吴婷校长硅谷演讲：AI is Global, Reality is Local

在AI产业化元年的开端，吴婷校长代表嘉宾商学以及嘉宾生态的企业家校友们，在世界创新前沿的硅谷进行了一场题为AI is Global, Reality is Local的英文演讲。
当地时间2026年1月10日，在美国计算机历史博物馆，吴婷校长应邀出席"Smart Living In the Era of AI"主题研讨会，同台嘉宾包括加州财政部部长Fiona Ma、美国计算机科学家Peter Norvig等。
在演讲中，吴婷校长以嘉宾校友企业Genspark、地平线、擎朗、比亚迪等为例，分析AI产业图景与致胜关键，对比硅谷“雨林模式”与中国“草原模式”，并欢迎全球科技从业者来到嘉宾商学，学习中国、感受中国、寻找中国产业机会。
以下是英文演讲全文及中文翻译，分享给大家。

Hello everyone, I am Wu Ting, the founder of Jiabin Business School, the only business school in China focused solely on case studies.

I established Jiabin 10 years ago. That was the era of "Mass Entrepreneurship and Innovation" in China. Everyone with a dream was trying to change the world through business. Therefore, countless fresh and vivid business cases were emerging every day.

I studied at a traditional business school before. I found that the cases in classrooms were often old and closed off. You couldn't dig deeper with questions, you couldn't see the details, you couldn't verify if they were true, and you couldn't feel the atmosphere or the energy. And the biggest thing is, you couldn't take students to study cutting-edge cases in such a fast-changing era.

So, after spending a decade in the media industry, back in 2016, I decided to switch career paths and started my own business school, focusing entirely on a new way of learning cases.

We created a new form of online case teaching. From 2016 to present, we have released thousands of video business cases, attracting 4 million high-quality followers across the web.

We also pioneered a new model for offline learning: In-depth On-site Studies. We take entrepreneurs "To the Edge, To the Core, To the Truth". This is indeed a highly innovative and unique way of learning.

I have led over 600 top entrepreneurial alumni from various industries, including AI, the Internet, new energy vehicles, manufacturing, biotech, and retail, to visit hundreds of global leading companies and learn directly from their founders.

Nearly every quarter, I join an alumni entrepreneur in ringing the opening bell at a major stock exchange. You should know that the Chinese capital market uses a strict approval-based IPO system. To get listed, your revenue and profit scale must meet high standards. It is not easy. The frequency is actually quite astonishing.

We have been to events like CES many times. I remember the first time was in 2017. Back then, new energy vehicles were just sprouting, and Nvidia's market cap was less than 100 billion dollars. Today, it is around 5 trillion dollars.

Now, in the hottest AI industry, China and the US are competing intensely. Jensen Huang summarized the AI industry as a "5-Layer Cake". From bottom to top: Energy, Chips, Infrastructure, Models, Applications.

The bottom four layers require massive capital or manufacturing tech—that is a game for the giants.

To my mind, for most of the companies, the biggest opportunity lies in the Application Layer, and we saw exactly this trend at CES, including many of our alumni enterprises.

Now, the key point with the application layer is this: The global air force is AI, but it's not enough. To win in applications, you just need to ground it in reality—that's your localized ground force.

This is the key message I want to share with you today: AI is Global, Reality is Local.

1、AI is Global

Let me quickly break down why AI is Global. By AI here, I mainly mean Large Language Models (LLMs).

Although there are thousands of languages in the world, and Chinese and English grammar differ, for the self-learning of LLMs, it's just a difference in input data. The underlying technology is universal. And our goal in using large models is the same: to improve information processing efficiency and lower the barrier to acquiring knowledge.

So we see ChatGPT has 800 million monthly active users globally, and Gemini has 350 million. In January 2025, Chinese DeepSeek stunned the world, quickly becoming the No. 1 downloaded app on the App Store. Countless professionals called it the "DeepSeek Moment". Later, Qwen, Kimi, and MiniMax also went global. They used open-source models to export China's AI power to the world.

Our Jiabin alumni, Eric Jing, went from China to Silicon Valley and founded Genspark. In just 18 months, it became a unicorn. Their most popular product Super Agent launched in April 2025, reached 10 million dollars in Annual Recurring Revenue (ARR) in just 9 days. In 45 days, it soared to 36 million dollars, setting one of the fastest growth records in the AI industry.

AI is global. It is like the air force—surveying the earth from above, and empowering us with cross-border insights and opportunities.

2、Reality is Local

But that is only the first half of the business story.

AGI lowers the cost of starting, but not the cost of finishing.

Once you get into real-world applications, you'll discover the true importance of "going local".

