高通 CEO 谈机器人、芯片与对华竞争

Cristiano，欢迎来到 WP Studio。

Cristiano, welcome to WP Studio.

不，谢谢你邀请我。很高兴来到这里。

No, thank you for having me. Happy to be here.

我们直接开始吧。我们想谈谈你的公司，它涉足很多领域，而且越来越多。多元化对你来说几乎是家常便饭。个人电脑芯片、智能眼镜、工业系统等等。当你考虑未来 5 年时，高通最大的研发赌注是什么？

Let's just dive right in. We want to talk about your company, which is involved in so much and increasingly in more. Diversification is kind of the name of the game for you. Chips for PCs, smart glasses, industrial systems, much more. When you think about the next 5 years, what are the big R&D bets for Qualcomm?

看，这是个好问题，我认为有几个。它多少遵循了我们所进入的不同行业的趋势。也许我提供一些背景信息会更容易。我们的多元化之旅始于几年前，当时我们意识到我们为移动行业开发的一些技术可能与许多行业非常相关，然后我们就开始了，随着我们深入这些行业，我们最终建立了一个平台，现在我们的路线图多少是由此驱动的。所以在这个背景下，你知道，我们从手机到汽车，正如我所说，到个人电脑，到工业，到我们现在称之为个人 AI 设备、机器人，一直到数据中心。与同行相比，我们可能很独特，因为我们从通信和计算，从 5 瓦的耳机芯片一直到现在的 500 瓦数据中心芯片。所以在这个背景下，当我展望未来 5 年时，在这个范围内，我们今天正在进行一些大型研发赌注，我认为这是正确的问题，因为有些技术投资我们必须提前几年进行。高通的一个独特之处在于我们多少是标准的创造者，因此第一个大赌注是 6G。6G 将在本十年末开始进入市场。我认为 2028 年预启动，然后 2029 年你将开始看到发布，并且

Look, it's a good question and I think there are a number of them. It kind of follows the trends of different industries that we enter. Maybe it's easier for me to just provide some context. Our diversification journey started a few years ago when we realized that some of the technologies that we have developed for the mobile industry could be very relevant to a number of industries and then we started that way and as we went deeper into those industries we end up building a platform and now our road map is kind of driven by that. So with that context, you know, we went from phones to automotive as I said to PC to industrial to what we call now personal AI devices, robotics and all the way to the data center. We're one of the probably when we compare ourselves with our peers, we're very unique given that we go from communication and compute and we go from five watts like to an earbud chip all the way to now a 500 watts chip for the data center. So with that context when I look at the next 5 years within this surface area there's a number of big R&D bets that we're making today and I think that's the right question because some of the technology investments we have to do several years in advance. One of the unique things about Qualcomm is we're kind of a creator of standards like and with that is first big bet is 6G. 6G is going to start coming to the market at the end of the decade. I think with pre-launch in 28 and then 2029 is when you're going to start to see launches and out of

你个人的，你知道，你的手机到本十年末将是 6G。你是这个意思吗？

Your personal, you know, your cell will be a 6G by the end of the decade. Is that what you're saying?

你的手机和其他设备也是。但我们肯定可以

Your cell and other devices as well. But we can definitely

不管是什么设备，

Whatever the devices are,

我们乐意详细说明。第二个大赌注是，随着我们继续深入汽车行业，你会看到我们刚刚推出了辅助驾驶的技术栈，我认为这将继续向全自动驾驶演进，加上汽车中不断增加的算力和智能体式用例。第三个赌注是移动设备如何演进，我们谈论这个新类别的个人 AI 设备。我们在这个十年中有一个重大变化是人机界面发生变化，因为现在 AI 理解我们所看到、所说、所听到的一切，然后计算也随之改变，我认为我们现在对移动设备如何成为个人 AI 设备有了愿景，这将从根本上改变可穿戴设备。机器人对我们来说是一个大赌注。物理 AI 全是关于 AI。

We are happy to unpack that. The second big bet as we continue to go deeper into the automotive industry you'll see that we just launched our stack for assisted driving and I think that will continue to evolve towards full self drive plus all of the computing that continues to increase in the car agentic use cases. The number three bet is how mobile devices are evolving and we talk about this new category of personal AI devices. One of the big changes we have in this decade is the human computer interface changes because now AI understands everything that we see we say we hear then the computing is changing as a function of that and I think we have now vision about how the mobile devices are going to become personal AI devices and it's going to fundamentally change like wearables Robotics is a big bet for us. Physical AI is all about AI.

