Thank you. Good morning, good afternoon, good evening, depending on where you are.

It's that time of the year, great to have everybody on the call. We're going to go through the same thing, frankly.

I'm going to talk a little bit about the highlights, and then Wood is going to take us through the the financials, and then we'll go through gaming with Lin Chen. I'll talk about Mynd and education a little bit, and then I'm going to wrap it up with the outlook, and then we'll go into, I think, very active, hopefully, Q&A.

Today's theme is besides going through all the businesses, the progress and all that, I think you're going to hear a lot about AI. We've been kind of talking about and praising AI.

We coined a term called "AI Now." You will see what's going on in 2025, but more importantly, what's going to happen in 2026.

If you look at the AI technology, 2026 is an inflection year. It It really, because of the the the coming of AIGC, is actually it's going to make a huge difference.

Before, AI is there to assist you, but now AI is going to doing is going to do things for you. Lin Chen going to talk a little bit about that, how it's going to help us to the future growth.

I'm going to talk about in the education side, how does it create more opportunity for us. As far as we are concerned, we're going to take advantage of that transition or inflection.

I would say we're going to go from optimization is that to our cost. Now, by the way, we're going to continue to optimize our cost, but more importantly, how do we leverage AI for growth.

With that, I'm going to go into the highlights, start with gaming. As you can hear, we spend a lot of time optimizing our our gaming business.

I'm not going to steal Lin Chen's thunder by going on all the details. If you look at a lot of the key statistics, MIU MAU and all these different things in growth is that it is there.

While we are, optimizing our our cost structure, which is actually quite difficult. My analogy is actually you're flying a plane and you're changing the engine at the same time.

Congratulations to Lin Chen and his team by making it happen. I think more importantly is how do we create native AI games going going forward.

Again, Lin Chen is going to talk about it. The next one, I'm going to talk about frankly is education rather than just Mynd.

It's actually the good news is that we're making progress on Mynd. It is, frankly, in a challenging environment because of the the worldwide economy, the uncertainty of some of the policies in the US.

US being our biggest market. By coming into 2026, we felt pretty optimistic at the beginning of the year, which is January, and then all of a sudden we are seeing a war in in Iran.

What we're going to do is actually we're going to continue with our cost optimization. AI is going to play a key role in that.

I'll talk about our outlook for 2026 for Mynd. At the same time, we extended our our product line, we can expand our addressable market, and again, that is leveraging our big install base of a million classrooms around the world so we can maintain our revenue.

We're seeing growth, and then we are also going to be looking at how we can get back to profitability very quickly. Really encouraging, we're seeing a lot of country opportunities for us, both in Africa and in in Asia, which I will go into a little bit more detail later on.

Some specific opportunity for us leveraging the AI technology, which is developing around the world, and the AI expertise that we have is going to give us a lot of real opportunity in those markets. With that, I'm going to turn it over to Wood to go through the financials.

Wood, it's Thank you, Simon. Good morning, everyone, and let me walk you through our financial highlights in 2025.

Despite a challenging macro environment and revenue pressure, we still achieved a meaningful recovery in the second half of 2025. In this half, profitability improved significantly.

We can see both operating profit grew 47% YoY to 164 million RMB, and half-over-half is also grew 41%. Our operating margin in 2025 second half goes to 7.9%, almost 3% percentage point YoY and H-over-H.

More importantly, our OPEX in second half dropped 38% YoY and 7% H-over-H to 1.3 billion. This demonstrates our cost-saving initiatives and operating efficiency improvements that is transforming into our margin recovery.

Now turning to the full-year performance. Revenue for 2025 was 4.5 billion, down 26% YoY.

A break into this, gaming and application services decreased by 13% to 3.3 billion, and this due to the optimization of our flagship IP to support its long-term development and sustainability. Mynd declined by 43%, primarily due to the disposal of the Singapore early education business in the second half of 2024, and as well as a prolonged demand adjustment cycle.

As mentioned by Simon, once the market condition stabilizes and improves, we will be expecting customer demand to gradually recover. Gross profit was 3.1 billion RMB, down 20% YoY.

Out on our OPEX side, both selling and marketing expense decreased by 13% due to the proactive measures to tighten our marketing spending, and administrative expense down 19% YoY, and the R&D expense down 29% to 1 billion due to the foreign organization adjustment to better align with the AI. When you see the bottom line, profit attributable to the owners of the company in 2025 is 1 151 million.

Bear in mind, we have a two one-off adjustment here. One is the lease impairment in 2025 was about 59 million, and versus the disposal gain in the 2024.

