Reading the CES 2026 Tea Leaves
As CES 2026 approaches, the real themes are already emerging from the pre-show noise.
With just about a week to go before CES 2026 officially begins, some clear themes are already emerging via the steady stream of analyst and media outreach landing in overstuffed inboxes. This isn’t the polished, on-site story yet. This is the pre-CES story – the one companies tell while they are still trying to frame the conversation, not simply reinforce it.
Building the Bridge Between AI and Reality
The most obvious pattern is that nearly every firm is trying to position itself as a bridge. Specifically, a bridge between the promise of AI and the physical world we actually live in. The messaging is less about abstract intelligence and more about embodiment – how AI gets eyes, ears, hands, and some form of persistent presence outside the cloud.
That bridge shows up in several distinct ways, depending on where a company sits in the ecosystem.
One version of the bridge is wearable. Smart glasses, wrist-based controls, gesture systems, and other “eyes-up” interaction models are appearing in far greater volume than in past years. The pitch is familiar by now: ‘AI is powerful, but it needs a more natural input-output mechanism to be useful in daily life – and it won’t be the mobile device you already own.’ Touchscreens and keyboards are framed as legacy interfaces, while wearables are positioned as the missing layer between humans and intelligent systems.
What’s notable this year is not simply that these devices exist – they’ve been teased for years – but that companies are now talking about them as modular, cross-compatible components rather than singular, closed products. The implication is that the future isn’t one dominant headset or one magic wearable, but an ecosystem of loosely coupled devices that together create a usable AI experience. Whether that ecosystem actually materializes, or collapses under its own fragmentation, remains to be seen. After all, recent history has given us more than one reminder of how quickly “AI-first” hardware can fall apart once it leaves the stage, with the Humane AI Pin and Rabbit R1 standing as cautionary examples of promise vastly outpacing substance.
When AI Shows Up as Robots
Another, more literal version of the bridge shows up in robotics. Some companies aren’t bothering with metaphor at all. Instead of talking about “connecting AI to the physical world,” they are rolling out machines that already live there. Autonomous systems that can move through a space, interact conversationally, and perform a physical task are being framed as proof points rather than concepts.
These products are interesting not because they are perfect – they aren’t – but because they force the conversation out of theory. Once AI has a physical presence, even in something as mundane or playful as a service robot, expectations change quickly. Reliability matters. Context matters. Memory matters. And the gap between a clever demo and a product that can survive real-world use becomes very obvious.
The examples go far beyond the semi-serious toilet-paper delivery robots of a few years ago and arrive in the shadow of early category leaders, including the original robotic vacuum company now needing to declare bankruptcy protection amid intense follow-on competition.
When Embedded Intelligence Breaks the Surface
A third theme coming through loudly this year comes from companies that historically operated behind the scenes. Chipmakers, microprocessor vendors, and embedded technology firms are no longer content to stay buried inside someone else’s product. They are pushing their way into the foreground and making the case that none of this works without them.
If AI is going to run locally, last all day, and respond in real time, it needs silicon and software designed specifically for those constraints. What’s different this year is how forcefully these firms are making that argument. Instead of quietly supporting other brands’ announcements, they are positioning themselves as the core enablers – the engines behind wearables, robots, sensors, smart infrastructure, and edge-based intelligence.
Many of these companies will have a larger presence at CES than ever before, and they are no longer content to be hidden components. Embedded intelligence, long assumed and rarely discussed, is very deliberately breaking through the surface.
The AI PC Arrives – But For Whom?
Closely related, but distinct enough to stand on its own, is the renewed push by PC manufacturers. This year’s pre-briefings include a noticeable surge in messaging around next-generation PCs with onboard, edge-based AI capabilities. The promise is that local AI processing will make personal computers smarter, faster, and more useful without relying on the cloud.
That pitch raises two uncomfortable questions. First, will this actually help most users in a meaningful way, or is it simply a new spec-sheet story in search of a killer application? Second, after hundreds of thousands – if not millions – of white-collar layoffs across the tech sector, is there really an audience eager to refresh hardware right now?
There is a strong case to be made that the market is already saturated with underutilized PCs, both in enterprises and home offices, and this new capability alone may not be enough to drive a meaningful replacement cycle.
Still CES – The Weird, the Wonderful, the Physical
Layered on top of all of this are the occasional standalone products that don’t fit neatly into any category. These are the things that still make CES feel like CES. This time around, these are expected to include a chocolate 3D printer that takes a physical object and recreates it in chocolate, turning novelty and personalization into manufacturing efficiency. New user-interface models for desktop computing that rethink familiar workflows. Devices designed not to replace smart homes, but to manage the sprawl of smart devices people already own.
This is also where the show’s spectacle still asserts itself. Tech-celebrity appearances, carefully scheduled meet-and-greets, and surprise cameos remain a reliable draw for exhibitors looking to create buzz. And alongside that theater, the industry’s more serious traditions continue as well, including the naming of this year’s IEEE Medal of Honor recipient – a reminder that beneath the noise, there is still deep, foundational work being recognized.
In some cases, these products and moments are delightful curiosities. In others, they quietly point to real problems that haven’t been solved particularly well yet. The common thread is that they are less concerned with AI as a headline and more interested in solving a specific, tangible problem – sometimes with AI, sometimes merely enabled by it.
Cutting Through The Hype
Taken together, the upshot of this year’s pre-CES impressions is clear. There is going to be a lot of AI-washing on the show floor. There will be plenty of promises about connecting virtual intelligence to the physical world, along with no shortage of claims that it is one company’s technology or another’s platform that truly makes all of this work.
The harder – and more interesting – job will be navigating through those claims to find the products that actually deliver. The ones that work outside a controlled demo. The ones that don’t require heroic assumptions. The ones that still make sense once the buzzwords fade and the hype cycle inevitably cools.
That has always been part of CES. This year, however, the distance between promise and reality feels wider than usual. When much of the AI noise settles, a smaller group of real products and real winners will remain. Those are the ones worth paying attention to – and they rarely announce themselves as loudly as everyone else.
For me personally, this is familiar territory. I’ve been attending CES for many decades, long enough to have seen entire themes rise, crest, and quietly disappear. I’m still waiting for my five-second movie downloads that were once promised about 5G with absolute certainty. I’m also still waiting for the metaverse, so loudly touted just a few years ago, to become something more than the useless buzzword it ultimately proved to be.
Over the years, I’ve watched genuinely transformative products emerge from unlikely corners of the show. I’ve also seen ideas that were, in hindsight, deeply misguided – sometimes amusing, sometimes baffling – arrive with great fanfare and go absolutely nowhere. I expect all of that to happen again this year. (That crashing sound we all heard after the pandemic was AR and VR glasses missing their exit, despite millions of people stuck at home still failing to find a reason to put them on – and yet, sadly, they’re still trying.)
What keeps CES worth attending isn’t the headline announcements that will be recapped everywhere else within minutes of keynotes and press conferences ending. You can look just about anywhere for those. For me, it’s the quieter moments. The hidden booths. The smaller exhibitors still refining something real. The technologies that aren’t quite finished yet, but clearly hint at where things may actually be going.
That’s where the signal usually hides, beneath the noise. And that’s where I’ll be spending my time once again – not chasing the loudest promises, but looking for the quieter, emerging pieces that make CES worth coming back to year after year.