我们暂停一下。我想问，当你审视机器人技术时，确实存在算力问题，对吧，技术、AI，但这也是一个物理问题。你需要机器人技术有哪些阶跃变化才能实现广泛普及，你知道，彻底改变养老护理，让每个人家里都有机器人。在你看来，物理世界还需要什么才能实现这一点？

Let's pause on that for a second. I want to ask when you look at robotics it's right there's the compute issue right the technology the AI but it's also a physical problem. What are the step changes you need in robotics to get widespread you know revolutionizing elder care having everyone having a robot in their home. What more do you need in the physical world to enable that on your end?

是的。顺便提一下，我也提到了数据中心和糟糕的力量，但让我回到你的机器人问题。我喜欢在辅助驾驶和自动驾驶中看到的与机器人技术之间做个类比。例如，创建机器人出租车的问题，你必须做大量的训练。你可以让 AI 和能力从 0 到 95%，然后有司机在那里以防万一。如果你移除司机，那么你需要达到 99.9999%。所以额外的部分需要大量训练。为你提供了一个巨大的市场机会，为路上的每辆车提供不同级别的辅助驾驶。为什么司机还在那里，然后最终你会有没有方向盘的完全自动驾驶系统，这需要大量的技术开发。我现在把这个应用到你的机器人问题上。一个将在你家里和你在一起并做你要求的一切的机器人将应对每个房子都不同。我们会处理它，学习新任务。我认为这将需要大量训练，并且需要一点时间。然而，有一个巨大的商业机会，让你得到一个机器人并让机器人在一个特定任务上完善，这实际上是一个更容易解决的技术问题。这容易得多，因为你可以在虚拟数据上训练机器人，然后你可以做我们所说的远程操作，让人类完成任务，然后你知道那是物理训练，然后你让机器人完成那个任务，我们现在看到一些有趣的应用，例如当你想到一个 AMR 机器人，一个有两个轮子然后知道导航门和两个手臂的机器人，你可以把它放在医疗保健中，你可以把它放在零售中，零售会在半夜去上架。所以你可以在那个特定环境和任务上训练。我认为那个技术现在可能就在我们面前。让你的机器人伙伴进入你家的技术，我认为还需要几年时间。

Yes. And by the way I mention also data centers and the bad force but let me go back to now your robotics question. I like to do an analogy between what we seen in assistant driving and autonomous driving in robotics. The problem of creating like a robo taxi for example, you have to do a lot of training. You can get the AI and the capability like from zero to 95% and then you have the driver there in case something happens. If you remove the driver then you need to get to 99.9999%. So the extra requires a lot of training. There's a massive market opportunity for you to provide assisted driving at different levels for every car in the road. Why you have the drivers still there and then you have eventually you're going to have fully autonomous systems with no steering wheel and that's require a lot of technology development. I'm going to now apply that to your question on robotics. A robot that is going to be with you in your house and is going to do everything you ask your robot to do will deal with every house is different. We'll deal with it, learn new tasks. I think that's going to require a lot of training and that's going to take a little bit of time. However, there's a massive commercial opportunity for you to get a robot and get the robot perfected on one particular task and that is actually a much easier technology problem to solve. It's much easier because you can train the robot on virtual data then you can do what we call teleoperations to get the human to do the task and then you know that's physical yo training on that and then you get the robot to do that task and we see now a number of interest applications for example when you think about an AMR robot a robot with two wheels then knows got a navigating door and two arms you can put it in healthcare You can put it in retail and retail will go in the middle of the night and stock the shelf. So you can train on that particular environment and task. I think that technology is probably now right in front of us. The technology to get your robot buddy in your house I think that's a few years away.