Another one-off, severance payment in 2025, is amounting to 220 million. This paves the way of our cost-saving, OPEX-saving.

Excluded this item, adjusted profit attributable to owner actually was higher than that. Turning to the segmental performance, the gaming and application services, the revenue down 13% to 3.3 billion, and however, our gross margin is still hit at a high level at 87%.

Our core segmental profit is only decreased 3% to about 697 million [goals?] .

Due to we have suffered a lower revenue, but at the same time, it's largely offset by the cost saving. The core segmental profit margin increased by 3.9 percentage point to 17 21.4%.

Talking about the OPEX, the segmental OPEX expenses declined 22% YoY to about 22% due to our AI strategy to fully capturing the cost-saving banner. Overall, you can see our gaming and application services is still highly profitable and cash-generative.

Talking about the Mynd business, the revenue declined 43% to 1.2 billion, primarily due to the disposal of the Singapore early education business and a prolonged demand adjustment, and the gross margin we still keep stable at 25% this year. While the core segmental loss enlarged to about 358 million RMB, but our operating expenses is declined 21% YoY.

The Mynd business is at a better position for the growth in the next few years. In summary, while 2025 was a year of revenue adjustment, we have substantially restored profitability momentum in the second half of this year and through the cost optimization and strategic focus.

Our gaming business is remain profitable and highly cash-generative, while Mynd business is at a better position for the once the market condition is improved and stabilized. Now I hand over to Lin Chen talking about the gaming business.