Now, going back 80 years to 1945, the famous economist Friedrich Hayek wrote an essay called The Use of Knowledge in Society. One of his most profound points was:

• Scientific knowledge occupies now so prominent a place in public imagination that we tend to forget that it is not the only kind that is relevant......We need to remember how valuable an asset in all walks of life is knowledge of people, of local conditions, and special circumstances.

Later, scholars summarized this as "on-site knowledge". It is spread out in the minds of every market participant, not just in the mind of a central planner.

This insight has new meaning in the AGI era.

Even if AI has global intelligence and a global perspective, when it lands in applications, AI lacks local context, and "on-site knowledge" is changing every moment.

In a digital sandbox, AI is like a god; but in the real world, sometimes it acts like a fool. To avoid being a fool, AI must step out of the sandbox, go deep into reality, and use "Local" to make "Reality" smarter.

Take Horizon Robotics, one of our Jiabin alumni companies.It is the first stock in the auto driving sector in China and the largest provider of ADAS (Advanced Driver Assistance Systems) solutions in China.

Horizon Robotics released its first chip back in 2017. At the time, Mobileye was at its peak, and Nvidia was investing heavily in this field. The field was already crowded with global giants.

How did Horizon break through in "AI Reality"? By using the power of Local.

At that time, giants like Mobileye won by sheer computing power. They sold "Global" products. Their underlying algorithms were completely closed. Adaptation was the customer's own problem. They didn't get involved.

Horizon did the opposite. They opened their underlying algorithms and sent teams of dozens of engineers to work alongside customers, polishing the product together to ensure it met "Local" requirements. In the words of Dr. Yu Kai, founder of Horizon Robotics, this is "standing outside the range of the giants".

Everyone here today is an entrepreneur or a founder. You know clearly that customer needs are always Local. There is no "one-size-fits-all" need. The AGI era is no different.

Back to Horizon. It's not just about service. In hard-core technical competition, it also relied on "Local" to dig its moat deep.

Horizon's moat is its extreme understanding of Chinese roads: weaving riders, low-speed microcars, complex interchanges, and aggressive city driving.

Horizon used this massive, non-standard, extremely "Local" road data for training. Finally, it turned its system into an "experienced driver" that knows Chinese roads best.

In contrast, Tesla's FSD, which works incredibly well in the US, faced unique challenges during its China tests in February 2025—running red lights, driving the wrong way, and crossing solid lines. This has nothing to do with computing power or algorithms; it is only about "Local".

This does not, however, mean that Local is closed. Global companies can still share the "Reality" cake through Local-style cooperation and innovation.

Take, for example, Volkswagen Group, a giant of the fuel-powered era, which counts China as its largest market. But China's shift to EVs and smart driving was far faster than Germany's. In 2023, they chose to invest in Horizon Robotics to jointly develop smart solutions specifically for the Chinese market.

There is also Keenon Robotics, another alumni company. It is the world's largest commercial service robot company. In Japan, where labor is short, Keenon focuses on dining delivery robots, replacing waiters at 1/5th to 1/3rd of the cost. In Europe and the US, they launched lawn mowing robots, making mowing easy.

In the Jiabin ecosystem, there are many robot companies. For example, UBTech, the world's first humanoid robot stock; ENGINEAI, the first in the world to achieve a backflip; DEEP Robotics, a pioneer in commercial robot dogs, and Sanhua Intelligent Controls, which supplies joints for Tesla's Optimus.

In every niche track, they use a local mindset to polish products and a global vision to open markets.

The more global technology becomes, the more local reality matters.

This is glocalization.

3、Rainforest & Savanna

The tech innovation models of China and the US are also a weave of Global and Local.

US tech innovation is centered in Silicon Valley. The book The Rainforest describes this as the "Rainforest Model"—a highly connected ecosystem where people interact constantly, and new ideas are born every day.

On the other hand, China possesses the world's most complete industrial system. Its industrial GDP accounts for over 23% of the global total—exceeding the US, Japan, and Germany combined. China's tech innovation doesn't rely on a single highland; it is spread across various regions.

Instead, it's more like a vast Savanna—one that's teeming with all kinds of players, like lions, zebras, wildebeest, and buffalo. Each has its own role. I call this the "Savanna Model".

Let me focus on the lions here. The male lion sets the direction and protects the territory, representing China's unique "ecosystem leaders".

Take BYD, a Jiabin alumni company, and I am the brand consultant. It is the King of New Energy Vehicles in the world. With nearly 1,300 suppliers and investments in over 100 external companies, they form a massive Lion Pride.

BYD leads the charge in opening global markets, while suppliers deliver parts and solve specific problems.

Giants like Huawei, Midea have all assembled their own Lion Prides.

The Rainforest is lush, the Savanna is fierce.

The rainforest creates new ideas (0 to 1). The savanna scales them up (1 to 100). Together, they make up the real business world.

Neither is better. They just fit different stages and local realities. Top global companies know how to get inspiration from the rainforest, and how to use the savanna's strength to win new markets.