你知道，你考虑的是如此广泛的行业，现在 AI 加速器芯片，你们不是唯一挑战 Nvidia 的，但他们是主导者，他们是那个市场的大佬，给我一个你雄心的规模感，与现在的大两家相比，你希望这个市场几年后是什么样子？

You know that's such a wide range of industries you're thinking about and now AI accelerator chips you all are not the only ones challenging Nvidia but they're the dominant they're the big guy right in that market give me a sense of the scale of your ambition what do you want this market to look like a few years from now compared to you know the big two right now

是的，首先我将回答这个问题，看，对我们来说，你知道，边缘有巨大的机会。我认为我们在边缘 AI 流行之前就开始谈论它了。现在每个人都在谈论它，但云将继续是一个巨大的可寻址市场。所以对我们来说，我们不需要得到很多。我们只需要得到你知道的一部分，我认为有很多参与者的空间。有很多竞争的兴趣。但我现在要说的第二部分实际上更相关。

Yes so first I'll answer the question look for us you know there's a massive opportunity at the edge. I think we started talking about AI at the edge before it was popular. Now everybody's talking about it but the cloud will continue to be like a massive addressable market. So for us we don't need to get a lot. We just needed to get you know a portion of that and I think there's room for many players. There's a lot of interest in competition. But the second part of what I'm going to tell right now is actually more relevant.

数据中心的演变：当你从 AI 的创建——比如训练——走向将 AI 投入生产时，所有这些数据中心、所有这些投资，都渴望开始进行推理，因为只有这样你才能将全部投资变现。这将推动架构变革。我们一直很清楚，数据中心会进一步 disaggregate（解耦），不同任务会有不同的架构。

The evolution of the data center: when you go from creation of AI—let's say that's training—into putting AI to production, all of those data centers, all of those investments, they are eager to start doing inference because that's how you're going to monetize all of that investment. That is going to drive architecture changes. We always had clarity that you're going to further disaggregate the data center and you're going to have different architectures for different tasks.

我们在数据中心的投资：我们说过我们并不专注于 GPU。我们有 GPU。我们在所有移动设备、所有汽车芯片、PC 中都出货 GPU。但我们创造了这个神经处理单元，即 NPU，它针对推理的高计算密度和低功耗进行了优化。我认为 Nvidia 最近对 Groq 的收购验证了我们的论点：GPU 不是唯一的解决方案，你会为每件事找到不同的优化计算解决方案。

Our investment in data center: we said we're not focused on the GPU. We have a GPU. We ship GPUs in all of our mobile devices and all of our automotive silicon, PCs. But we created this neural processing unit, the NPU, which was optimized for high density of compute and low power consumption for inference. And I think the recent acquisition that Nvidia did of Groq validates our thesis that the GPU is not the only solution for this and you're going to find a different optimized computing solution for everything.

我想讲一个故事，有助于说明这个视角。当你审视当今数据中心的情况，计算需求曲线非常陡峭，而且会持续下去。但能源曲线却与之不匹配。电力不够。所以你现在必须应对——也许在很多情况下，电力将成为一种有限资源。这正是手机上发生过的情况。

And I want to tell you a story which will help put that perspective. When you look at what's happening in data centers today, you have this very steep curve of compute needs which will continue. But then you have the curve on energy and they do not align. There's not enough power. So you have now to deal with—maybe in many cases power is going to be a finite resource. This is what happened on phones.

