions. We are very busy on the execution of those projects, and we have a larger team focused on those projects and work with customers very closely and engineering hand-to-hand and very closely to execute those projects.

All right, you're saying that 2027 we might see some revenue contribution?

We should see. We have multiple projects.

We should surely see the revenue contribution.

All right, okay, thank you.

Hi, Jack, can you hear me?

My first question is regarding to the short-term revenue guidance and demand, and the midpoint of the first quarter revenue guidance suggests revenue will remain roughly flat quarter over quarter. Could you share with us which product lines are seeing stronger demand in the first quarter and which are trending towards a sequential decline?

The strong point is the Q1 coming quarter is our high-speed device. As our demand, our TTD device, those are pretty large contribution, and major our customer, are doing historically in the Q1 doing some inventory adjustment, and we have been known for a long, long time.

We think that overall they were back to what they forecasted given to us. That's expected, it was not a surprise at all.

Got it. How does the product mix for the first quarter for the segment will look like compared to that of the fourth quarter?

Frankly, I don't have that number yet. As I described, we see the peers might have a couple percentage higher on the intensity of our mix, and surely our TTT, which carry our TTD device, will have a few points above the current quarter, the last quarter.

Got it, very clear. My second question is regarding to the Standard Plus customer and the ASIC-like business.

Regarding our Standard Plus client, we have strong cooperation in panel ICs and some high-speed interface chips. Could you share details on the further collaborations regarding ASIC-like high-speed interface chip with this client?

Are there specific trends that could lead to further revenue gains from the Standard Plus client? Thank you.

We have been working with our Standard Plus customer in every front, right? On the display, on the high-speed, and every front there.

We continue work very closely on the display side. We have a few projects on the display side in the execution now, and we continue foresee we may have more there.

On the high-speed side, we are not necessarily only work with our Standard Plus customer, even though previously we already worked on something, and the project has been in production already. There are new customers, and the chip on the high-speed side are relatively big, and also multiple chips we're working on.

That's as much as I can share with you. We're using our high-speed IPs, which is most important, or high-speed proven IPs, and will be executed our Deepsemicron, the FinFET technology.

That's as much as I can share with you.

Thank you, very clear. My last question is regarding automotive revenue and the market outlook.

Could management share the approximate automotive revenue for last year? With the overall automotive market remaining quite modest this year, what is your outlook for the automotive revenue?

Additionally, are there any developments in automotive products that you can share, such as winning new customers or launching new products?

Yeah. Our automotive, especially high-speed automotive, both high-speed converter and our DP 2.0 hub, we won in the last quarter and the 2024, quite a few major customers, whether the very high-brand, the Europe large-volume customers, so as towards the China EV side.

Our converter device, and especially the DP hub, doing quite well in the field. I think I understand our DP hub and the automotive-grade DP hub device can support multiple panels at the same time, and you can swap.

Those high-speed devices, you come in 10 gig, and you can spread many lower speed to go out. That's pretty unique.

We have multiple wins in the last quarters. We continue to see those high-speed device start to MP more in the quarter to quarters.

We also just announced our first ever EVP, the TTV automotive device, we call it TC1316V, in low for panel to do the qualification for automotive application. We put together a total solution for the display, so as those USB box, what are they called.

We are pretty optimistic for the automotive revenue generation. I think this year we might have a few percent in our revenue, or more, and that depends on the customer, how they introduce their car into the market, right?

We have the U.S., the number one EV car, we have the number one China EV car, we have the number one luxury, the high-volume, the Europe car on the DP hub. That part many there.

The automotive, the only thing automotive is the design cycle long, and it takes time to the market, but a few of them, it's already in the shipment.

Thank you. I'll be back to the queue.

Thank you.

Hi, Chairman. I have two questions.

The first one is for this year display product, because we're expecting that you will provide growth for next year. This year you face the consumer products shipment challenge, DRAM issue.

I'm wondering what kind of product trend do you see that you may benefit from this trend rather than suffer? Because for the TTD panel, is that offering cost-competitive solutions that customers, may want to adopt more?

That's why you are confident that this year can still grow? Does that make sense?

