Want a future for AI networks? Want to realize the real value of AI? Look for an AI agent approach that empowers the smartphone.
For network professionals pondering the future of AI, there are two important questions: Where is AI headed, and what will AI do to networks?
Enterprises have to plan for the future, and the future of AI is important beyond AI itself, since it touches the technology they buy, the networks they build, and the profits they generate. Since we can’t be absolutely sure what the future holds, it’s smart to look at the range of options, from most conservative to most radical. And yes, it’s also more fun.
The most conservative vision of AI and its impact is that all the hyperscalers that are deploying enough GPU servers to challenge the power grid (and residential electric bills) are right — that hosted AI services you can chat with and play with are the way things will always be.
The middle-range vision for AI is that AI agents, which are essentially like software components, will inject intelligent insight into applications, help improve operations, and even facilitate real-time processes and work. They extend the “automation” paradigm we’ve had for half a century or more, modernize it, make it better.
The radical view? Metaverses and virtual or augmented reality. Where we work, where we live, all end up inside a giant computer game, and AI is the dungeon master. Critical real-world properties, like the fact that the thing in front of you is a wall, are reflected into the game, but non-critical properties can be manipulated. The wall can have an ad pasted on it, telling you that the store on the left has an item you’ve been looking for on sale. Or maybe that there’s an exit from the burning building you’re in, if you just crawl to your right. Everything you need to know is literally before your eyes.
If you’re an investor, you probably have whiplash from watching how all the AI stocks move. In the last month, they’ve been up and down as hedge funds see risk, then see opportunity, then go back to risk again. It would be nice to be able to judge which of our three views is most likely to fit the market in the long term, but stock price obviously isn’t going to offer us much clarity. If you’re an enterprise network professional, or work for a network operator or ISP, each of these three futures creates a range of network opportunities and issues. Can we at least see how each of our views could impact the network?
Our first, conservative, view says that AI’s network impact is largely confined to the data center, to connect clusters of GPU servers and the data they use as they crunch large language models. It’s all “horizontal” traffic; one TikTok challenge would generate way more traffic in the wide area. WAN costs won’t rise for you as an enterprise, and if you’re a carrier you won’t be carrying much new, so you don’t have much service revenue upside. If you don’t host AI on premises, you can pretty much dismiss its impact on your network.
Contrast that with the radical metaverse view, our third view. Metaverses and AR/VR transform AI missions, and network services, from transaction processing to event processing, because the real world is a bunch of events pushing on you. They also let you visualize the way that process control models (digital twins) relate to the real world, which is critical if the processes you’re modeling involve human workers who rely on their visual sense.
Could it be that the reason Meta is willing to spend on AI, is that the most credible application of AI, and the most impactful for networks, is the metaverse concept? In any event, this model of AI, by driving the users’ experiences and activities directly, demands significant edge connectivity, so you could expect it to have a major impact on network requirements. In fact, just dipping your toes into a metaverse could require a major up-front network upgrade.
Networks carry traffic. Traffic is messages. More messages, more traffic, more infrastructure, more service revenue…you get the picture. Door number one, to the AI giant future, leads to nothing much in terms of messages. Door number three, metaverses and AR/VR, leads to a message, traffic, and network revolution. I’ll bet that most enterprises would doubt the value of the first door and fear the up-front cost impact of the third, which leaves us with the middle door, to the AI agent.
AI agents say that AI is really just a software component, a pathway to a new kind of automation. It doesn’t foreclose a role for giant AI hosts and public services, nor does it foreclose a world of AR/VR and metaverses. It’s really not about where AI ends up at all, but about how it gets there. Agents move along slowly, a project at a time, in the same way that computing has entered the workplace and our lives all along.
Meta epitomizes the most exciting, most disruptive, most investment-dependent, and ultimately the most profitable vision of AI. Google and Microsoft offer the most accessible, easiest to achieve, version. What’s in between, the agent story, isn’t glamorous and surely isn’t instantly gratifying to any buyer or seller, but it bridges the two, or it could. The success of the bridge lies in its foundation: the chips. And it’s not the super-GPUs that it depends on, it’s the chips in your phone.
Metaverses and AR/VR demand collecting and correlating enormous amounts of data, and some of this has to happen with very low latency, meaning close to the real-world action that’s being visualized. Well, how much AI could we stick in a pair of glasses or in a sensor? But almost everyone carries a smartphone. An AI future that bonds AI tightly to our lives and our work has to be with us while we’re living and working, and there’s one device that already fits the bill. Give the phone a big role in the AI future, and you not only enable real-world, real-work, real-life coupling, you build edge traffic, promote edge computing. Best of all, you do it in a way that doesn’t break anyone’s bank. Many of us, including me, already have phones with AI chips, and if I decided to get AR/VR glasses, I’d want them to link to my phone, wouldn’t you?
If we end up metaversing even with agents, though, aren’t we confronting that issue of network impact? In the end, probably. But we’d get to ease into it. You can step toward the metaverse future without buying all your workers AR/VR glasses and networking them wherever they are. All business applications of a metaverse don’t demand immediate personal visualization. You can start just by putting an AI agent on each phone and letting it cooperate with the rest of your IT applications and data, and with the workers. It truly is the bridge we need for AI to prove its value.
Want a future for AI networks? Want to realize the real value of AI? Look for an AI agent approach that empowers the smartphone. When it reaches out to touch us all, that future will touch through a transformed network.