这是一个非常有趣的工程挑战，因为一代又一代，如果你看看今天的智能手机，再看看疫情时期的智能手机，计算能力已经大幅提升——每年增长 10%、20%，计算量不断增加。但手机上发生了一件非常有趣的事：外形尺寸不变，它必须能放进口袋。无论我们塞进多少计算能力——比如手机上的 GPU 做光线追踪等等——电池必须能用一整天，电池尺寸不变，它不能发热。我无法对它进行液冷，而且你会把它贴在脸上，它不能发烫。这就迫使像我们这样在这个领域的公司去解耦计算。

It's a very interesting engineering challenge because generation over generation, if you actually look at a smartphone today and look at the smartphone just like at the time of the pandemic, the computing power has increased dramatically—it goes 10, 20% every year, increasing the amount of compute. But there's something very interesting that happened in phones: the form factor size is the same, it has to be in your pocket. No matter how much more compute we're putting in—like our GPUs on phones do ray tracing and all those things—the battery has to last all day, the battery is the same size, it cannot get hot. I cannot liquid cool that thing and you're going to touch your face with it, it cannot be hot. So then it forces companies that have been in this space like us to disaggregate the compute.

一个例子是——回顾 PC 时代，当我们都用笔记本电脑，用 iTunes 和 MP3 播放器时，PC 会在 CPU 上解码 MP3 文件。就是这样，PC 是以 CPU 为中心的。我们意识到在移动端，当你在手机上播放音乐时，你不能那样做。它耗电太多。你开发了单独的硬件来做 MP3 解码。拍照时你有单独的硬件来做 JPEG 编码。你有点在解耦计算。

An example of this is when you look at—if you go back to PCs, when we all had our laptops and were doing iTunes and doing MP3 players, PCs will decode an MP3 file on the CPU. That's it, PCs are CPU-centric. We realized in mobile when you start playing music on a phone, you cannot do that. It burns so much power. You develop a separate hardware to do MP3 decode. You have a separate hardware to JPEG encode when you take a picture. You kind of disaggregate the compute.

我们现在开始看到数据中心正在发生同样的事情，因为你有这个——这是计算与能源的问题——我们开始看到同样的事情以不同的规模发生。你开始看到推理被解耦。我需要一个用于预填充的架构。我需要一个用于解码的架构。我认为这就是我们的机会。

What we're starting to see happening in the data center right now, because you have this—this is the compute-energy issue—we're starting to see the same thing at a different scale. You're starting to see the inference being disaggregated. I need an architecture for prefill. I need an architecture to decode. And I think that's the opportunity for us.

所以我们现在的雄心是——我们非常兴奋能进入这个领域，从而分得一杯羹。这对高通来说意义重大。但我认为机会在于进一步解耦，你会寻找的不仅是竞争，还有更高密度和低功耗的解决方案。我认为这正符合我们的 DNA，我认为这对公司来说是一个巨大的机会。

So our ambition right now is—we're very excited to enter the space so we can get a piece of it. It's going to be significant for Qualcomm. But I think the opportunity is to further disaggregate and you're going to look for not only competition but solutions that have higher density and low power. I think that's right in our DNA and I think that's a massive opportunity for the company.

达沃斯是一个独特的地方，你能把来自如此多领域和行业的人聚集在一起，对于像您这样拥有全球公司的人来说，我能想象来到这里特别有用。另一面是，你会看到很多世界上正在发生的事情，以及你所涉及的众多供应链、技术和行业。当您环顾全球图景、地缘政治图景时，在思考公司的未来时，最让您担忧的是什么？

Davos is such a unique place where you can bring together people from so many fields and industries and for someone with a global company like yours, I can imagine it's particularly useful to be here. The flip side of that is you see a lot about what's happening in the world and with so many supply chains, technologies, industries you're involved in. When you look around at the global picture, the geopolitical picture, what worries you the most when you're thinking about the future of your company?