Yeah. The TTED device for the panel are unique.

If you integrate a source driver, T-Con, and touch together. We are doing so, we are not only, you have a single-touch style.

Customer likes a lot. The previous, the two we are shipping, all are on the relative high-end so-called LPTS panel, right?

We expect the so-called low-temperature growth. Because USB adapter rate increase will continue going up, right?

More important, not only the high-end LPTS panel, we will go to the oxide amorphous panels as well. Those TT device not only support the lower border, the narrow border, but also support thinner panel.

People talk about a lot of narrow border, right?

And

the single-support panel becomes another factor. You probably can only achieve this using the TTD to achieve it.

You look at those panels, you can achieve 1. something millimeter panel thickness.

For industrial design, those are very attractive. That's why we see not only one OEM, multiple OEMs are jump onto those areas.

Is that easy for your peers to duplicate similar solution? Is that easy for competitors to duplicate this kind of similar solution?

That's the beauty of TTD, because you have a touch there. There is a firmware there.

Somewhere you have a significant difference. The behavior will be very different.

I think it's a lot more harder to penetrate compared with TCom to other type of thing, or plus On-Cell panel, right? Literally

you will see onsale

probably will be cut a lot. I don't see why people need On-Cell anymore.

The cost, everything, it's much lower on the in-cell.

Yeah. And

Because, as current, we are in the LPTS, the panel itself, we have a solution breakthrough to going to the oxide, going to... Those will be cross-border change.

I see. My second question is about SE like.

I think last quarter I asked the same question, and I think I got the sense that the projects that you are working on are multiple projects, multiple customers, and if some of them come through, then it will provide probably the strongest growth ever. Compared with the historical.

Is that still the case for you, or after three months, are you more confident that your SE like will grow meaningfully from next year?

The previous one we talked, we were in the phase, we were talking or planning for projects. Now we are in the phase with customer forecast.

We're in the execution, with a clear schedule, with clear, a lot more details on the projects. I think if we are, we hope both parties all put together want to make it successful.

The revenue will be significant. Excellent.

I think the share price now reflects this kind of positive, do you consider to share buyback to show that your confidence?

Actually, our

board

today, our board meeting, I think it might be announced. I don't know whether announced or not announced yet.

We will do the share buyback as we will do, and we will shares to return the cash back to our investors.

Excellent. That's great.

For the last one, for ASIC-like, I think you start with consumer-related devices. Is that possible that you will expand to server, or even AI-related, probably in the long term?

We're always chasing that, right? That's part of our strategy to chasing this to AI-related.

We are doing the share with the investor community. We are doing those PCIe for storage, using our retimer and redriver.

We have a few of the design win case, and those are the data center type of storage. We have those design winning out there.

We're also working on those cable solutions, whether PCIe cable or Ethernet cable. Those are our silicon Germany device.

We're doing, especially storage, we work with a few of customers on those storage solutions. So

you already penetrate into server storage?

Server storage, and utilize...

Like Dell EMC?

Dell EMC and those kind of thing, yeah.

I see. How many percentage that, probably very low now, but I think it will provide high growth.

We have not really carved those in because the design cycle is a little bit long, and also those customers have not really shared with us what is the potential volume and those kind of things. It's unlike the ASIC-like device, because you negotiate very heavily on both parties, and those one is just design wins.

Excellent. Is that strong now, this year, for the storage thing?

We already confirmed design wins. I don't know when they start to volume production, want shipment.

I see. Great.

Thank you, Jack.

Thank you.

I would love to speak. Thank you.

Hi, Jack. Can you hear me?

Thanks for taking my question. My first question still regarding the ASIC-like.

Jack, you just talked about right now we are moving into the execution phase. Can I ask your view on your market share, especially on the high-speed interface-related projects?

Because you talked about there are multiple projects, but I'm just wondering, based on your view, how big are the TAM, and how would you expect your share in the next three years? That would be my first question.

Thank you.

Frankly, I can only, based on the customer in the negotiation, give us a forecast, right? Because those projects we invest heavily.

I think very heavily. The customer or our partner give us the forecast, and we did our investigation as well.

