Nvidia’s GTC conference will be held in Berlin, Germany, from Oct. 20 to 22 local time, with company founder and CEO Jensen Huang scheduled to deliver the keynote on Oct. 21. It is Nvidia’s highest-profile technology and ecosystem event of the second half, and the signal from this year’s gathering is already clear: the chip giant is shifting its narrative focus from “selling GPUs” to “selling AI factories.”
According to the official agenda, this year’s GTC features more than 100 sessions spanning the full AI technology stack, with key areas including accelerated computing, agentic AI, physical AI, robotics, open-source models and AI infrastructure. For investors following Nvidia, the event is both a testing ground for its technology road map and an important window into the pace at which its European strategy is being implemented.
On the hardware side, actual deployment progress of the Vera Rubin platform is widely seen as a key focal point. At the San Jose GTC in March this year, Nvidia positioned Vera Rubin as a full-stack computing platform for agentic AI, spanning CPUs, GPUs, networking and storage. In Berlin, the market is no longer expecting another road map but rather whether major cloud providers are truly entering the large-scale deployment phase, and whether per-token inference costs can post a substantial decline. Those two indicators bear directly on whether the AI factory model can move from concept to a replicable commercial loop.
More than hardware specifications, Nvidia’s capital moves in Europe may be worth tracking more closely. Judging from the company’s recent string of actions, Europe is being shaped into a key market for its next phase of AI ecosystem expansion.
On Oct. 13, Huang said at an event in Paris that Nvidia would announce about $2 billion in new European AI investments or deals within the coming week. Over the past year, the company has already invested about $8 billion in Europe. Combined, Nvidia’s European investments and commitments will approach $10 billion.
While in Paris, Huang announced that Nvidia was providing financing to four French startups: health-tech company Doctolib, robotics AI company Genesis AI, software company Pigment and AI materials company Entalpic. Later, at an event in London, he named British AI companies Boltz, Prima Mente, Hiverge, Riverlane and Foresight. Judging from the distribution of investment targets, Nvidia’s European buildout is not purely financial investment but an effort to weave an ecosystem network around physical AI, materials discovery and industrial software.
Germany is likely to be the main battleground for this round of expansion. According to German business daily Handelsblatt, Germany is planning an AI partnership with Nvidia and is expected to receive several hundred million euros in government funding, with a formal announcement possibly this week, almost in tandem with the GTC conference.
The partnership is based on Nvidia’s Nemotron family of language models and Cosmos physical AI models, with two interrelated goals: developing “physical AI” for robots and autonomous vehicles, and training industrial-scenario AI models with international competitiveness. More than a dozen German companies are expected to take part, including Siemens, Audi and Black Forest Labs, while Deutsche Telekom’s industrial AI cloud platform will provide training infrastructure.
If the partnership materializes as planned, its significance goes beyond a single order. For Nvidia, it means its models and computing platform will be embedded in Europe’s manufacturing system in the form of a “national-level industrial partnership” for the first time; for German industry, it means bringing in an external technology foundation in areas where its own software and AI capabilities are relatively weak.
From a broader perspective, Nvidia’s decision to step up in Europe now is closely tied to changes in the local regulatory environment and industrial policy. The European Union has in recent years pushed ahead with a digital sovereignty agenda, demand for localized computing infrastructure has risen, and governments are willing to trade subsidies for AI capabilities. By entering through government partnerships combined with startup financing, Nvidia can avoid the policy friction that comes with simply selling hardware while locking in ecosystem partners early.
In terms of market impact, this year’s GTC could shape expectations along three threads. First is the deployment pace of Vera Rubin: if major cloud providers confirm large-scale adoption, it would strengthen market confidence in the revenue ramp for Nvidia’s next-generation platform. Second is per-token inference cost: a clear decline in this indicator would mean improved economics for AI applications, a positive for the entire downstream software and services chain. Third is the implementation of European investments and government partnerships, which offers room to imagine Nvidia opening a third growth pole beyond North America and mainland China.
Of course, there is still a distance between announcing investments and generating meaningful revenue contribution. The maturity of Europe’s AI ecosystem, the procurement pace of local companies and the subsidy approval processes of individual governments could all affect how quickly projects advance. How many verifiable details Huang provides in his Oct. 21 keynote will be the first yardstick for judging the substance of this European push.