Why Telco Took Top Billing at GTC Berlin

GTC Berlin represented this year’s European edition of Nvidia’s flagship event series and a showcase of how the company works across diverse verticals. Industrial AI was an obvious focus given Germany’s aspirations to evolve its manufacturing legacy in line with modern tech (IE, AI and robotics). Telco, however, was also front-and-centre across the opening keynote, product launches and partner announcements. 

Across the myriad verticals all rushing to leverage AI – and all working with Nvidia – telecom’s prominence is noteworthy. The “why” of this prominence speaks to the ways in which telco can benefit from AI and ways in which Nvidia is positioning itself to benefit from the telco industry’s requirements. 

• Mobile & Standards. Nvidia likes to think of itself as a company that creates standards, at least in practice; time and again at GTC, CEO Jensen Huang referred to various Nvidia products and solutions as de facto industry standards. Telco networks, mobile in particular, are dominated by standards that drive interop and scale. While they are more formal than de facto ones, the imminent arrival of 6G means there’s room for Nvidia to make its mark if it moves quick. 
• Network Innovation. Even where standards tightly dictate a technology’s development, there’s still room for innovation. If that wasn’t the case, you would see much less vendor activity at trade shows like MWC. For its part, NVIDIA has claimed that AI in the RAN can drive new efficiencies and network capabilities, proving that this is more than just theoretical is the challenge. 
• The AI Layer Cake – Part 1: Scaled Networks. Nvidia has discussed AI as a “layer cake” comprised of: land and power, chips, infrastructure, models, and applications. It’s best known for delivering chips which support infrastructure that power models and applications. A compute-intensive industry, telecom stands as a vertical into which Nvidia could sell lots of chips and infrastructure if it manages to get the capabilities and form factors right.
• The AI Layer Cake – Part 2: Filling. Nvidia’s layer cake analogy largely ignores the filling that connects layers. Some of that will take place within compute or infrastructure solutions. Much of it will be delivered by telcos and their networks. 
• Demanding Workloads. Every vertical believes that its activity is critical. Telecom, however, is famous for engineering networks to deliver 99.999% uptime in order to meet customer requirements. As governments and societies increasingly look to those networks as critical infrastructure, you can add in new “sovereignty” requirements. Demanding workloads, in turn, suggest a premium to be paid for compute or broader solutions which can meet telco needs. 

Beyond any theoretical reasons for Nvidia’s interest in telecom, we can’t ignore it’s $1 billion investment for a 2.9% share in Nokia almost a year ago – a bet on the telecom opportunity as much as the vendor. Today, Nokia’s market cap is just about $47B and Nvidia’s 2.9% share is worth $1.35B. On this front, it seems that the company’s plan is paying off. We’ll be looking at how it’s doing with the others – how it’s addressing standards, scaling, network innovation and demanding workloads – in a follow-on analysis.