NVIDIA's GTC Berlin 2026 to Test the Economic Viability of AI Infrastructure

According to Zhitong Finance APP, NVIDIA (NVDA.US) will hold its GTC conference in Berlin, Germany, from October 20 to 22 local time. CEO Jensen Huang will take the stage on October 21 at 11:00 a.m. local time (5:00 p.m. Beijing time) to deliver a keynote speech. According to NVIDIA’s official website, the conference will cover topics including accelerated computing, Agentic AI, robotics, open models, Physical AI, sovereign AI, and AI infrastructure.

A few years ago, this might have been just another highly anticipated tech event in the artificial intelligence (AI) industry. But at this point in time, AI trading is moving beyond merely betting on the expected growth of computing power demand and entering a more results-oriented “accounting phase” focused on return on investment.

As global investment in AI infrastructure continues to expand, capital markets are shifting their focus on AI trading from whether demand for computing power will grow to more practical questions: Can massive AI capital expenditures ultimately translate into sufficient revenue, cash flow, and productivity?

Therefore, the GTC Berlin conference may serve as a crucial “stress test” for the current AI industry. What the market truly wants to hear from Jensen Huang is no longer just how much performance will improve in the next generation of computing platforms, but three more fundamental questions: How long can AI compute demand continue to grow? Where will the return on investment for AI infrastructure come from? And can this round of capital expenditure ultimately be translated into real economic value?

Key Point One: Where Will the $2 Billion European AI Deal Go?

Before the official opening of the conference, NVIDIA had already planted a significant悬念 for GTC Berlin.

On October 13, Jensen Huang revealed at an event in Paris that NVIDIA will announce approximately $2 billion in new AI investments or deals in Europe over the coming week. Previously, over the past year, NVIDIA had already invested about $8 billion in Europe. Combined, NVIDIA’s total investments and commitments in Europe will approach $10 billion.

During his time in Paris, Jensen Huang announced that NVIDIA is providing funding to four French startups, including healthcare technology company Doctolib, robotics AI company Genesis AI, software company Pigment, and AI materials company Entalpic; subsequently, at an event in London, he named UK AI companies Boltz, Prima Mente, Hiverge, Riverlane, and Foresight.

This means that during GTC Berlin, the market should focus not only on what NVIDIA announces in terms of technology, but also on where the additional $2 billion in European investments will ultimately flow, and whether these investments can translate into greater demand for AI infrastructure.

Germany could be a key destination. Reports indicate that Germany is preparing to advance an AI collaboration with NVIDIA, focusing on Physical AI tools for robotics and autonomous driving training, as well as AI models for industrial applications. If this collaboration further materializes during GTC, it will strengthen NVIDIA’s presence in Europe’s industrial AI sector.

Therefore, Europe is not merely the host location for this conference but may be emerging as a key market for NVIDIA’s next phase of AI ecosystem expansion.

Point Two: Vera Rubin and the AI Factory—Can They Prove That the Compute Cycle Is Still Ongoing?

For NVIDIA, hardware remains one of the core highlights of GTC, but the way the market understands NVIDIA’s products has changed.

In the past, investors discussing NVIDIA focused primarily on the performance, memory, and manufacturing process of individual GPUs; today, NVIDIA increasingly emphasizes the concept of a complete "AI factory," integrating GPUs, CPUs, networking, storage, software, and data center infrastructure into a unified system.

At the GTC conference in San Jose this past March, NVIDIA unveiled the Vera Rubin platform, positioning it as a full-stack computing platform for Agentic AI, encompassing Vera CPUs, GPUs, networking, and storage. NVIDIA also announced the DSX AI Factory reference design and DSX Air, which enable digital simulation before building physical data centers.

Therefore, what may be more noteworthy at this GTC Berlin event is not whether new chips are being introduced, but rather the actual deployment speed of Vera Rubin, the strength of customer orders, and the extent to which unit token computing costs can be reduced.