Lin Chen,

please. 大家好, 我是林琛, 很高兴为大家介绍网龙游戏2025年的业绩。那首先在游戏业务上,

我们保持打造健康运营的长线IP的战略。那从游戏的MMO品类来讲, 用户的活跃、用户的回流跟新内容的用户接受度是我们一直以来最重要的MMO的观察的指标。那我们很高兴地看到, 对于一个长线运营的MM MMO的游戏来讲, 这些数据在2025年都是非常正向的。那这也我们也对主动对用户的口碑, 既用户的口碑、用户的接受度主动做了一些调整, 这个为我们未来长线的运营打下了很好的基础。那今年以来呢也是《魔域》20周年, 那作为一个长线运营的游戏, 我们也非常 魔域IP在未来能够充分发挥更大的活力。下一页。那在今年的话呢, 我们很高兴看到我们主动做了一些调整, 那尤其是在无论是付费门槛、回流门槛上。那在我们的核心的幻兽, 无论是我们的年兽还是国庆宠的售卖上, 我们都很高兴地看到说随着我们的用户的一些门槛的降低, 那用户购买的购买率是显著提升的。那同时呢, 我们在用户的资料片上也降低了回流的门槛, 那我们也可以看到说用户是对我们的回流的数量有一个很大的促进。那这也给我们的长线的用户的在线跟运营带来了一个很大的帮助。下一页。那在这个游戏的MMO品类的长线的内容营销上面, 那我们在AI的帮助下, 那在今年我们不仅在每个资料片去推出一些新内容, 我们有能力在我们的长线运营的资料片上同时去优化我们过去的一些已有的系统, 去推出新内容, 去推出文化向跟内容向的组合式的内容营销内容。这个对我们的游戏的长期的稳定跟用户的生态是一个很大的一个助力。下一页。那可以看到在我们的《征服》IP上, 事实上我们在之前就主动做了一些优化跟调整。那这个从我们的运营数据上就有一个很及时的反馈, 那就证明说我们在MMO赛道的积累是可以复用是而且是可验证的。那首先会看到是说游戏的内容的出海, 那也是我们整个国家所推的一个很大的一个战略。那我们随着这个借助国家这个IP的一个东风跟文化的东风, 把我们的中国式的内容更好地推到了全球很多个国家。那这个就推动了我们在海外收入的同比增长, 并且克服了汇率的一些困难。同时呢, 在MMO这个大品类上, 我们的MMO加休闲的一些融合也得到了用户的认可。那我们把这些的玩法单独抽离出来的做了一个单独的游戏, 那《Alpha Code》。那前期在埃及、印度等落地, 已经得到了非常正向的反馈, 拉动了我们的收入增长。在下一步的话呢, 在更多的地区, 我们也会跟我们本地的一些代理商去合作, 把更好地把我们这些的玩法去带到更多的国家。所以说大家会看到在《征服》IP上下半年同比增长的幅度更大。那随着我们这些策略的推动, 在今年也会延续一个非常好的一个发展的趋势。OK, 下一页。那接下去是我们的这个《英魂》IP, 也就是我们在MOBA赛道的一个布局。那其实从3年前, 我们就基于长线运营的这个策略主动去优化了我们整个的内容策略。那我们也很高兴在内容加赛事的双引擎的驱动下, 我们做到了这个今年整个《英魂》IP的收入新高。那这个不光是我们的这个内容营销的部分, 同时呢是我们的整个的商业化是有所提升的。那整个从ARPPU的上面也可以明显看到我们的商业化内容被更多的玩家所喜爱。那我们前期在游戏加文旅、游戏加公益等内容上的一个铺垫, 也会对我们的用户在今年度的一个对IP的接受度的提升跟收入的提升打下了良好的基础。那我们也希望推动这种趋势的不断地扩大, 再延续到今年。下一页。那接下去是我们持续坚持的这个游戏作为一种文化产品在内容上的深度的不断地拓宽。游戏无论是游戏加文旅、游戏加公益, 还是游戏加非遗, 那过往呢我们其实已经在单点上获得狠狠高的一个收益。那今年很重要的的话呢, 我们是把这个游戏加文旅做成了一个系列化 的一个工作。那我们在去年其实已经在政府的帮助下去整合获得了整个福建省的10个文旅IP。我们有机会把在今年把它做成一个完整的系列, 去推动我们整个IP的的用户的扩盘跟发展。那我们也很高兴今年在今年的春节, 福建的文旅无论从收入的同比、人流量上都有一个很明显的提升, 也证明我们去布局的这个文化内容的眼光是没有问题的。同时的话呢, 随着线下的经济的发展, 也能带动我们整个游戏O2O的联动去扩大我们的IP影响力。同时在AI生产力的助力下, 无论是IP的衍生品、AI的漫剧, 还是AI的互动小游戏, 我们都可以基于这些文化内容推出更多更好的游戏产品。下一页。那接下去就是我们最重要的在AI上的投入。那第一点是非常高兴地看到我们其实从前年开始就对AI有了布局跟整体的转型。那在上次的这个发布会上, 我们其实有提到我们的目标是做到已经做到了15%的AI增效, 那目标是25%。那事实上在2025年, 我们也很好地达到了这个目标。那会看到无论是我们的这个利润率, 还是我们的研发投入, 在我们的整个长线运营的对它的一个降本增效的情况下, 都从数据上达到了我们预期的目标。那同时的话呢, 我认为还有很关键的一点是在去年2025年或者之前, AI是作为一个工具存在的, 那它更多的优势发生在所谓的增效上面。那在今年呢, 在去年的下半年, 我们重点是搭建了我们这个AI的生产中心。那AI的生产中心已经不是把AI当做一个工具, 而是把AI当做一个Agent或者当做一个员工来去建设, 与我们的未来的AI型组织去构建我们整个游戏的AI生产能力。那也就是在今年, 它也会产生一个很重要的一个质变。在今年我们其实目前已经上线了200多个的AI员工或者AI Agent。那我们很有信心在今年把这个数据做到1,000以上, 甚至甚至是5,000的水平。那甚至那这个时候呢, 对我们的AI工作量在整个研发的工作量占比就已经可以拉动到50%。而且在AI的这个趋势下, 我们很很有信心在明年可以把这个比例做得更高。那这个对我们意味着什么呢? 也就是说这个会对我们在整个游戏研发方式上是有一个很大的创新。那就会可以同时对我们整个研发在生新游戏的研发能力可以做到2到3个人的团队, 在一个月之内就可以完成我们一个轻度研发的一个急速的交付能力。那同时的话呢, 基于我们一些擅长的赛道, 尤其是MMO跟MOBA赛道, 我们很有机会把一些休闲跟RPG的融合的玩法的这个know-how能够放大到更大的范围之内, 去拓展更多的一些市场机会。那在将来面向未来的话呢, 无论整个集团还是游戏业务, 最重要的就是往这个AI型组织去转型, 去充分发挥我们在AI上的一个更好地发挥的一个竞争力。那因为我们在这件事情上的布局从两年前就开启了, 所以说我们现在也有一个很好的一个优势用位置。OK, 下一页。简单给大家举个例子。那事实上其实我们在过往曾经也有立项过无论是Roguelike还是三消的游戏。那样子的一个游戏在传统的模式上都是需要几十上百人的投入, 而研发周期可能做一个demo也需要3个月这样子的一个时间。而对我们实际上在去年下半年跟今年上半年已经实现了一到两个主创, 通过十几个人AI Agent的帮助下, 我们在两到三周之内就可以拿出一个可玩的一个demo。那这个在原来我们是不可想象的。那这至少是10倍以上的效率提升。我们会看到一个美术出身的一个制作人, 他可以在我们的AI创意中心里面跟我们的AI的系统策划、AI的数值策划、AI的关卡策划, 在用自然语言去讨论三消游戏应该是如何设计的, 并且呢AI通过跟这个主策划、跟制作人的互动, 就能够产出让他满意的一个案子, 而这些案子就可以直接进入到我们的AI生产中心去生成我们的美术资源跟我们的代码, 而最终就能产出一个可玩的demo。而如果这个在我们的今年也会跟我们的AI运营中心的智能投放系统整个的串联起来, 就可以形成一个完整的闭环, 去更好地去去加速我们整个AI的落地的进展跟商业化的一个计划。这个在我看来是一个从一个量变到质变的过程。当我们的渗透率研发的这个替代率到了50%甚至70%, 马上对我们就是一个很大的一个优势跟帮助。下一页。那除了这个AI对我们的整个研发效能上的一个提升, 那其实在AI我们非常另外一个非常看重的也就是对玩家体验的一个优化跟提升。那AI呢目前来讲我们把它分成两个维度。第一个呢, 它可以通过反外挂做一些这个陪玩去很大程度降低玩家的负面体验, 那这个已经是通过一些技术上已经实现并验证的。那我们还有一个很重要的也就是我们的重点就是我们的智能NPC。那在过去呢, 我们是对它一个智能能力的研发, 比如说它如何可以更好地跟玩家对话, 如何更好地去交互。那在今年的话呢, 事实上在上个月, 我们就已经把1,000个智能NPC去投入到了我们单个的MMO的服务器里面。那这些的这个NPC产生已经把我们原来的智能的能力转换为我们的互动行为。那我们也很高兴说在这1,000个NPC的加入下, 玩家是非常愿意跟这些NPC产生一些互动, 甚至能够产生我们在设计之外意想不到的一些有趣的体验。那这个对我们来讲的NPC的产品化, 包括下半年会更好地产生一些收入上的帮助是非常重要的。下一页。那总结看来得话呢, 整个AI的推动在今年的2026年, 它的效果是一个不断放大的过程。那我们也希望在在这个整个产生闭环以后, 能够实现很大的一个商业化的一个一个助力。那把它分成两个维度: 第一个呢, 对于我们长线的IP的长线运营来讲, 那它无论对我们的内容研发的持续的降本增效, 还是对于我们的体验的提升, 都有一个很大的助力。尤其是智能NPC在今年下半年我们是非常有信心把它转换成我们对我们实际留存跟商业化收入的提升上。那同时第三个也是一个极具想象力的内容, 那这个也是整个研发方式的颠覆。当我们把 我们的运营创意、把我们的研发能力跟我们的运营能力做一个闭环完以后, 那AI可以是一个自行运转的系统。它可以通过最新的AI的搜集信息跟内容跟制作, 给我们的游戏源源不断地带来创意跟内容。那相信在AI的这个助力下, 今年我们下半年的长线运营游戏, 它的这个收入跟利润肯定是会有一个很大的一个助力。那接下去在新机会的探索上, AI对我们带来的还有两个很重要的维度。那第一个呢就会它极大能够拓展我们在我们已有赛道上的一些的 想法的扩展。比如说为什么我们去尝试三消跟Roguelike赛道呢?