Jiabin Business School is a global business school. We have global mentors, global students, global offices, and a global vision. Yet we also focus heavily on localization. Only local learning can truly help everyone feel the world as it is, and thus make the best decisions and applications.

Many Chinese Americans and American friends are here with us today. I believe you all have a good understanding of the opportunities and challenges that China is facing right now. At such a time, if you wish to come to China to learn more, explore more, and seize more opportunities, Jiabin Business School will be your best choice.

Thank you.

中文版

大家好，我是嘉宾商学创办人吴婷。我在10年前创办了这所中国唯一的全案例教学机构——那是在中国的双创时代，“大众创业、万众创新”，每个心怀梦想的人都在尝试用商业的力量改变世界，因此，无数的鲜活案例每天都在涌现。

在那之前，我曾经在传统的商学院里学习过，我发现，教室里学习的案例是陈旧和封闭的，你无法追问主角，你无法获知细节，你无法验证真伪，你无法感受现场的氛围与能量。最重要的是，你无法快速带领学员获取时代前沿的案例。

于是，从事了十年媒体的我，在2016年决定转变身份，创立了一所商学院，专注于全新的案例学习。

我们开创了线上案例教学的新形态，从2016年至今，累计发布了数千个视频商业案例，吸引了全网400万高质量粉丝。

我们更是开创了线下案例教学新范式——近场深度访学。我们带企业家们“行至最前沿、访到最深处、学到最真实”。这确实是一种极具创新且独特的学习方式。

10年里，我累计带领600多位各行业头部的企业家校友——包括AI、互联网、新能源汽车、高端制造、生物医药、消费零售等领域——到数百个全球的顶尖公司，向创始人和CEO学习。

几乎平均每个季度，我都会陪一位校友敲钟。朋友们，中国资本市场实行严格的核准制，想上市，营收和利润规模必须达到要求，这并不是一件容易的事。所以我们校友的敲钟频率，是高得惊人的。

CES这样的盛会，我们来了很多次。记得第一次是在2017年，那时新能源汽车产业才还是萌芽状态，英伟达市值还不到1000亿美元，如今已经到5万亿美元级别了。

在当下最火热的AI产业，中美也在激烈竞逐。黄仁勋将AI产业概括为“五层蛋糕”，从下到上依次是——能源、芯片、基础设施、模型和应用。下面四层，都对资金或制造技术有严苛要求，是巨头们的游戏。

在我看来，对于大多数公司而言，最大的机会在于应用层，我们在CES上也确切地看到了这一趋势，其中就包括我们许多校友企业。

关于应用层，关键点在于：AI 是全球化的空军，但这还不够。要在应用层面致胜，你必须将其扎根于现实——这就是你的在地化陆军。

这就是我今天想分享的核心信息：AI是全球的，现实是在地的。

1、AI是全球的

我先简单阐述为什么AI是全球的。这里的AI，主要指大语言模型。尽管全世界有数千种语言，中文和英文在语法上不同，但对于大语言模型的自我学习来说，无非是输入的语料不同，但底层的技术是全球通用的。并且，我们使用大模型的目的是一致的——提升信息处理效率、降低知识获取门槛。

于是我们看到，ChatGPT全球月活8亿、Gemini 3.5亿；2025年1月，DeepSeek 惊艳世界，迅速成为App Store下载量第一名，被称为DeepSeek时刻；后来Qwen、Kimi、MiniMax也都走向世界，他们用开源模式，向世界输出中国AI力量。

我们嘉宾商学校友景鲲，从中国来到硅谷，创立了Genspark，只用了18个月就长成了独角兽。他们最受欢迎的产品，是2025年4月推出新一代AI搜索工具Super Agent，上线9天就实现1000万美元年度经常性收入(ARR)，45天飙升至3600万美元，创下AI行业增长最快纪录之一。

AI是全球化的。它就像空军——从高空俯瞰地球，赋予我们跨越国界的洞察力与机遇。

2、现实是在地的

但这只是商业故事的上半场。AGI让开始变得容易，但完成则不然。

一旦进入现实的应用，你又会发现在地化的重要性。

80年前（1945年），知名经济学家哈耶克发表了一篇文章叫《知识在社会中的运用》，其中最深刻的观点是：

“科学知识在公众想象中占据了至高无上的地位，以至于人们往往忘记，它不是唯一重要的知识……在各行各业中，关于人的知识，关于当地环境的知识，关于特殊情况的知识，是多么宝贵的财富。”

后来，学界将之总结为“现场知识”，它分散在每个市场参与者的头脑里，而不仅在一个中央计划者的头脑里。

这一洞见，在AGI时代有了新的意义。即便AI具备全球化的智力、全局视角，但在落地应用中，AI又缺乏在地化的语境，现场知识也时刻在变。在数字沙盘中神一般的AI，在现实世界里有时却像一个傻子。