If NVIDIA can demonstrate that major cloud providers have entered the scale-out phase and prove that the new generation of platforms can significantly reduce AI inference costs, concerns about overinvestment in AI computing power may be somewhat alleviated.

Conversely, if the conference continues to primarily reveal forward orders, data center planning, and capital expenditure commitments, without synchronized validation of actual deployments or end-user revenue growth, it could instead reinforce investors' concerns about a "self-reinforcing cycle" in AI infrastructure investment.

Therefore, what the market needs this time is not just a technical roadmap, but the ability to deliver on business outcomes.

Point Three: Can Agentic AI Become the Next Engine for Compute Demand?

If the AI infrastructure demand over the past two years was primarily driven by large model training, then the next chapter NVIDIA is now trying to tell is likely Agentic AI.

Traditional chatbots are primarily "one-question, one-answer" systems, whereas AI agents capable of autonomously completing tasks require continuous model invocation, database searching, tool usage, file reading, code execution, and multi-step reasoning.

This means AI computation could shift from one-time "request-response" interactions to continuous computing processes.

From NVIDIA’s perspective, this will create a new demand curve: whereas AI computing demand was previously driven primarily by the scale of model training, it may now be driven jointly by pre-training, post-training, inference, and agent workloads.

This is also a key reason why NVIDIA has positioned Vera Rubin as an "Agentic AI platform."

For investors, a key question is whether AI agents can truly achieve large-scale commercial applications, thereby making inference computing demand the successor to training demand.

If, in the future, AI evolves from "training a model" to becoming "24/7 digital employees," data centers will no longer be merely one-time, long-term capital assets, but may continuously expand in response to growing numbers of agents and increased usage frequency.

This directly relates to one of the market’s biggest concerns: whether AI capital expenditures are a one-time surge or will establish a long-term, sustainable infrastructure cycle.

Point Four: Can Physical AI unlock new computing power markets such as automobiles and robotics?

Compared to generative AI, Physical AI is still at an earlier stage of development, but NVIDIA is significantly elevating its strategic importance.

The topics at GTC Berlin this year cover robotics, autonomous vehicles, digital twins, and industrial AI, with a participant lineup that strongly reflects an industrial focus.

From the lineup of participants, this year’s GTC Berlin has clearly expanded beyond the traditional AI chip domain. The involvement of companies such as Mercedes-Benz, Siemens Energy, Nokia, and Wandelbots has made automotive, energy, power grids, communications, and robotics integral parts of the conference.

Particularly noteworthy is that Mercedes-Benz and Wandelbots align with Physical AI applications such as autonomous driving and robotics, while the participation of the head of Siemens Energy’s digital grid business reflects the growing electricity demand driven by the expansion of AI data centers.

This means NVIDIA is driving AI further into Europe's traditional industrial sectors, beyond the cloud and internet companies.

If robots, autonomous vehicles, and industrial AI begin to enter large-scale deployment, the AI market will expand further beyond software and data centers into global manufacturing and automation industries.

For NVIDIA, this could mean a new source of computing power demand; for the market, it means the valuation logic of the AI industry has the potential to expand beyond "large models" to encompass a productivity revolution in the real world.

Point Five: As AI inference costs decrease, could this actually lead to increased demand?

“Lowest Cost per Token” is also a key term to note for this GTC.

After AI entered the era of inference, investors' focus has shifted from peak computing power to cost per unit of computation.

As agent-based AI continuously invokes models, complex tasks may require a large number of tokens. If the computational cost per token is too high, it becomes difficult for AI applications to achieve large-scale commercialization; conversely, as computing efficiency continues to improve and token costs keep falling, the barrier to using AI will also decrease.

There is an economic logic behind this that deserves the market's attention:

Increased efficiency does not necessarily mean a decrease in overall demand.

If technological advancements significantly reduce the cost of AI computation, businesses and consumers may increase their use of AI, thereby driving continued growth in overall computing demand.

Similar logic has previously appeared in the internet, cloud computing, and mobile communications industries.