也就是说轻加重的游戏模式, 无论是业界的《弓箭传说》啊或者各种各样的游戏, 它都证明了一个前端非常适合玩家的休闲的一个入口, 加上后端的一个成熟商业化的模式, 已经是被市场验证的赛道。那在AI的助力下, 我们可以快速不断地去迭代这些的休 休闲加重度游戏的一个组合, 去拓展我们的新的产品机会。那同时呢, AI在一些经典游戏的IP复活上, 也可以让我们去把游戏 我们原来网龙历史的IP去更好地放大。那同时呢, 基于智能NPC, 那原生AI的原生游戏加上AI的制作能力, 就会使我们在的探游戏的新游戏赛道的探索能力有一个显著的增长。那总结起来看呢, 游戏业务呢通过长青游戏的IP的策略, 我们有信心保持基本盘的不断的稳定跟健康发展。而在又AI的助力下呢, 我们在今年下半年通过AI在实际的产品化数据上的落地跟新新机会的探索, 我们相信在今年下半年跟明年, 游戏业务都能保持一个很好的稳定跟恢复增长的一个趋势。好, 谢谢大家。 Thank you, Lin Chen.

Now now you know what I meant by AI everywhere and the AI Now, like I mean everyday Lin Chen talk about AI a lot. I would imagine he probably wake up with AI in the morning and then sleep with AI at night.

