NVIDIA’s investment in Nokia: How it will impact future of 6G telecom networks

After making an unlikely and unprecedented investment in Intel in September 2025, NVIDIA has followed it up with an equally surprising $1 billion deal with Nokia – yes, that same legendary mobile handsets brand of the early 2000s we all loved. 

Survey

The investment gives NVIDIA a roughly 3% stake in Nokia. What for? The stated aim is that together NVIDIA and Nokia will co-develop and deploy AI-powered telecom networking technologies of the future, especially one that will help global telco networks transition from 5G to 6G smoothly. 

Add As A Trusted Source For Google.

Add as a preferred source on Google

NVIDIA’s advanced chips, which have become the heartbeat of AI’s all-conquering march on all things digital, will form the bedrock of Nokia’s RAN (radio access network) products and infrastructure. In case you didn’t know, Nokia is the third largest RAN provider in the world (behind Huawei and Ericsson) – and excluding China, Nokia directly competes with Ericsson for the top spot in telco RAN deployments across most of the Western world.

Shift to standardized silicon for RAN telecom networks

As much as we all depend on mobile networks for 4G / 5G internet access, not many think about the hardware that’s deployed to keep those networks always-on. Traditionally, most telecom network hardware has been proprietary and highly customised for only radio communications. 

Unlike standard CPU, GPU, RAM, etc, that go into your desktop PC, laptop or even smartphones – which have over time become modular in design, with standard sockets, ports and interface interconnects – our telecom networks still need specialized hardware (like custom ASICs or FPGAs) to work. Where network functions are rigidly built into the hardware itself, not so much on the software layer that goes on top of it. 

Also read: Nvidia invests $1 billion in Nokia to fuel 6G and AI infrastructure collaboration

Making any changes (software or hardware) or upgrades to these RANs operated by telecom providers around the world is a painstakingly lengthy and costly affair. Needless to say, this status quo feels outdated, especially in the age of AI.

This is where the NVIDIA and Nokia partnership attempts to mark a fundamental shift. By running on NVIDIA chips integrated into standardised commercial hardware, telecom operators around the world will have the ability to instantly make their network infrastructure programmable – both with easy software and hardware upgrades. This also makes RANs more scalable as compute resources, where the network’s footprint can be more dynamically adjusted – by adding or removing CPUs, GPUs, etc or with a few lines of code – depending on demand at any given time, location or region.

As Nokia’s aim is to build, trial and ultimately deploy software that will power 6G networks of the near future, the NVIDIA partnership will boost its ability to make highly scalable, programmable and reliable 6G networks for telecom operators around the world – including here in India. The NVIDIA hardware + Nokia software package aims to make 6G networks more plug n play and less convoluted.

As Nokia’s CEO Pekka Lundmark put it in the official announcement, “Our partnership with NVIDIA, and their investment in Nokia, will accelerate AI-RAN innovation.”

AI-RAN and OpenRAN disruption of telecom networks

This reminds me of what Sachin Katti, Intel’s Chief Technology and AI Officer, told me recently in an interview earlier this month. Katti’s been a vocal advocate for OpenRAN and software-defined networks for some time now, and during our chat he summed up the direction of the entire telecom industry which reads like a prophetic footnote to NVIDIA’s recent moves.

Also read: NVIDIA’s Kari Briski believes open models will define the next era of AI

“At this point, OpenRAN is more of an inevitable conclusion to me… it’s a question of when it flips completely rather than an if, and I think it’s just the economics of silicon, right?” Katti said, when I spoke with him earlier in October 2025, just a few weeks before the NVIDIA-Nokia partnership announcement. “The industry is just not big enough to keep doing custom silicon. You have to leverage the economies of scale of a general-purpose platform,” Katti summed up.

While pitched from Intel’s vantage point, Katti’s statement aligns perfectly with what NVIDIA and Nokia are building towards – which is to bring modular, GPU-powered computing muscle to the telecom world. And replace the old guard of black-box, custom-built network equipment with brand new scalable, programmable alternatives that are of a heterogenous nature.

Intel, it should be noted, is betting on its own CPU and accelerator platforms to power cloud-native 5G networks. NVIDIA, on the other hand, is pushing its GPU stack into telecom – the same stack it’s optimized for AI and large-scale computing. Two different routes aiming for the same destination, and the conversation lately among chipmakers isn’t about competition but forging partnerships.

For decades, mobile networks were locked down by design. Now, they’re about to be opened up – not just by software, but by silicon as well. And this NVIDIA-Nokia partnership marks an important step in that direction.

Also read: Qualcomm hints at 6G rollout by 2028, says AI will soon be everywhere

Follow Us

Jayesh Shinde

Executive Editor at Digit. Technology journalist since Jan 2008, with stints at Indiatimes.com and PCWorld.in. Enthusiastic dad, reluctant traveler, weekend gamer, LOTR nerd, pseudo bon vivant. View Full Profile