Therefore, what needs to be observed at the GTC Berlin conference is not whether NVIDIA can reduce the cost per token, but whether AI usage can grow at an even faster rate after the cost reduction.

If the answer is yes, then a decline in hash rate prices could instead become a catalyst for the next wave of demand growth.

Point Six: Will sovereign AI become a new source of computing power demand in Europe?

Sovereign AI is another key theme distinguishing GTC Berlin from conventional AI technology conferences.

Over the past few years, global AI infrastructure has been highly concentrated among large U.S. tech companies. However, as AI is increasingly viewed as part of national competitiveness and industrial security, multiple economies in Europe, the Middle East, and Asia are now seeking to build their own AI infrastructure.

For Europe, sovereign AI means not only having local AI models but also requiring local computing power, data centers, networks, and data infrastructure.

This precisely creates new market opportunities for NVIDIA.

Topics at this year's GTC Berlin conference include sovereign AI, and prior to Jensen Huang's keynote, events invited executives from European companies and AI firms such as Deutsche Telekom, Dassault Systèmes, and Mistral AI to participate in discussions.

Meanwhile, the participant list includes the European Sovereign AI Center, high-performance computing institutions, and major industrial enterprises, indicating that NVIDIA is seeking to connect its AI infrastructure with Europe’s local industrial ecosystem.

For NVIDIA, if future European countries, companies, and industry groups all seek to build their own AI infrastructure, the growth in AI capital expenditures will no longer rely solely on a few U.S. tech giants.

On the other hand, the European market also brings more complex regulations, data sovereignty requirements, export controls, and competition from local AI chips. Thus, GTC Berlin is also a crucial window into NVIDIA’s global strategy.

Final Exam Question: Can AI capital expenditures achieve an "economic value loop"?

Ultimately, the most important question at the GTC Berlin conference is not whether NVIDIA can once again unveil a more powerful computing product.

Over the past few years, the AI industry has established a very strong growth chain:

Model capability improvements → Increased demand for computing power → Cloud providers increase capital expenditures → NVIDIA's revenue growth → Market raises AI expectations → More capital flows into AI infrastructure.

But by 2026, the market began questioning whether this chain could truly be closed.

Especially as AI infrastructure becomes increasingly reliant on debt financing, the issue has expanded beyond equity valuation to the credit markets.

The financing and profitability pressures recently faced by AI infrastructure companies are a microcosm of the market’s renewed scrutiny of AI capital expenditure returns. Investors are now focusing not just on “strong AI demand,” but also on revenue, free cash flow, capital returns, quality of customer contracts, compute utilization, financing costs, and ultimately, how much end users are willing to pay for AI.

In other words, the AI industry is transitioning from a technology cycle to an economic cycle.

Therefore, the signals that may be released by the GTC Berlin conference can be summarized into five questions:

First, are the Vera Rubin orders and deployments strong enough to demonstrate that the AI infrastructure cycle is still ongoing?

Second, can agentic AI drive sustained growth in reasoning demand, shifting AI compute requirements from training cycles to more stable usage patterns?

Third, can the AI factory and the next-generation computing platform continuously reduce the cost per token and create greater demand for AI usage through improved efficiency?

Fourth, can robotics, autonomous driving, and industrial AI generate genuine large-scale commercial orders, opening new computing markets for NVIDIA beyond cloud computing?

Fifth, and most importantly: Can large tech companies' AI capital expenditures ultimately translate into revenue, cash flow, and productivity gains?

If Jensen Huang provides sufficiently compelling answers in Berlin, recent market concerns about an AI bubble, excessive capital spending, and peak demand for computing power may be somewhat alleviated.

But if the conference continues to primarily reveal technical roadmaps, long-term orders, and capital expenditure plans, while the actual commercial returns from AI applications remain unverified, GTC could instead become a key moment for the market to reevaluate AI valuations.

Therefore, what this year’s GTC Berlin conference truly needs to verify is no longer whether AI can become even stronger, but whether AI, once stronger, can generate sufficient economic value to justify the ever-growing capital investments.