We think those forecasts represent the major shares, and those forecasts are correct. That's why those are part of in the agreement.

That's as much as I really don't know more than, like, you probably know more than what the shares, what it is. I only can talk based on what we knew, what we signed, what the customer gave us a forecast.

I hope that answers your question. I hope it answers your question.

We also did our investigation as the best that we do.

Got it. Sir, I think it's already very clear.

My second question, still regarding the gross margins. Jack, can I ask you, is there a KGD in TCON, timing controller?

Because you talked about OSD testing. I'm just wondering, for example, for med TCON, is there a KGD, or is there any of your products involving the KGD you need to buy the memory from zero?

No, we don't buy DMEM. That's good luck.

Those DMEM, the purchase is long gone, okay? We don't have those one anymore, okay?

Otherwise it would be horrible, right? No.

It's that EOL many years ago. You're correct.

A few years ago, we did have, but now we don't. We don't buy the DRAM for those high-end TCON.

The memory is integrated together. What you said on the margin, it's not particular on the TCON.

It's mainly on the OSD, the price adjustment. You probably heard quite a lot, we take into consideration this.

It's not a particular product or particular something. We hope we would overdone, but I think for us it is prudent to take into

consideration. Thank you, Jack.

Just one follow-up from me. Jack, can I consult your opinion?

Five years ago, when there was COVID, we also had foundry cost hikes, material cost hikes, logistics cost hikes, but a lot of the cost hikes at that time can be passed through to the customers, or even with some markup. This time, what's the difference?

What would be your view? Is there any chance we can still negotiate the pricing with our customers?

Because small sample parts, as you mentioned, are going up.

This time is pretty sudden, I would think, okay? We will negotiate with some of our customers on those one, and hopefully all rational to take into consideration.

We understand the situation and the market, because our OEM also faced many pressure there. Some of our parts, particular, we already reflect to our customer those increase there.

That we probably have to take into consideration, how do we negotiate with the customer under those price increase. That's the unfortunate this time.

Even we gone through last time, this time is a lot more quicker than was.

I see. Thanks, Jack.

Very clear. I'll be back in a queue.

Thank you.

Jack，小明，你们好，谢谢接受我的提问。那我的第一个问题是，像现在 AIPC，它对于屏幕的刷新率或者是对省电的要求，Z 的增加，那我们作为 ETP 的领导者，那我想请问 Jack，现在我们看市场上面呢，比方说像 ETP 1.5 Tcon，目前升级转换的进度如何？那或者是说，在未来这个 ETP 的升级，那对我们在每一台 PC 上面的 content value，大概可以比现在会增加多少？这是我第一个问题。 ETP 1.5，实际上我们已经有很多都已经在 production 上面的。那确实会 Tcon 本身带来一定的省电，对吧？那我觉得我们更多的省电是牵于刚才我们介绍的用 TED 或者 TTED。那这个省电比起 Tcon 跟其他的 architecture 要省电多很多，因为你少掉好几对的 interface 的 power，对吧？我们 customer 更喜欢看到的是这些 TED 跟 TTED。那它的 interface remain EDP，还是 EDP 1.5 的，所以这都已经在 adding to this consideration。那我觉得假如是说 competition wise，那

TTED based on EDP 1.5 的这些 competition 要 relative 更少一点，或者说这个 barrier for newcomer 进入这个领域更加难一点。