I think that's what he does. What I'm going to do is that yeah, I'll go through Mynd and education a little bit, and talk about our AI Now initiatives, some of our commitment, and then we'll go through the outlook of the year.

This is, actually, the financials. Not pretty, but it's encouraging because we're improving year-over-year.

We're going to be taking very aggressive cost-cutting both from the optimization of people, using AI, leveraging AI, exactly, in a way is no different from what we've been doing with with gaming. We're going to do it exactly the same with Mynd and with education going forward.

While we're waiting for the market to recover, it will recover. We're going to see some movement in 2026.

If the Middle East situation improves, I think we'll be even more obvious growth coming back. The next, while we are doing a lot of optimization, we also bring out new product.

Our flagship AP10 came out last year. This year, it's going to be the first full year that with AP10 in in action, which is, actually, very encouraging news.

The reason we spend so much time on AP10, it's actually, again, it's related to AI. In order to deploy AI services on a panel, you cannot rely on the old architecture because it's, actually, is Android-based, very limited.

What we did is that we unbundled the architecture from or we unbundled the the OPS from the panel, we have flexibility of either putting in doing an Android OPS or a PC or even a Mac to allow us to deploy services on the panel to help students and teachers teaching and learning along the way. Again, that's AI everywhere.

Next. To address the the low-end market, we help us to expand the market, we brought out two things.

One is, LE, which is at the lowest end that we have, we can bring it into Asia-Pacific, Africa, and those emerging countries. Also, we brought in a digital signage that we can complement our panel to go into the school, because there's some requirement for digital signage in different schools in the developed market.

Next. Another very encouraging news at Bett, one of the largest education show in in the world, it's actually that was held in London every year.

We won the Company of the Year Award, so that's very encouraging to the team, and also we have the Best Show Award at ISTE, which is in the US. We continue to gain recognition from the industry.

Next. Let me change gear and talk about some of the company-wide, at the group level, the initiative.

Next. The first one is, again, on related to education, but also related on some of the opportunities that I talk about either in Africa or in Asia-Pacific.

I'm going to give you two examples. I think you have you heard of one of them, or maybe you have heard of both of them before, but I'll go into a little bit more details.

It's moving really fast. In Thailand, you know we've been working with the Higher Education Ministry to provide training, to help them to expand their their industry focus, we help them to move into EV, to move move into semiconductor.

While they are changing the government, we never stop. We are working with them in creating new learning for them.

The big one is, actually, is helping young people in Thailand to learn how to use AI, which is, actually, very important. I think everybody knows about AI, but how do you use the different tools?

How do you use it to apply to your day-to-day work? More importantly, we are also helping them to assess the your learning, because it's after taking a course, how do you know you're progressing?

Assessment and keeping track of your improvement is very important. Ultimately, how do we help you after learning all the AI tools to to help your career is an important element.

Not only we work with the ministry, but we also work with big corporations in Thailand, because hiring is quite different now. Before, you hire a course according to your job spec.

Now, a lot of people are looking for candidates with the right quality, and then you bring them in, and then you put them into different jobs so they can flourish, because with all the AI capabilities that they have. The next one is, actually, is Cameroon, which I personally attended.

This is the opening of one of the intelligent classroom. What they struggle with in Cameroon, because being a developing countries, is actually teacher quality, is ability to bring software and features into the more remote area.

This whole idea of intelligent classroom, it started in Egypt, and we're now bringing it into Cameroon. Actually, I'm in the picture.

You may not see it. It's really satisfying sitting there looking at the teacher telling the Prime Minister, who sat right next to me, how are they leveraging our technology to address one of the key needs they have, which is education.

I'll talk a little bit more about the new concept that we have to help to help them even address that education need further in Cameroon and in other emerging countries. Next.

This is a country-wide, sorry, a company-wide initiative. We call AI employee, which this is kind of like a workflow.

This is, actually, the workflow for our AI content manufacturing, if you will. It's actually how do we leverage AI to do a lot of the work for us.

This applied both in education and also in what Lin Chen was talking about. Just to give you an idea of how we approach everything as a company.

Next. Let's talk about outlook a little bit.

I know you you want us to to do that, next page. We'll continue to commit to enhance the shareholder return with all of the asset and with the trust that you have in us.

We'll commit to provide a dividend of no less than $1 a year. I think we announced $0.50 for the for last year, the second half.

To go forward, we're going to commit the same, if not more. The core segment is, actually, we're going to maintain our profitability while we are applying AI, and we will return to to to growth is, actually, towards the end of the year.

We're going to maintain our net cash. Between dividend and also buyback, we're going to return no less than $600 million to to shareholders.