看，我认为现在每家半导体公司——我认为这是真的——如果你看看美国半导体行业，所有非常成功的半导体公司都是了不起的公司。我认为我们都需要一个非常有韧性且强大的半导体供应链。我认为我们都严重依赖制造体系。我们的大部分芯片由台湾的台积电制造。其中一些芯片也由三星制造，三星历史上一直是我们的供应商。所以我们都需要的是半导体制造的能力和可用性。在这方面，我们就像小政府：我们想要一个强大、有韧性的供应链。这对我们非常重要。

Look, I think every semiconductor company right now—and I think it's true—if you look at the United States semiconductor industry, all of the very successful semiconductor companies are fabulous companies. And I think we all need a very resilient and strong semiconductor supply chain. I think we all have strong dependencies on the manufacturing system. Most of our chips are made by TSMC in Taiwan. Some of those chips are also made by Samsung, which has been our supplier historically. So the thing that we all need is capacity and availability of semiconductor manufacturing. And in that way we're like mini governments: we want a strong, resilient supply chain. That's very important for us.

对我们非常重要的另一件事是美国的创新经济。它需要研发规模——研发规模非常重要。也许我们在某个方面有点独特。我认为高通一直——凭借我们对标准的贡献、我们的长期研发——我们一直是一家研发密集型公司。举个例子，2025 年，顺便说一下，在所有美国公司中——甚至规模远大于高通的公司——我们的专利申请量排名第一。但你需要规模。要有规模，你需要确保你的技术栈是全球性的。如果你最终面对的是一个小得多的市场，那就有问题。所以我认为另一件总是让我们担忧的事是如何保持我们技术的全球规模；这也是重要的事情。

The other thing that is very important for us is the innovation economy of the United States. It requires scale of R&D—scale of R&D is very important. And maybe we are kind of unique in a particular way. I think Qualcomm has always been—with our contribution to standards, with our long-term R&D—we have been an R&D-heavy company. As an example, in 2025, just to send it off, all American companies—even companies significantly higher in size than Qualcomm—were number one in terms of patent applications. But you need scale. For you to have scale, you need to make sure that your technology stack is global. If you end up with a much smaller market, that's a problem. So I think the other thing that always worries us is how to maintain global scale of our technology; that is an important thing as well.

直接相关的是，《芯片法案》起作用了吗？现在说还为时过早吗？

And directly related to that, did the CHIPS Act work? Is it too early to say?

看，它起作用了，但——我觉得当你问这个问题时很有趣，问它是否真的起作用。我的答案是它正在起作用。这需要时间。我认为这是朝着正确方向迈出的一大步。我们非常高兴有地理位置和产能的多样性。我们一直是它的坚定支持者，我们将使用在美国、日本和其他地方建设的产能。但当你真正看大局时，因为半导体将无处不在——每个行业都将有大量的计算和连接——如果你只看 10 年内，全球先进半导体的制造能力将需要翻一番。所以我们还有很长的路要走。我认为这是朝着正确方向迈出的极好一步，我们需要继续建设更多产能。

See, it worked but—I think it's interesting when you ask this question, asking did it really work. My answer is it is working. It takes time. I think it's a big step in the right direction. We're very happy that there's diversity of geographical location and capacity. We have been big proponents of it, where we are going to be using the capacity being built in the United States and Japan and other places. But when you actually look at the bigger picture, because semiconductor is going to everything—everything is going to have a lot of computing and connectivity going to every industry—if you look just within 10 years, we're going to have to double the manufacturing capability of advanced semiconductors globally. So we have a long way to go. I think it's an excellent step in the right direction and we need to keep building more capacity.

坊间有传闻说，台积电在亚利桑那的经理来自台湾，他们会说，这些不是我们习惯共事的人——文化标准、工作标准都不同。我的意思是，你觉得美国能在同一水平上竞争吗？我们能否重现那种在最前沿——最新、最好的芯片——并在这里大规模生产？还是说答案似乎是“我们还不知道”，或者你不同意这个前提？

Anecdotally you'll hear stories, you know, TSMC in Arizona managers come from Taiwan and they're like, these aren't the people we're used to working with—different cultural standards, work standards. I mean, do you think the US can compete on the same level? Can we recreate that right on the edge—the newest, the best chips—and build them here at scale? Or because that answer seems to be we don't know yet, or would you disagree with that premise?

