When people talk about the artificial intelligence boom, they usually mean chips, especially the powerful processors used to train and run AI models. But those chips are only useful if they can talk to each other quickly. That is where networking comes in. The switches, cables and connections inside data centres have become one of the fastest-growing parts of the AI spending story, and investors are paying close attention.
What AI networking means
Training a large AI model needs thousands of processors working together as if they were one giant computer. To do that, they constantly swap huge amounts of data. Networking equipment links them up. It includes switches that direct data traffic, specialised cables and optical connections that carry data at very high speeds, and the software that manages it all. If the network is slow, expensive processors sit waiting, which wastes money.
Why demand has surged
As AI models have grown larger, the number of processors in each data centre cluster has grown too. More processors means more connections, and faster processors need faster links. Big technology companies have been spending heavily on data centres, as we explain in our guide to AI capex cycles, and a meaningful share of that money goes on networking rather than chips alone.
Companies that make switches, optical components and high-speed cables have reported rapid growth. Arista Networks, for example, reported its fastest revenue growth since listing in 2014 in results published in November 2026, citing demand for AI networks.
How it fits the bigger picture
For investors, networking is sometimes seen as a second wave of the AI trade. Chipmakers benefited first. Networking, power supply and cooling companies followed as the physical build-out of data centres expanded. These businesses depend on the same customers, the large cloud and technology companies, so they share the same risk: if those customers cut spending, demand can drop sharply.
The risks to watch
Customer concentration is a big one. A few very large buyers account for much of the demand, which we cover in our explainer on customer concentration risk. Technology can change quickly, and today’s leading product can be replaced. Valuations can also run ahead of reality when a theme becomes popular. Even strong results may not lift share prices if markets are nervous, as the reaction after the 2026 US midterms showed in our Morning Market Brief.
What investors look for
When assessing AI networking companies, investors watch order growth, guidance for future quarters, profit margins and how dependent they are on a handful of customers. They also follow what the biggest technology companies say about their own spending plans, since that drives demand across the supply chain.
The beginner’s takeaway
AI networking is the plumbing that lets thousands of processors work together. It has grown quickly as data centres expand, making it an important part of the AI investment theme. But it shares the same dependence on big technology spending, so understanding that link is key.
Separating real trends from hype is one of the most valuable investing skills. If you want to see how yours measures up, try the free Samuel & Co Trading Trader Assessment.