不想变傻，就要走出沙盘，深入现实，用在地让AI更智能。

嘉宾商学的校友企业地平线机器人，是中国智驾第一股，也是中国最大的智驾解决方案提供商。

地平线发布第一代智驾芯片是在2017年，当时Mobileye如日中天，英伟达也把这个领域视为第二曲线，重金投入。这个领域已经挤满了全球巨头。地平线是如何在AI现实突围的？靠“在地”的力量。

当时Mobileye等国际巨头主要靠大算力取胜，只卖Global的产品，底层算法完全封闭，落地适配是客户自己的事，他不参与。而地平线开放底层算法，还派出数十人的团队，与客户一起打磨，来确保达到在地化的要求。用地平线机器人的创始人余凯的话说，这是“站在巨头的射程之外”。

今天在座的都是企业家、创业者，你们一定清楚，客户需求都是本地的，没有千篇一律的需求，AGI时代也一样。

讲回地平线，不只是服务客户，在真刀真枪的技术竞争上，它同样是依靠在地化，才挖深了护城河。

地平线的护城河，是对中国道路场景的极致理解：穿梭的外卖骑手、低速的老年代步车、复杂的立体交通、一些人彪悍的驾驶风格。地平线用这些海量的、非标的、极度本地的道路数据进行训练，最终，把自己打造成了一个最懂中国路况的“老司机”。

相比之下，在美国表现极好的FSD，在2025年2月进入中国测试时，接连发生闯红灯、逆行、压线等等交通违规事件。这无关算力的大小、算法的优劣，只关于在地化。

当然，在地并不等于封闭，全球企业通过在地式的合作和创新，依然可以分享现实的蛋糕。

比如德国大众集团，燃油时代的巨头，它的第一大市场是中国。但中国的电动化、智能化发展速度远超德国。2023年，它选择入股地平线，针对中国市场，共同开发智能化方案。

还有嘉宾商学校友企业擎朗智能，它是全球出货量最大的商用机器人企业。在日本，劳动力严重短缺，它主打送餐机器人，用1/5-1/3的成本替代服务员；在欧美，它推出割草机器人，让割草变得轻松简单。

在嘉宾商学的生态系统中，有很多机器人企业。比如全球具身机器人第一股优必选、全球第一个实现机器人前空翻的众擎、机器狗商业化先锋的云深处、为特斯拉Optimus供应关节的三花智控等等。

他们在各个细分赛道，用在地的心态打磨产品，用全球的格局开拓市场。

技术越是链接全球，现实越要扎根在地。

这就是全球在地化。

3、雨林模式（硅谷）与草原模式（中国）

中美两国的科技创新模式，恰好也是全球化和在地化的交织。

美国的科技创新集中于硅谷，《硅谷生态圈》将其称为“雨林模式”，整个生态系统高度连接、密集互动，不断产生新物种。

而中国有全球最完整的工业体系，工业GDP占全球总量的23%以上，超过美、日、德等国之和。中国的科技创新并不依赖某处高地，而是分散在各个区域。

它更像是一片广袤的草原，上面有丰富多样的种群——狮子、鬣狗、角马、水牛等。它们有各自的生态位。我称之为“草原模式”。

这里重点讲一下草原上的狮子。它们结成一个个狮群，雄狮领导方向、保护领地，雌狮配合作战、外出狩猎。它们共同生存、进化。狮群里的雄狮，代表了中国特有的产业链“链主企业”。

比如由我担任品牌顾问的校友企业比亚迪，它是全球新能源汽车之王。有近1300家供应商，还对外投资了100多家企业，他们组成一个庞大的狮群，比亚迪带头在全球开拓市场，供应商负责提供各类零部件、解决各类难题。

还有华为、小米、美的等等巨头，都集结起了各自的狮群。

雨林繁茂，狮群凶猛。雨林孕育了很多“从0到1”的奇思妙想，狮群发动了很多“从1到100”的饱和攻击，它们共同构成了商业世界的真实版图。

两者没有高下之分，只是合适不同的阶段、不同的 Local Realities。而顶级的 Global Players，既会在雨林中找灵感，又能带领狮群打江山。

嘉宾商学是一所全球商学院。我们拥有全球的导师、全球的学员、全球的办公室以及全球的视野。 但同时，我们也极度重视在地化。 只有在地学习，才能真正帮助大家感知世界的本来面目，从而做出最好的决策和应用。

今天，许多华裔朋友和美国朋友也来到了现场。我相信，大家对中国当下所面临的机遇与挑战都有很好的理解。在这个时刻，如果您希望来中国学到更多、探索更多、把握更多机会，嘉宾商学将是您的最佳选择。

谢谢大家。

演讲 | 吴婷