The last one is, actually, we have been investing in different companies. We are seeing the possibility or the profitability of us realizing some of the some return.

I put XXX. In hindsight, I should put dollar sign rather than X.

I think that will look more sexy. Hope that's the point.

We're expecting hundreds of millions of return in that one. Cannot commit, but it's, actually, is a line in sight.

Next one. Another very encouraging news is, actually, every company is trying to strive to, which is ESG.

We got rated A, which is very, very encouraging, and very We're really happy with that. Next.

Last but not the least, I'll make some comments from our directional standpoint. I think we talk a lot about gaming.

We we believe we're going to continue to apply AI to help us to optimize our costs and to position us position us to growth towards the end of maybe the second half of the year, so that's very important. From education, for Mynd, we're expecting a break-even EBITDA year.

That's through very aggressive cost optimization, and also we are only assuming a slight market recovery. We know the Middle East is, actually, is uncertainty, so that's why we are using very aggressive cost optimization direction to make sure we go into a EBITDA break-even situation.

We optimize our hardware product portfolio by going deeper and wider, i.e., companion product with our panel. We talk about AI.

The key is, actually, how do we bring new services into our classroom solution, i.e., leveraging our AP10 to put services into, we can build our SaaS revenue stream. We are eyeing our country strategy, like Cameroon, like Thailand.

I can assure you those are not the two countries that we're working on. Last but not least, I want to touch on only a little bit.

I want to tear it up. I do intend to take everybody through the entire strategy going forward, both the investor community and also the PR community or the media, which is what we call a sovereign education large language model.

The reason for that is, actually, in twofold. One, education is, actually, is every country's asset going forward.

Most countries wants to be in control. Not everybody can own his own ChatGPT, or Gemini, or DeepSeek, or Qianwen, and all that.

Having their own is important. We're going to leverage the open-source large language model, but we localize it either for the country or even for the school.

I think that's the other important message. Going back to some of the developing countries, because they don't have the telecom infrastructure, having a local large language model, i.e., like a DeepSeek sitting on a PC to support the whole school is going to be very, very powerful.

We're moving towards that, and actually we're working with Cameroon on that one to address one of their key element of problem that they have, which is the lack of infrastructure. Now this also applies to a developed country or area.

Hong Kong just announced a I think it's HK$2.5 billion initiative on AI, exactly, each school will get maybe $500,000. Having a large language model for in each individual school tailored to their individual needs is going to be very powerful.

We are looking at how do we take this not only to the developing countries, but also to the more developed either region, countries, or what not. This is really a global initiative that we're working on.

To sum it up, we did a lot of work in 2025. We'll continue to do a lot of work leveraging our commitment and our knowledge of AI.

The change in 2026 is not only we use AI to optimize our costs, but more importantly, to use AI to to leverage AI for growth. With that, thank you so much.

I think it's time to to open it up. Thank you.

I'll answer the first one, the easy one, and then Wood answer the next one. If you can put the chart back up, the one with all the commitment to I think this is how we do it.

Exactly, let me read to rate. 2026 dividend at no less than $1, okay?

Exactly, the total return will be no less than HK$600 million. Exactly, it's kind of the same commitment as last year, but yeah.

How do we fund it, exactly, you look at the other three columns. Exactly, we maintain our segmental profit in gaming, where you go into EBITDA break-even in Mynd.

Exactly, we continue to optimize our costs. Exactly, because of that, we expect to maintain our cash level, if not doing better.

The last one is, exactly, the upside, right? Now, don't know whether this is going to be next year, because that's why we said next 12 months.

Because now we're in March, we will see some of the investments that we have made mostly in the AI area that will come into fruition. That's what I committed.

I kind of anticipated that question what will come up. Wood, do you have any other comment to add to the cash flow?

