I don’t really do predictions — at least not in any big, fiery, pie-in-the-sky sense. The trap I've fallen into with prior attempts is investing too much into the notion of how technologies should progress, rather than the relative likelihood that things will turn out that way.
Technology's trajectory is not the simple product of future breakthroughs, it's how such breakthroughs are utilized, monetized, deployed, and how the interests of various invested parties are met. Predictions can thus be useful in helping to form our individual notions of the possible and plausible. They can broaden our context and parse out the likelihood of different scenarios.
I respect individuals with a knack for framing the current moment on a historical timeline, in a bid to extend lines of sight. As someone still prone to the occasional school anxiety dream, I'm glad I don't have Rodney Brooks grading my work each semester. I am, however, grateful that he issues and retrospectively evaluates his own forecasts at the top of every year (more below).
When a friend and former colleague asked me for my own 2026 robotics predictions, I wrote myself an essay. It was useful for gathering my thoughts, but more than anything, it highlighted my circuitous approach to predicting the near future.
If you’re at all interested in my state of the union on robotics heading into 2026, go read Lance’s piece for TechRadar linked below — he was my editor many lifetimes ago and still does yeoman’s work of surfacing the important points among my meanderings. The more closely I cover the space, however, the more I'm inclined to frame desired outcomes in terms of the obstacles they face.
Back to meandering: In 1958, KTLA introduced the Telecopter, providing Los Angeles a wholly new perspective on the news. From the sky, you can see the accidents that cause the bottlenecks that shape the traffic patterns. The video from the helicopter didn't offer any single viewer a one-to-one picture of what their commute might look like on a given morning, but educated audience members could adjust their departure times accordingly.
The end of 2025 saw an uptick in mainstream press coverage around some of the challenges facing wide-scale humanoid deployment. Tempering expectations is a net positive — the human brain is a complex tool capable of processing excitement and pragmatism simultaneously. There are mornings I wake up and would love to be all Han Solo "Never tell me the odds” about where the industry is headed, and then I remember how things went for him in the sequel trilogy. (Tech history is retconned by the victors.)
Counterintuitive as it might seem at first, waking up one morning to a 2,000-word New York Times piece about why humanoids are really hard is its own bellwether for a potential bubble. It's an indication that coverage around the limitless possibilities of general-purpose robots has been wide enough spread to have prepared readers for its counter-argument.
If you needed another indication that some folks are prepared to discuss humanoid robots like they're smartphones, you also could have just waited a few more weeks for CES to kick off.
I’ve had more than one person float CES 2026 as the “Year of the Humanoid.” In terms of numbers, fair enough. As recently as a year or two back, these robots were viewed as strange outliers at consumer-focused events like CES. In 2026, the various humanoids are competing with one another for column space.
Along with the sheer volume of robots present at the show, plenty of good actors find themselves competing with competitors' lofty promises. This is especially difficult to navigate for an audience that still doesn’t have a great barometer as far as what these systems realistically can — and can’t — do.
The closest thing I’ve gleaned as far as a consensus among the players themselves is that we’re still at the very early stages of this. The market is going to evolve a lot — and in unexpected ways. As the category matures, some of the biggest names will bow out, and be replaced by companies we’ve not yet heard of. All of which, to be fair, makes my job a heck of a lot more interesting.
How's that for a prediction? |
|