不，我完全不同意这个前提。听着，毫无疑问，你会在美国找到人才。而且我认为美国也一直是吸引人才的好地方。但这些事情在短期内是行不通的。你必须建立一个生态系统。所以这需要时间。你不能说我们之前没做半导体制造，你建了一个工厂，就指望第一天就有一支能实际操作它的劳动力队伍。但它是会随着时间建立起来的。我打个比方——举个例子吧。有很多电子制造业——现在电子制造业，我认为历史上大部分电子制造都在亚洲。你看到 80 年代和 90 年代日本有很多制造业，然后你看到台湾制造业的发展，你看到韩国，然后今天是中国，而且它还在不断转移。如果你看大约 8 年前，有很多公司开始在越南制造电子产品。当然，一开始你没有生态系统——很多组件会从中国运来。现在达到规模后，你开始看到组件的生态系统。现在，例如，越南是一个制造业中心。你现在也看到一些电子制造商去印度。你看到一些公司说我们要在印度制造手机，然后你开始——它还不那么成熟，但你开始发展一个生态系统。随着你在台积电亚利桑那、三星泰勒，以及像美光这样已经存在的制造企业继续建设产能，你将开始发展一个生态系统——不仅是熟练劳动力的可用性，还有为该地区周围发展的晶圆厂提供材料的供应商。所以这需要时间。重要的是要有政策方向，并准备长期投资。

No, I completely disagree with that premise. Look, there's no question—you're going to find talent in the United States. And I think the United States has also been a great place to attract talent. But those things don't work in the very short term. You have to build an ecosystem. So it takes time. You're not going to say we were not doing semiconductor manufacturing, you built a facility, and you expect day one you're going to have a workforce that is actually capable to operate in that. But it's going to get built over time. I'm going to do a parallel—let's just give an example. There was a lot of electronics manufacturing—right now electronics manufacturing, I think historically the majority of the manufacturing of electronics has been in Asia. You saw a lot of manufacturing in Japan in the 80s and 90s, then you saw the development of manufacturing in Taiwan, you saw in Korea, then it's China today, and it keeps moving. If you look about 8 years ago, there were a lot of companies starting manufacturing electronics in Vietnam. Of course, in the very beginning you didn't have an ecosystem—a lot of the components would come in from China. Now at scale, you started to see an ecosystem of components. Now, for example, Vietnam is a hub for manufacturing. You also see now, for example, some electronics manufacturers going to India. You see some of the companies say we're going to make phones in India, and then you start—it's not as mature, but you start developing an ecosystem. As you continue to build capacity on TSMC Arizona, on Samsung Taylor, on things that already existed like Micron that does manufacturing, you're going to start developing an ecosystem—not only the availability of skilled labor but suppliers of materials into the foundries that are going to be developing around that region. So it just takes time. Important in those things is to have direction of policy and be prepared to be investing for the long term.

你说需要时间，能否让我们更清楚地了解规模有多大？因为对很多人来说，当他们考虑去风险化时，看到现在台湾海峡发生的事情，完全不可预测，但我们谈论的是几十年、几个月、几年？我的意思是，我知道这有点难以量化，但你怎么看待这一点与那个地区的风险相比？

And when you say it takes time, could you give us a better sense of how much scale? Because for a lot of people when they think about de-risking, you see what's happening in the Taiwan Strait now and it's completely unpredictable, but are we talking in terms of decades, months, years? I mean, I know it's somewhat intangible, but how do you think about that versus the risks in that part of the world?

听着，我认为如果你算一下——我们只是从更高层面来说，我不会很精确，只是数量级，对吧？所以如果你看所有已宣布的项目和实际正在建设的东西，例如台积电或三星在美国的产能，你知道，那些现有的已宣布项目达到约台积电在台湾产能的 10%，然后最终会从 10% 到 20% 和 30%。我认为这些是令人难以置信的进展，但你谈论的真的是以五年为增量的视野，也就是大约 5 到 10 年。它不会在 12 个月内发生，对吧？所以这就是时间范围。

Look, I think if you do the math—let's just go higher level, I won't be very precise, just level of magnitude, right? So if you look at all of the announced programs and what is actually being built, for example, capacity from TSMC or Samsung in the United States, you know, those existing announced programs get to about 10% of the capacity that exists in TSMC in Taiwan, and then eventually it's going to go from 10% to 20% and 30%. And those are incredible, I think, progress, but you're really talking in horizons of increments of five years, which is like a 5 to 10 years. It does not happen in 12 months, right? So that's kind of the time horizon.

