Trending Articles
Trending Topics
Trending
People are split on Las Vegas – some love it, others hate it, but every January large chunks of the tech sector have to go there either way.
The gravitational pull is, of course, the Consumer Electronics Show. This year over 148,000 people attended, the largest number since lockdown. There’s always an unofficial theme to the show and it’s occasionally very wrong. At CES 2010, 3D TV was going to be the next big thing, but you’ll notice their absence from your living rooms 15 years later. This year robots were everywhere, and in many cases rather bad.
Just under 40 companies showed off vaguely humanoid robots, with Chinese manufacturers making up the majority of expo demonstrators. This does not bode well for American ambitions in the field, although there are some interesting designs.
Probably the most impressive robot on display was Atlas from Boston Dynamics, but you’d expect that: the project has been under development for over a dozen years now. The Atlas robot, originally funded by the US military’s research arm, the Defense Advanced Research Projects Agency (DARPA), was designed to create a machine that could operate in environments designed for humans.
Previous models of Atlas were hydraulically powered, but the business is now demoing its electronic version. And the demo video looks damn good. Judged from this alone, Amazon bosses will be wanting these machines in its warehouses pronto, since they look able to sort and stack equipment 24/7. All without the need of those pesky human workers, with their need to eat, relieve themselves in bottles to meet quotas, and rest occasionally. What a shame, then, when BD admitted that the Atlas demo was controlled by a human operator this year.
Worse still came from South Korean manufacturer LG. Its robots failed on stage to fold a tea towel correctly, even with human help. One towel badly folded in a minute suggests we are far from a Terminator situation, or even a useful home helpmate. It’s not helped by the fact that the chaebol’s robot looks like a template for Marvin the Paranoid Android. No release date has been scheduled as yet.
Atlas is due to be augmented with Google’s Gemini AI and will apparently go to work at Hyundai’s US mega-car manufacturing site in 2028. Meanwhile, at Davos, Elon Musk predicted his firm’s Optimus robots might be functioning in factories by next year – but this is an Elon deadline, so don’t hold your breath.
While robots might be gaining from AI, the same isn’t looking likely for human workers. At the end of the month Anthropic’s CEO Dario Amodei published an extended essay predicting that 50% of white collar entry-level positions will be wiped out within five years, and probably sooner than that.
“In the short term, being creative about ways to reassign employees within companies may be a promising way to stave off the need for layoffs,” he wrote.
“In the long term, in a world with enormous total wealth, in which many companies increase greatly in value due to increased productivity and capital concentration, it may be feasible to pay human employees even long after they are no longer providing economic value in the traditional sense. Anthropic is currently considering a range of possible pathways for our own employees that we will share in the near future.”
Now Amodei has a product to sell, so this kind of message isn’t unsurprising. And analyst firms like Forrester are happy to reinforce that, predicting this month that 10.4 million jobs would be replaced by AI by 2030. That’s more than were lost in the Great Recession of 2008, and Forrester added to the gloom with by noting that, unlike in that case, those jobs won’t be coming back.
But others disagree. Deloitte reported this month that only one in five of IT bosses it talked to have seen any meaningful return on investment on AI spending. Fellow analysts at PwC put the figure even lower, with 52% saying they’d seen no cost savings or revenue growth. And pollsters at Gallup found that AI use in the workplace flatlined in the last quarter of 2025.
These are not the figures that the AI industry wants to see. Revenue might be growing at AI companies but profits, or even a diminution of losses, are not yet apparent. Maybe the industry doomsayer Ed Zitron might turn out to be right after all.
We’ll know a lot more in the coming weeks on how the so-called Magnificent Seven tech firms – Alphabet (Google), Amazon, Apple, Meta Platforms, Microsoft, Nvidia, and Tesla – propping up the US economy are getting a return on investment.
On Wednesday Meta’s earnings call the social network predicted it would spend between $115 and $135 billion on AI capital expenditure, with Microsoft reporting on the same day it will be investing a more modest $37.5 billion on capex. The markets rewarded Meta with a share price rise, backed largely by advertising sales growth, while Microsoft’s fell sharply, not helped by less-than-stellar Azure sales. But there’s no sign yet that the market is comfortable with huge amounts of money being spent on AI infrastructure.
Both insisted that AI was becoming increasingly integrated on their platforms, which it is, but Microsoft said only 15 million customers were actually paying for its M365 Copilot product, and around five million were buying the GitHub AI product. Meta was more coy but said its Meta Superintelligence Labs was making significant progress.
Apple isn’t spending much on AI, instead buying in Google’s Gemini system. This more conservative approach seemed welcome on Wall Street at its reports call on Thursday, helped by stronger than expected iPhone sales. These might be hit, however, by increased memory costs as AI buyers try to secure components Cook & Co need.
Tesla, as is often the case, was the wild card. In its earning call it said it was transitioning from a “hardware-centric business to a physical AI company,” and said it was cancelling the Model S and X electric cars and converting the factories to building its Optimus robot. Quite when this robot will be ready for service remains to be seen. There are also rumours that xAI and SpaceX will merge – not a bad strategy since SpaceX is America’s only reliable orbit business and sales of Tesla’s cars are tanking.
Iain Thomson will be writing a monthly “View from the Valley” article for TechFinitive. To make sure you never miss it, sign up to our newsletter.