Jack 我的第一个问题是，刚刚很多先进都已经有问到

ASIC like，但是我想再聊一些细节。那我想讨教的是说，像我们在类比电路这边的设计蛮强的，在一些讯号的衰减补偿是同期最佳。但是我们最主要的对手呢，他跟美系客户在一些底层的协定相容性很高。就是说双方都有一些优点。那我想问说，那我们可以看到这家美系的这个潜在客户呢？它刚推出的这种新的平台架构，它都在推动里面更高的这种 IO 的带宽。那这个技术上面的趋势呢，有没有机会让我们在争取未来的 ASIC like 产品是一个加分？还是说，到底是我们这种过去做一些 high speed 的进去，做这种比方说做一些 ASIC 机会大，还是说有一些底层的 particle，它可能过去合作很久，那它的客户更改这种 supply chain 的意愿不高。就是说我们应该怎么去看这种各自的优点。 我觉得我们只能谈我们自己看到的。那我们看到的是客户喜欢我们的 proven 的 high speed 的这些 IP，包括 PCIe Gen5，包括 PCIe Gen6，包括 USB4，USB4.2，这些 proven 的 IP对他们 long-term 非常 attractive。这个是特别是在 proven 在 12 或 6 nanometer 上面，这些对 customer 是有最大的吸引力的地方。那，这是我们看到的。我觉得这是我们 invest for long, long time 的这些东西，对吧。那，我觉得对我们说最主要的客户比较喜欢，beside 我们的团队的 execution 的 capability，这个我们大概可能在美国也有一些 advantage，对吧。因为我们有 team，在美国，蛮大的 team 在美国，有些 advantage，因为我们在同一个时区，说同样的 language，这边都 more better，对吧。但是最重要的，最最重要的还是 high speed IP，

proven high

speed IP。OK，Jack，那上游供应链的多角化会是这种客户考量的点吗，还是最主要是回到产品的本身？我不觉得，对最大的客户这些都不是 key，是 proven 的 IP。好，谢谢。现在所谓的这个做 business，做 sourcing 的，会有很多的供应链的多样性，对吧。在做其他的 more engineering，我觉得是 less。因为这些 high speed 实在是两边都需要花 significant amount investment engineering time 去做这些东西。所以，我觉得是大家把事做好是最 critical。好，谢谢

Jack，我的问题先到这边。谢谢。下一位提问者是富邦

Jack 你好，那我们这边请问有关继续涨价的后续效应，就是我们因为目前还是看到蛮多终端的平台方可能有提到涨价的趋势，那这可能会影响我们在 PC 规格升级的速度吗？然后也包含在一些周边配件上面，例如说 USB 4 产品会不会有一些负面影响。我听到的 OEM 是，他们想要 accommodate 现在 situation，是 move 更多的产品到 high-end 去，所以 fully utilize 他们的这个 resource for

memory。这个大概也比较容易懂的就是说，他们尽量去做这些 high-end 的机器，对吧。那这个对我们是 good，因为 high-end 的机器一定会要 USB 4 requirement，一定会要 USB 4 requirement。那会不会他们去涨价，这个机器涨价？我觉得应该会吧，因为你这个这么多 cost increase，对 OEM 大概不得不去做一些 price 调整吧。除了把这个 memory 减少，equipment 减少，或者 equipment 再涨价的。那我觉得这是我自己的观点，我不知道这个 OEM 到底是怎么想法，但是确实我也知道一点，就是他们希望尽量 move 到 high-end 的 segment 去。这个是我被告知的，这个会不会涨价，我觉得这个是我自己的观点，我觉得他们觉得这个大概很难不去涨价的。了解，谢谢。那想再问另外一个问题是在服务器领域，目前是有近期光进同退的趋势又再次引发一些市场关注，那这一块对于公司在服务器蓝线的业务未来会不会觉得说有什么 wanted 的影响，或是观点的改变？我大概同意你啊，这个光进同退的说法。对啦，我们也在观察，也想，因为这个 high speed 这么高速了以后，这个铜线变成越来越困难，对吧，越来越困难。这个确确实实是一个方向，我们也在关心关注当中，但是，yeah，我们只能说我们在关心当中啦。那这个做光进的东西，大概是一个也是蛮大的技术 barrier，要变成这个 fully production 这个 working 的东西。那铜的毕竟是已经比较成熟的这个 service，这个 architecture 也成熟。那得到光的这个要 everybody 要 build 这个新的 infrastructure。但是从理论上说应该是，应该会是这个样子。好，了解，那我没问题了，谢谢。谢谢。下一位提问者是