你知道我们是《华盛顿邮报》，不幸的是在华盛顿，很多人当他们谈论——

And you know we are the Washington Post and unfortunately in Washington a lot of people when they talk about—

嗯，我们来自圣地亚哥。

Well we're from San Diego.

嗯，我自己就是加州人。所以这就是为什么我们在大部分对话中我更关注未来。但在华盛顿，似乎每次对话都会回到与中国的竞争，大问题是出口管制。对吧。稍微谈谈我们讨论的内容，几乎有一种哲学分歧，我相信你在过去几天里有过很多这样的对话。我第一次听说——

Well I'm a Californian myself. So that's why I focused more on the future for most of our conversation. But in Washington it seems like every conversation comes back to competition with China and the big questions are export controls. Right. A little bit about what we discussed and there's almost a philosophical divide and you I'm sure you've had a lot of these conversations over the past few days. First I heard of—

嗯，我很高兴我们在结束前谈到了这个。

Well I'm glad we got it before the end.

对，对，对，对，对，对，对，对，对，对，对，对，对，对，对，对，对，就在你即将回到阳光明媚的圣地亚哥之前。嗯，在美国存在一种分歧：我们是试图超越中国，还是试图拖慢他们？也许并不总是这种二元对立，但你认为在我们仍在建设这些生态系统的时候，向他们出售我们最先进的芯片是个好主意吗？或者如果不是，任何特定的芯片或任何特定的政策。请带我们了解你如何看待来自中国的竞争以及对像你们这样的美国公司的挑战。

Right, right, right, right, right, right, right, right, right, right, right, right, right, right, right, right, right before you're about to head back to sunny San Diego. Um there is a divide, you know, in America: do we try to outcompete China or do we try to slow them down? And maybe it's not always a binary that way, but do you think that selling our most advanced chips to them is a good idea at this point as we're still building out these ecosystems? Or if not, any particular chip or any particular policy. Take us through how you think about the competition from China and the challenge to American firms like yours.

好的。所以这是个有趣的问题，也许我会提供高通的视角，对吧？因为你可能经常听到——我认为你在行业中经常看到，如果你看某些行业，它们一直在应对大量竞争，例如美国公司和中国公司之间，你看到竞争尤其发生在拉丁美洲、东南亚、中东、非洲、东欧等市场。你看到那些地方有来自中国的全套公司 and 能力。你看到来自欧洲的竞争——最好的例子，如果你看现在的汽车行业，我们非常幸运能为所有汽车公司提供芯片，实际上所有中国公司都在他们的电动汽车中使用我们的芯片，你看到那些市场上的竞争，欧洲汽车制造商、日本汽车制造商、美国、中国。而且经常你谈论一些中国竞争对手的速度——他们谈论中国速度。所以有趣的是，回顾移动行业，在从功能手机到智能手机的转变中，很多中国公司进入了手机市场，他们变得相当成功，我认为与苹果、三星等公司并驾齐驱。你看到像小米这样的公司，移动市场一直是一个超级竞争的环境——我们也是竞争对手——有趣的是我们活下来了。如果你真的去中国，看看今天的中国国内市场，你看到很多中国手机使用骁龙。然后他们非常有竞争力地进入了汽车领域。

