Lin Xiumin: Semiconductors See the World》What is Germany’s dilemma in developing the AI ​​industry?

At the 2026 Berlin Fireside Chat event held in Berlin, Germany, on October 26, Huida CEO Huang Renxun and German Chancellor Friedrich Merz talked about the future of Germany's development of AI. Huang pointed out that "physical AI starts a new industrial revolution, and there is no second chance." So what is the dilemma of Germany's development of AI?

First of all, Huang Renxun reviewed Huida’s deep relationship with German industry. As early as 2000, NVIDIA mainly focused on the field of PC games. It was through cooperation with the German graphics card manufacturer ELSA that Huida successfully entered the workstation market, and thus started all-round cooperation with major German automobile giants.

Huang Renxun pointed out that the software computing revolution in the past 50 years has reshaped the global industrial structure, but the next wave of AI is "Physical AI" - that is, the deep integration of artificial intelligence with the physical world, robotics, smart manufacturing, and heavy industry.

"All the digital technology in the world cannot replace electromechanical physics technology, and electrons will never replace atoms. In this industrial revolution, the power is so powerful that you will not have a second chance."

Jen-Hsun Huang told Merz and advocated that Germany needs to ensure that its development is safe and responsible while cultivating its local AI industry. Huang Renxun believes that policy makers should focus on supervising the application side of AI, just like supervising various network products, rather than trying to comprehensively control the underlying technology.

Jen-Hsun Huang hopes that Germany will have a world-renowned "Small and Medium Hidden Champion Enterprise (Mittelstand)". These companies have accumulated massive industrial manufacturing data. If they can combine this profound physical industry experience with artificial intelligence, it will inject strong new momentum into German industry.

The development of this year's AI craze actually did not fall from the sky overnight, but has accumulated based on the past development of semiconductors and information communications. If Germany wants to develop AI, it will have to re-examine its past deficiencies in the development of semiconductors and information communications before it can catch up.

There were two major failure factors in Germany's industrial development in the past: relying on Russia for energy and relying on China for cheap energy.

The shortcomings of relying on Russia for energy were fully exposed in the Ukraine-Russia war in recent years. If you want cheap energy, you will end up putting your national security neck on the knife of a threatening country that is eyeing you. In the face of the more energy-consuming AI industry, if this problem is not faced squarely, Germany's AI industry will always be controlled by Russia.

As for relying on China for cheap products, it is also a fatal factor. Since 2018, when the United States has actively lobbied relevant allies to implement clean networks, Germany has actually been unwilling to do so. After intense pressure from the United States, Germany has only begun to face up to the harm of Huawei base stations to German national security in recent years, and has slowly removed Huawei's presence in Deutsche Telekom's core network.

In the AI ​​world, China is now providing a large amount of cheap and low-cost Chinese AI model computing power, which is just like a replica of the base stations Huawei provided back then. Therefore, if the German government does not face up to the threats posed by Russia and China to Germany, its AI industry may be like its declining semiconductor industry. It will have to wait for a day in the future when companies like TSMC can set up factories in Germany to save the German semiconductor industry.

Of course, this dilemma is not just for Germany. In fact, all countries including Taiwan should think about the above issues.