可以啊。就是那个你看到是今年我们

就是我们的这个就是那个今年的现金流其实是3亿 3亿多,

就可能有些投资者担忧就是说我们发1个亿的那个股息是怎么半, 但可以想象就是说, 第一就说其实我们对于我们游戏的分类其实还是有一个9亿的一个core

我先我先来介绍一下吧。就是从首先从AI的这个增效的这个效果上来看, 那我们前年是做到了15%,

去年做到了25%, 但今年我们的预期是可以做到50%。其实从这个这个比例上就可以看到,

它其实今年也是会有一个显著的变化。那从这个实际上的AI技术来讲, 它其实也是一个非常��速发展的一个一个一个趋势。尤其在一线我们看, 比如在AI编程上, 随着大模型的进步, 那我们在之前AI AI编程可能可以帮助我们这个增效20%, 但只要大模型正 这个更新了, 它马上这个数据可能就会变成30%到40%。所以说的话呢, 可以可以看到说我们为什么对下半年是很有信心的, 因为随着我们对AI的使用率跟所有的同学对原 AI能力的技能掌握的提升, 包括我们AI员工的搭建, 它下半年会有一个这个加速的一个一个趋势。那尤其是说呢, 游戏的品质它需要突破一定临界值, 它的效果才会成倍地放大出来。过往大家会看到一些AI做的一些原画可能用户对它不是那么买单, 是因为它可能离那个标准只就差了一丝, 所以问 玩家觉得说你AI做的东西品质会是不是不用心。但如果说我们很用心地把AI用好, 这件事情一旦突破临界值, 玩家对我们现有的AI产生的内容的接受度其实是有个质变的过程。那所以说游戏环节的这个无论真效的这个数据, 下半年可以看到至少会有还有个10%到15%的一个增幅的这个效果, 可能就相当于去年的一个比例。而在品质上的接受度, 那我们也很有信心在下半年AI的快速发展下, 品质一旦突破临界值, 用户的接受度就会成倍的的这个提升。那这个是在我觉得是从研发的这个成本结构跟收入结构上带来的下半年我们的预期。对, 那财务的部分可能就有请Wood来进行做一些回复, 谢谢。 OK, 我插一句吧。因为AI对我们来讲是有两两个重要的那个环节哦。第一个就是那个 cost optimization, 但是另外一个就更重要, 就是怎么帮助我们增长哦。所以这个大家希望注意一下。所以从某个角度我们用AI帮助我们减成本, 但是另外一个方面呢我们也会投入在AI帮助我们以后的发展, 是我们拿一个平衡。在教育方面呢, 因为其实老实讲那个游戏走的比教育比较快, 所以我们要学习利用它的经验,

所以那个我觉得在教育里面呢我们的空间会比游戏更大一点。Wood, 有没有其他的comment? 对, 如果是从财务影响的话,

刚才提到是说我们2025年我们的OPEX, total

可能我讲的不够清楚了。AI绝对可以帮助我们, 所以我讲很多次就说不光是帮我们减低成本, 也会帮助我们是就是发展。我们在Mynd最大的优势就是我们有我们的install base,

所以呢我们都可以把AI的服务放进去, 在教室里面帮助那个老师来教学, 也帮助那个学生学习, 然后把那个服务也把它搬到家里, 所以在家里也帮助学生在家里学习, 但是也更好的去帮助那个家长怎么去跟那个去跟孩子去学习。所以这个是我们在Mynd来做一个变现的一个非常非常好的机会, 所以你看我们整个公司为什么说我们真拥抱那个AI就是这样。美国市场一定会回来的, 因为他们是那个他们很需要那个新的科技, 有两个原因,

这回答为什么也是level

Wood, you want to take the CAPEX? Yes,

yeah.

Good.

Of course we know we are AI now, put our CAPEX in the AI area. When you look at the 2025, our CAPEX is about 150 million RMB, we are on us to double our investment in the including the AI area, we will be expect to have 300 million RMB in 2026, the CAPEX including the AI area.

I'll go back to the first one. Number one is, actually, it's going really well.

We're in the verge of announcing a few new products not only for Hong Kong, but for overseas. I'm not going to steal Cherrypicks' thunder.

When that happened, which is imminent, by the way. I invite you to join us in some of the conferences and all that.

The other one is, actually, when I talk about the sovereign large language model, which is, actually, a localized large language model. I did use an example of maybe DeepSeek, but definitely Wenge's large language model is part of the consideration.

We're very happy with the progress, and we are very excited about the prospects. I probably described it at the high level right now because I don't want to steal the thunder from Cherrypicks.

I can rest assure you, it's actually our relationship with Wenge is going to be coming to fruition sometime this year.

I have to say I thought I'd talk about all that, but Can you put the chart back up so on Thailand? This is actually what's happening.

Exactly, we are moving forward on what we did, but we're expanding into the area of very important is to use AI to teach people how to use AI tools, and then at the same time we helping them to assess. We're also helping the enterprise to look at how they can kind of pivot to a new way of recruiting talents.

Like I said, before, you have a job spec, you look at the experiences. I think now the new way of recruiting, you're looking for talents, you're looking for capabilities.

That's what we do. It's moving really fast forward, because the encouraging news is, actually, the election is over, actually the party that's committed to working with us is, actually, returning to be the main part of ruling party in Thailand.

We fully expect this to continue and to expand. I can't talk about some of the other ones.