Okay. So it's an interesting question and maybe I will provide a Qualcomm perspective on it, right? Because you probably hear often—I think you often in the industry, if you look at certain industries that have been dealing with a lot of competition between, for example, American companies and Chinese companies, you see like the competition happens especially in markets such as Latin America, Southeast Asia, Middle East, Africa, Eastern Europe. You see that in those places have the full stack of companies and capabilities from China. You see that the competition from Europe—competition like the best example of this, if you look at the automotive industry right now, and we're very fortunate to be providing chips to all of the automotive companies where actually all the Chinese companies use our chips into their EVs, and you see the competition in those markets across the European automakers, the Japanese automakers, the American, the Chinese. And often you talk about the speed of some of those Chinese competitors—they talk about China speed. So the interesting thing is go back in the mobile industry and there was a transition from feature phone to smartphone, a lot of Chinese companies entered the phone market and they became quite successful, I think, alongside companies like Apple, Samsung. You see companies like Xiaomi, and the mobile market has been a super competitive environment—like we also competitors—and the interesting thing is we survived. If you actually go to China and if you look at the China domestic market today, you see that a lot of the China phones use Snapdragon. If then they went very competitive to cars.

他们用我们的芯片，你知道，其中一些——我们说的不是数据中心芯片，而是消费电子芯片，他们是有替代品的——来自中国、来自台湾的头部替代品。而我们活了下来。我觉得实际发生的是，我们学会了如何在中国保持竞争力。我们一直认为，竞争永远都会存在，但如果我们不仅在本土有竞争力，在中国也有竞争力，并且能在中国市场取胜，那我们就知道自己具备全球竞争力。所以某种程度上，我其实认为我们是受益的。

They use our chips, and, you know, some of those — we're not talking about data center chips. We're talking about consumer electronics chips, which they have alternatives — the head alternatives from China, from Taiwan. And we survived. And I think what happened is we learned how to be competitive in China. And I think we always thought, you know, there's always going to be competition, but if we can be competitive not only at home but also be competitive in China, and we win in the marketplace in China, then we know we're globally competitive. So to some extent, I actually think we have benefited.

放到全球技术栈的竞争里看，因为竞争其实是件大好事。它教会我们如何始终保持竞争力，如何不断重塑自己，不变得自满——尤其是在移动业务上。你知道，我常说移动业务就像角斗士的生意，因为我们每年都得做出一颗新芯片。今年有一颗很棒的芯片，说明不了什么，因为明年还得再做一颗。就像角斗士：你走进斗兽场，有三种可能的结果——赢、输、同归于尽。而如果你赢了，明年你还得再进斗兽场。这就是我们看待与中国竞争的方式。

For the global competition of technology stacks, because, you know, competition is a great thing, actually. It taught us how to always remain competitive, to always reinvent ourselves, not to become complacent — especially in the mobile business. You know, I often say the mobile business is like a gladiator business, because we have to do a new chip every single year. If we have a great chip this year, it means nothing, because next year we have to do another one. It's like the gladiator: you go to the Colosseum, three possible outcomes — win, lose, both lose. And then if you win, next year you have to go to the Colosseum again. So that's how we think about competition with China.

而且你看，竞争永远都会存在。美国和中国——这会是一段很有意思的关系，因为两国经济是相互连接的。有些依赖关系会继续存在，某些领域也会出现一定程度的脱钩。我认为这就是现实。企业需要去适应，而我们要做的就是学会在这个格局里保持竞争力。最后我想说，我们开始让公司多元化。我们越多元化，拿到的中国业务就越多。我想，也许未来就是这样。

And look, there's always going to be competition. US and China — it's going to be an interesting relationship, because the economies are connected. There are some dependencies that will continue to be there. There's going to be some decoupling in some areas. I think that's the reality. Companies need to adapt, and I think we're just going to have to learn to remain competitive in the space. And I'll just finish by saying we started diversifying the company. The more we diversify, the more China business we got. And I think that maybe that's what's going to be.

这是一个很鼓舞人心的收尾。可惜我们时间快到了，所以得收尾了。Cristiano Amon，感谢你今天来到这里，我们非常感激。

Well, that's an inspiring note to end on. Unfortunately, we are just about out of time, so we'll have to wrap it up. Cristiano Amon, thank you for being here today. We really appreciate it.