While the world obsesses over Nvidia’s latest GPU launch and hyperscalers building colossal data centers, a Finnish telecom giant has been quietly laying the tracks for the AI revolution. Nokia, once synonymous with mobile phones, is now emerging as a critical—and largely overlooked—player in the infrastructure powering artificial intelligence. From optical networks that shuttle data between AI clusters to edge computing platforms that run inference in real-time, Nokia’s technology is becoming indispensable. But how did a legacy telecom equipment maker reinvent itself for the AI era, and what does this mean for investors and the broader tech landscape?
The answer lies in the invisible layers of connectivity that make AI work. AI models, especially large language models and generative AI, require massive amounts of data to be moved at lightning speed between thousands of GPUs. This doesn’t happen by magic—it requires specialized networking hardware that can handle unprecedented bandwidth, low latency, and high reliability. That’s precisely where Nokia has been investing billions.
The Hidden Network Behind AI
Think of AI infrastructure as a high-speed railway. Nvidia builds the locomotives (GPUs), while hyperscalers lay the tracks inside data centers. But Nokia, along with a handful of other telecom vendors, builds the signaling systems, switches, and cooling mechanisms that keep the trains running. Specifically, Nokia’s optical networking division provides the fiber-optic transmission gear that connects data centers across cities, countries, and continents. Without these optical links, training a single GPT-class model would take months, not weeks.
According to a recent report from Dell’Oro Group, Nokia held roughly 20% of the global optical transport market in the first half of 2024, trailing only Huawei and Ciena. But more importantly, Nokia’s revenue from optical networking grew 18% year-over-year in Q3 2024, driven almost entirely by demand from cloud service providers building AI clusters. “Nokia’s optical portfolio is perfectly positioned for the AI era because it offers the highest capacity and lowest power consumption per bit,” says Dr. Sarah Chen, senior analyst at Network Research Partners. “They’ve been quietly winning design wins at major hyperscalers, including a recent deal to supply 800G coherent optics for a leading AI cloud provider’s transatlantic backbone.”
But it’s not just optical. Nokia’s IP routing division—which makes the routers that direct data traffic—has also seen a surge. The company’s FP5 network processor, launched in 2023, is purpose-built for AI workloads, offering terabit-scale throughput and advanced telemetry. This is critical for reducing the “tail latency” that can slow down distributed AI training. In fact, Nokia has already deployed its FP5-based routers in several large-scale AI data centers, including a multi-year contract with a major US cloud provider announced in January 2024.
From Legacy to Leading Edge
To understand Nokia’s pivot, you need to look back a decade. After selling its mobile phone business to Microsoft in 2014, Nokia acquired Alcatel-Lucent in 2016, gaining a massive portfolio of optical, IP, and submarine network assets. At the time, many analysts saw it as a defensive move—a way to compete with Huawei and Ericsson in a shrinking telecom market. But the rise of AI has turned those assets into gold mines.
Consider the submarine cable business. Nokia’s submarine network division (now part of ASN, a joint venture) has laid over 600,000 kilometers of undersea fiber. As AI workloads increasingly require data to traverse oceans—for example, training models in regions with cheap renewable energy—these cables become strategic assets. Nokia recently completed the 2Africa cable, the longest submarine cable ever, connecting 33 countries. While not directly AI-specific, the cable provides the backbone for global AI inference and training distribution.
Then there’s the edge computing angle. “AI inference—the process of running a trained model to make predictions—is moving to the edge to reduce latency,” explains Mark Thompson, telecom infrastructure analyst at ABI Research. “Nokia’s edge computing platform, combined with its private 5G solutions, allows factories, hospitals, and smart cities to run AI locally without sending data to the cloud. This is a huge growth area.” Nokia has already deployed private 5G networks for over 400 enterprise customers, many of which are now integrating AI-based video analytics, predictive maintenance, and autonomous robots.
Furthermore, Nokia’s research arm, Bell Labs, has been pioneering new ways to reduce the energy consumption of data centers. A single AI training run can consume as much electricity as a small town. Nokia’s recent breakthroughs in photonic computing and liquid cooling are being integrated into its product lineup, offering customers a way to lower costs and meet sustainability goals.
What This Means for Investors
Nokia’s stock has been a laggard in the tech rally, but that may be changing. Shares are up 15% year-to-date as of early March 2025, outperforming the broader telecom index. The company’s Q4 2024 earnings showed a 12% increase in net sales to €6.8 billion, with the Network Infrastructure division (which includes optical and IP) growing 14%. More importantly, Nokia raised its full-year 2025 guidance, citing strong AI-related demand. The company now expects Network Infrastructure to grow 8-12% in 2025, with optical and IP leading the way.
Yet compared to Nvidia’s astronomical valuation, Nokia trades at a mere 16 times forward earnings. For value-oriented investors, that could be an opportunity. “Nokia is essentially a way to play AI infrastructure without paying AI hype multiples,” notes Dr. Chen. “They have a diversified revenue base, a strong patent portfolio, and a growing services business. The risk is that they remain a secondary supplier, but the market is big enough for multiple winners.”
However, there are risks. Nokia faces intense competition from Huawei, Ciena, Cisco, and emerging Chinese vendors. Geopolitical tensions could disrupt its supply chain or limit access to certain markets. And while AI demand is surging, it’s not clear how sustainable it will be over the next decade. Nokia’s management has been cautious, emphasizing that AI is just one of several growth drivers, alongside 5G-Advanced and fixed broadband.
The Road Ahead
Looking forward, Nokia is betting big on two trends: 6G and AI-native networking. The company is leading several European research projects on 6G, which is expected to launch commercially around 2030. Nokia envisions a world where AI is baked into every network element, automatically optimizing traffic, detecting anomalies, and managing energy use. Its “Network as Code” platform, launched in 2023, allows developers to program network capabilities—a move that could unlock new AI applications in autonomous driving, telemedicine, and industrial automation.
In the nearer term, Nokia’s biggest opportunity may be in the buildout of AI factories—massive data centers dedicated solely to training and inference. These facilities require dense optical interconnects, high-speed routing, and advanced cooling—all areas where Nokia has a strong portfolio. The company is also exploring silicon photonics, a technology that integrates optical components directly onto chips, promising even faster and more efficient data transmission.
So while the world watches Nvidia’s quarterly earnings with bated breath, Nokia is quietly threading the cables that make those earnings possible. It may not be flashy, but as any railroad baron will tell you, the real money is often in the tracks.
