Jensen Huang’s blunt critique of Germany lands as Berlin prepares a state-backed AI pact with his company. Yet the data suggests the more consequential shift is not Europe’s lag, but the closing gap between the United States and China.
Berlin — Jensen Huang rarely minces words, and this week the Nvidia chief executive aimed his characteristic candor at Germany. Speaking at a press briefing in London, Huang said it “hurts me more than anything else” to see one of the world’s most technologically advanced nations fall so far behind in artificial intelligence. He attributed the gap to a European tendency to “overthink everything,” concluding with a blunt prescription: “Just get to work”.
The remarks, first reported by German tech outlet WinFuture and amplified across the country’s media, carry immediate political weight. On Monday, Chancellor Friedrich Merz is set to unveil a comprehensive AI pact with Nvidia at the “Invest in Germany” summit in Berlin. The federal government is prepared to commit a high three-digit million-euro sum, with companies including Siemens, Audi and Black Forest Labs expected to participate. The funds would purchase computing capacity in Deutsche Telekom’s Industrial AI Cloud, which already houses 10,000 Nvidia Blackwell-generation chips.
The subtext is unmistakable: Germany, the continent’s industrial engine, is being positioned as a recipient of American compute rather than a builder of its own. Huang’s criticism, however uncomfortable, is difficult to dismiss on the evidence.
The German Deficit, Quantified
Germany’s AI adoption figures tell a story of uneven progress. By 2025, 26 percent of German companies with at least ten employees used at least one AI technology—above the EU average of roughly 20 percent, but well behind Denmark at 42 percent, Finland at 38 percent, and Sweden and Belgium at 35 percent. Among large companies with 250 or more employees, adoption reaches 57 percent. But for small firms with 10 to 49 employees, the figure drops to 23 percent.
More telling is the quality of adoption. According to KfW data, only three percent of German medium-sized companies use sophisticated data analysis, and a mere one percent employ AI for autonomous machine movements—the applications most relevant to industrial productivity. Digitalization spending by medium-sized firms fell to €23.8 billion in 2024, a decline of €8.1 billion year-on-year, weakening the foundation on which AI must scale.
Venture capital tells a starker story. Of roughly $259 billion in global AI venture funding last year, $194 billion flowed to the United States. The European Union received $15.8 billion. Germany’s share of that is a fraction. Huang’s “overthinking” diagnosis may be impolitic, but the funding and deployment data lend it credence.
The American Lead Is Real—and Shrinking
Yet the more consequential story lies not in Berlin but in the broader trajectory of the US-China AI competition. Huang’s warning to Germany comes as American primacy in the field is itself under pressure.
The Stanford University AI Index, released in 2026, found that the performance gap between US and Chinese AI models has “effectively closed.” Leading systems from both countries have traded top positions multiple times since early 2025. By March 2026, the top US model led by just 2.7 percent. US Treasury Secretary Scott Bessent has acknowledged that the American lead may now be as narrow as three to six months.
China’s gains are not confined to model performance. In 2024, Chinese entities accounted for 74.2 percent of all AI patents granted globally—97,206 out of 131,121—up dramatically from a decade ago, when the US held 42.8 percent. China contributed 17.8 percent of global AI research papers, more than double the US share of 7.6 percent. Chinese institutions produced 41 of the world’s 100 most-cited AI papers in 2024, up from 33 in 2021, while the US presence declined from 64 to 46.
On computing power—the metric Huang knows best—the picture is more nuanced but trending in one direction. The US still accounts for nearly three-quarters of global AI computing capacity, according to Epoch AI, with China holding just over 14 percent. Nvidia alone represents more than 60 percent of global AI computing capacity among major chip designers. But in China’s domestic AI accelerator market, the shift has been rapid. Nvidia’s share fell from roughly 55 percent in 2025 to 34 percent in the first half of 2026, with shipments down 44 percent year-on-year. Huawei’s Ascend chips reached 28 percent share, and combined with Cambricon and Alibaba’s T-Head, domestic vendors now hold roughly 47 percent of the Chinese market—surpassing Nvidia.
The Adoption Paradox
There is a further irony in Huang’s critique of Germany. While the US retains leadership in frontier model development and advanced computing infrastructure, China is outperforming it in model adoption. Chinese models’ share of usage on Open Router climbed from about 1 percent in 2024 to over 60 percent in May 2026, according to US Representative Sara Jacobs, a California Democrat who sits on the House Democratic Commission on AI. Chinese models are also cheaper and use less compute than leading American systems.
This is the dimension Germany struggles with most: not building frontier models, but deploying AI across its industrial base at scale. Huang’s prescription—“just get to work”—is less an insult than an acknowledgment that the gap between pilot projects and productive deployment is where technological advantage is ultimately lost or won.
The Nvidia pact may buy Germany compute and political cover. Whether it buys the capacity to close the deployment gap is another question. And as Washington watches its own lead over Beijing narrow, the lesson from Berlin is one Huang knows well: in AI, the laggards are rarely those who lack ideas. They are those who lack the urgency to act on them.
