NVIDIA provides kick-starter for AI – is Europe ready for it?

What can NVIDIA bring to the table for companies in Europe to catch up in the AI race, and how is the discussion about AI evolving? NVIDIA’s GTC event in Berlin in mid-October 2026 was an excellent opportunity to explore these questions.

GTC Berlin was not the place for a hype announcement like the crazy flying roadster from Tesla, but NVIDIA presented new building blocks to jump-start the AI age. NVIDIA makes AI factory deployments faster and more sustainable, AI agents more secure and enterprise-ready, and sovereign AI available at the personal level. One question remains: Is Europe ready for all of this?

Here are the main announcements from GTC Berlin and what they mean for the European market:

AI Factories

• Modular AI containers to build AI factories faster. NVIDIA DSX POD allows companies to shorten deployment time from months to weeks. While this can help Europe catch up with other regions of the world, it puts even more pressure on policymakers to streamline approval processes and accelerate power grid expansion. If these issues remain unresolved, Europe will continue to be too slow to catch up.
• AI factories will be less hungry for natural resources. NVIDIA DSX AI Factory Platform significantly reduces the consumption of water (for cooling) and power. Both aspects are critical, not only for sustainability reasons, but also for meeting local requirements and speeding up approval processes for AI factory deployments across Europe. Access to water and renewable energy as well as energy grid stability have been a bottleneck for new big data centers; the Nordic region has so far been a real hot spot for AI factory deployments in Europe. This may change now.

AI Agents

• Securing agentic AI is a top priority. NVIDIA OpenShell for Safe AI Agents is a step in the right direction. Unsecured agents are a real threat to the entire AI ecosystem. When people, companies, and governments lose trust in these systems, adoption slows down, especially in Europe. We all know that building trust takes time but losing trust can happen in seconds. With this in mind, everyone should be aware that we are just one major incident away from a collective rethinking of AI.
• Connecting AI agents to enterprise data. NVIDIA Agent Knowledge Foundation and the new NVIDIA Kumo models help companies connect agents to structured and unstructured enterprise data. This enables the creation of agentic workflows in different areas, such as research labs, factory floors, and industrial engineering. The ability to create sovereign systems composed of internal data, enterprise-wide agents, and infrastructure makes them highly relevant to European companies.

Edge Computing

• NVIDIA is moving into AI PCs. Just a few days before GTC, NVIDIA launched RTX Spark laptops, based on Windows, for personal AI workloads that can run locally. This step explains why NVIDIA changed its reporting structure in February 2026 by introducing a new segment called Edge Computing. This underscores the relevance of this topic for NVIDIA and helps European users run AI locally for personal needs.
• Ready-to-use humanoid robot for researchers. Agile Robots announced Agile ONE SN, a humanoid reference platform built on NVIDIA Isaac GR00T. The aim is to provide researchers with a general-purpose robot to develop skills for many different tasks simultaneously. This announcement reveals a weakness in the current humanoid robot ecosystem: every provider invests in training highly use case-specific skills for their own robots. This limits scalability. Standardized hardware to create a software-defined system on top sounds like a good solution, but it will not work. All providers have a hardware-centric business model, and hardware differentiation will remain relevant in the coming years. As a result, we do not expect this announcement to have a significant impact on the European market.

Here are some personal highlights from discussions at the event:

• Investing in AI triggers even more investment in data and change management. The most interesting statement at the event came from an AI user company. Immanuel Schweizer, Enterprise AI Agentic Lead at Merck, said that for every Euro they spend on AI, two more Euros have to be spent on data and seven more on change management. This is an excellent way to describe the relevance of data and company transformation in the AI context. Many thanks for this statement – I totally agree.
• Turn AI from an efficiency engine into an innovation driver. The RoI of AI initiatives in companies is often not clear, and the unrestricted use of frontier models for everything leads to exploding token costs. However, the discussions at the event focused more on AI as an innovation enabler, rather than providing proof points for RoI. Examples of AI as an innovation driver include drug discovery in life sciences, design space exploration in industrial engineering, and autonomous driving/systems. A stronger focus on innovation is good, but so are better tools to measure efficiency gains. As always, you can only manage and optimize what you measure. AI needs better control systems.
• Form-factor discussions are moving away from humanoids. It is good to see that the hype about humanoids is slowing down, and discussions center more on mobile manipulators (sometimes still with a human touch to feed the hype). From a manufacturing perspective, we appreciate focusing more on the business value of mobile manipulators today rather than the long-term opportunity for humanoids (expected after 2030). European companies need efficiency improvements today, not some time in the future.