I can assure you Cameroon is not the only country that are working on. Just to share with you, not try to be too personal, my wife is actually not very happy with my travel schedules.

I'm traveling a lot to work on different projects. The other thing, I don't know whether you remember last September, I was in Paris with the team attending UNESCO conferences, so that was very good success.

As a result of that, there's a lot of countries in Africa, in Middle East, and also in Southeast Asia to come to talk to us. Our key is, actually, is not to work on every single one, then the key for us is, actually, to focus on the one that we can bring a lot of value to.

Thank you. 我们下一道问题,

OK, 我其实我也简单的讲过那个我们怎么去用AI帮助我们在教育发展, 我再讲一遍。其实分成两方面呢,

一个就是我们的install base,

我们就利用我们那个普及在100万的那个教室里面, 我们都有因为有机器, 我们把新的services放进去。譬如说我们怎么去协助那个老师来教学, 帮他备课啊, 帮他出题目啊, 要帮他改作业啊, 或者是更重要, 他在教室里面怎么去帮助他来那个教学的, 去配合他了, 好像是一个类似一个助教的一样在里面。然后呢, 这个在这个学生里面呢, 我们也可以看那个学生学习的态度啊, 他有什么不明白,

technology, 整套去卖给一些不同的学校也好, 一个地区也好, 国家也好。另外一个呢就是那个emerging

都把我们那个叫那个sovereign large language model for education把它嵌入, 集成到我们Promethean的那个offering, 可能用一些比较我想想低端的, 不是低端了, 就是比较value的那些panel, 也是去帮助他们。这个最关键的就是我们在那个设计的都不需要他们有网络的存在, 在那个学校一个large language model就可以支持整个学校的教学, 这个是空间是非常非常大, 所以那个为什么喀麦隆那么喜欢, 香港也有很多学校也很欢迎。所以那个是我们对这方面信心是挺大, 因为我们花了那么多心血在AI, 也是知道AI的发展是怎样, 所以利用这个AI继续帮我们的发展。好, 谢谢。 谢谢Simon。由于时间的关系呢, 我们接入最后一道问题。那最后一道问题呢也是刚刚Simon提到的这个主权教育AI这方面的问题。投资者对这个很感兴趣啊, 因为主权教育AI呢这个市场很广阔。那在Simon您的预想中, 我们未来可能会以怎样的一个合作或者收入模型来就是推广这个主权AI呢?

我们现在有没有一些诶很可能就会比较发生的具体发生的事情呢? 你再讲一遍后面那个, 我知道 repeat。

对, 我们未来我们即将我们很快可能会即将看到一个怎么样子的一个主权教教育AI这方面的一个progress? OK, 其实我们老实讲很多。但是我们知道我们一定要跟一些伙伴来一起做, 因为那个世界太大了哈。譬如说在那个美国我们当然是会充分的利用我们Promethean的那个能力, 但是Promethean也有很多那些partner, 譬如说那个CDW啊, 我们新进来的Synnex啊那些帮我们来做一起做。但是那个因为那个sovereign LLM是比较特别, 所以我们自己也有参与, 无论是在marketing方面啊, sales support我们都会做, 我们给地方会加强。譬如说在那个我们叫country strategy了, 我们会跟一些local的partner来合作。因为譬如说那个喀麦隆我也不能去很多次吧, 所以我去都是重要的人去, 其实我的同事也是受到很多折磨, 就是去那个喀麦隆, 从香港去要飞土耳其, 然后再飞过去。飞土耳其是12个小时, 那个从土耳其飞到喀麦隆有7个小时, 中间还要等, 这个是蛮辛苦的, 所以那个我们一定要利用那个本地的能力。其实这个也是那个为什么喀麦隆那个总理那么欢迎我们那个方法, 不光是他们可以解决他们一些本地教育的问题, 也可以帮助他在科技里面所有发展。所以我们也跟本地的伙伴呢, 也跟他那个教育部里面的IT的人, 我们也希望培训他们可以支持这个事情。

譬如说在香港来讲, 香港有2,000个学校, 有一些我们很熟识, 但是有些我们也不是那么熟识。所以我们都看情况, 我刚才讲了, 譬如说我们的LLM会放在一个server里面, 所以很自然一些硬件的厂家、有品牌可能第一个那个是伙伴。譬如说在一些提供一些譬如说通讯服务的给这些学校的一些公司也可能是我们的一个partner, 或者是做一些SI的是一个partner。所以在这方面呢, Cherrypicks跟闻歌在香港也是我们的partner。我希望都给你一些概念, 就是不好意思没有现在目前来讲不会详细来讲谁是我们的那个伙伴业务伙伴在谈这个事情。