Lu，请提问。Hey Jack，我想问一下，就是说公司的竞争优势一直在 high speed 这边的一些 IP 啊这些东西。可是状况就是说在 server 这边，高阶的

server，就是我们还在努力中。那 Edge AI 现在看起来需求还是相对弱势。那 consumer 好像也不愿意花钱买真正所谓高阶的机种，因为如果 compute 是放在 cloud 的话，我觉得你现在的客人就专心要做一些高阶的 T 型，好像也没有办法根本上解决这整个 structure 的 issue，对不对？这个局要怎么去破呢？就是说我们是放掉 consumer 这一块，专心做 high-end，然后直接把 data center 这边一路攻下去，还是说我们要看一下在 edge 这边要怎么去，不然在这边现在看起来是两边都是一个僧龙棋局，对不对？这不是一个非常 philosophical 的 good question，是吧。这 question 不光针对我们 Pre，还是针对所有这些做 OEM 的，对不对？无论是 edge 还是 cloud。那我不知道你有关心到这两天的一些 news，有一个叫 Multi-Board，你听到这个消息吗？他们是

run to set mini Apple mini 上面。然后他在 Apple

mini 上面一直在 update 它的 model，所以你就不用花钱你去可以，chat。这个 so hot 的 man突然间把 Apple mini 都卖完了。你听到这 news 吧？你听到这 news 吧？我有听到。你听到这个就是所谓 edge 跟 cloud 之间的一个 compete，对不对？那这个就是为什么说 Edge 有它的一定的意义在那边。当然它不能用 notebook，说它一直要在连着网络，一直要在 update，是吧？一直要在 update。那 quite interesting，当我读了这个东西，我也觉得 quite interesting。因为毕竟，whether 这个 ChatGPT，或者 Gemini，或者 Cloud，我都有去试，我觉得都非常好。那毕竟你还是要花钱嘛，对吧？也不嫌花钱嘛。这个 solution 你是 under control，你不用花钱的，对不对？而且你 can do a lot more。所以，yeah，所以这个就是回答你这个 question，大家都在看什么样最 efficient 的 solution，什么样最 efficient。这个，yeah，我也有朋友也去 try 的，觉得那个真的还不错，真的还不错。所以，当然从我们的观点来说，假如给 Edge 的人有一些这个 room，这个 definitely is a good direction to go，对不对？definitely good direction。那刚才我也跟这个前面各位先进介绍的，就是我们还是有进这些 data center 的 storage，对吧？storage 没有进这些 major computing，但是有进这些 data center 的 storage 上面，对吧？我们有几张位，这些我们的

retirement driver 都有几张位在那边。那我们当然也跟我们的客户去了解那些 inference device。那我觉得我们会有一些 room 去 play 这些 ASIC-like 的这些 device，到这种 room 上面去 play。因为毕竟在……对，那方面上，你在 Microsoft AI 这方面呢？你要放在你的这个 ASIC-like data center 的这些 storage peripheral 这些东西，还是在 edge devices 把它做到最高级？我觉得假如我有

opportunity 去做 data center，是 firm 的 business，我当然会把这个放在最高的，对不对？因为那个 return big。那 otherwise，对啊，otherwise 怎么会有发钱呢？没有没有，我是这样说，假如有这些 opportunity rights，这些 opportunity rights，我不能去随便，我们不能去随便 define 一个 device，然后没有，或者 multiple years，没有 investment return 的，这个我们一定会去找有 very firm 的 investment return 的，as long as we capturing correctly。我想我大概 made it very clear，我们的 investment has to have a return。那我懂，我就这个问题而已，谢谢。对对。但是你刚才的问题 philosophically very interesting，对吧？这个 whether 这个 edge 还是这个在 cloud 里面。但是在 cloud 里面还有一部分是 inference，这个我觉得很多人都已经把这一部分已经忘记掉了。了解。Yeah, inference computing，我觉得这一方面还是有很多的可以 explore 的，可以去做的地方。而且 customer 也 willing to work with you to do it。那里的

market 还 sizeable？Yeah，那两方面就不会要你去

optical，不会要请做这种事情，对吧？好，我知道了，谢谢。OK，谢谢。非常感谢各位的提问。目前没有其他的问题，我们将结束今天的会议。非常感谢各位参与普瑞科技股份有限公司2025年第四季度线上发布说明会。今天的会议将到此结束，您现在可以离线了，谢谢，再见。Thank you everybody.