For me, CES 2025 will be marked forever by the Los Angeles fires, which sent many of us scrambling home with the show half over to experience the terrifying firestorm with loved ones. Today, the power is back on, the air is still, and the sky blue, a stark contrast to the destruction just five miles east of us. The winds will be back next week. In the meantime, I’ve managed to write up my hot take on this year’s edition of the world’s biggest trade show, which was attended by 140,000 people in 2024. It didn’t feel that big this year, but official numbers won’t be available for some time.
This year’s CES was all about AI—or at least, that was the headline. Everywhere you looked, booths and panels touted the transformative potential of artificial intelligence. Yet for all the talk, there wasn’t much to see. Unlike XR, which dazzles with visual demonstrations, AI is invisible. Exhibitors leaned heavily on last year’s devices, repackaged as “AI-enabled.” Demonstrating AI’s full potential in real time was hampered by bandwidth limitations inside the convention center.
AI Everywhere – And Nowhere
This paradox—AI being both everywhere and nowhere—defined CES 2025. Sure, there were smart gadgets, predictive algorithms, and personalized assistants sprinkled throughout the show, but few felt groundbreaking. It’s clear we’re in the early days of consumer AI, where marketing outpaces practical deployment.
The undisputed highlight of the show was Nvidia CEO Jensen Huang’s keynote. Huang painted a sweeping vision of AI’s role across industries, from robotics to healthcare and automotive. His message was clear: Nvidia is not just a chipmaker; it’s the king of the AI revolution. Huang showcased Nvidia’s AI-powered robotics platform, emphasizing how it could reshape industries like logistics and manufacturing. For all the hype, the exhibition floor offered little evidence of anything but a self-declared revolution.
Examples of consumer AI robots included auto vacuums, lawn mowers, security bots loaded with sensors (but less efficient and much more expensive than an actual dog), and robot toys for kids.
The Big Boys Take Up A Lot of Space
The show floor in the LVCC was dominated by all the big consumer electronics companies: Sony, TCL, LG, and Panasonic all had enormous booths, showing off their diverse product lines, from smartphones to laptops with multiple screens.
Sony and Honda unveiled their auto collaboration, AFEELA, which we kind of got to demo. It’s $90,000, very cool, and packed with technology that has nothing to do with driving. LG was showing off TVs, including transparent ones (which they also demoed last year). Samsung’s theme was “AI for All,” and TCL’s was “Inspire Greatness,” which surely a cheap HD TV will do for anyone.
XR Isn’t The New Kid Anymore, But It Hasn’t Grown Up
The XR area of CES also reflected a quieter year for the sector. The most successful product in the space, Meta’s AI-enabled Ray-Bans, weren’t even in the same building with the other XR in Central Hall. Many players that once staked their claim in the XR space were conspicuously absent. In their place, Xreal dominated the conversation with its sleek glasses and 3 DOF (degrees of freedom) screen extenders. Rokid maintained a solid presence as well, but the real innovation was hard to spot.
Sony launched its XYN (/zin/) VR headset, an integrated software and hardware solution designed to facilitate the creation of spatial content, including 3D graphics for applications like the metaverse. Developed by Sony’s XR Business Development Division, XYN comprises three components: mobile motion capture sensors, an XR head-mounted display, and a spatial reality display. The new headset will retail for $4,750.
There were smart glasses everywhere that were knockoffs of Meta’s Ray-Bans, with the primary innovation being the integration of ChatGPT instead of Meta’s Llama models. While these glasses offer the same audio features and cameras, they don’t offer the quality, fashion, comfort and ease-of-use of the plug-and-play Ray Bans. They’re just fifty bucks less.
Halliday introduced its new smart glasses which feature a monocromatic monocular microdisplay in the frames, that can show notifications, or maps. This left me scratching my head. Monocular AR displays may make sense for a company like Everysight, whose $400 monochromatic AR glasses are a popular BMW motorcycle accessory.
The vision of AR glasses overlaying data on the real world in a way people have to have has proven elusive, both technically and in terms of consumer adoption. That’s why companies like Xreal and Rokid are focusing on screen-extender functionality, allowing users to project personal screens that feel like a 300-inch TV. Xreal’s 5 DOF model is a step forward in this space, offering enhanced immersion. But the question remains: Do people really want this? All day every day smart glasses like Meta’s Ray-Ban feels like it’s happening without the display. Most consumers already have large, high-resolution devices in their pocket. For most consumers under 30, it’s their main screen.
Hope for the Blind
Most people don’t know this, but my niece is legally blind, so I’m always on the lookout for companies that use wearable and spatial technology like computer vision to help the blind and partially sighted to see. Ocutrx was at the show, as they were last year, showing tw set o new devices. Using proprietary software running on a Qualcomm XR2 chip, OcuLenz provides vision correction to patients with Advanced Macular Degeneration (AMD), the leading cause of vision loss in adults over 60, and the ORLenz 5K XR headset displays 3D holographic images and drastically improves ergonomics by giving surgeons freedom of movement and positioning.
Conel Amariei was back at CES with his Lumen headset, which uses lidar sensors to activate vibrations inside a haptic headband to silently, invisibly guide the blind without the need for a cane or guide dog. I also sought out Soliddd, whose CEO I interviewed for an article on accessibility last year. They, too, are making a headset to help the brain “see around” blind spots caused by AMD.
I saw the reincarnation of the Wii in compact form: Playground, which was giving private demos in a Venetian hotel suite. That was, seriously, my best of CES, not to throw shade at Xreal or anyone else that impressed me.
A Future Under Construction
Amid the noise, the broader message of CES 2025 was clear: AI is the new frontier, but its road to consumer impact is still under construction. It’s easy to imagine a future where AI reshapes how we work, play, and interact with technology, but this year’s CES demonstrated how far we are from that vision.
For XR, the path forward is similarly murky. The technology remains niche, with few standout use cases beyond gaming and screen-extension. While Xreal and Rokid are pushing boundaries, the lack of broader industry momentum raises questions about the sector’s future beyond AI-enabled smartglasses, which may be the ulimate form factor, because AI, the web, and location awareness is all the functionality you need, and perhaps want, in a pair of glasses.
CES 2025 was a show of promises, not breakthroughs. AI is poised to change everything, but its presence at the show was more theoretical than tangible. XR continues to struggle for relevance, with flashes of innovation drowned out by a sea of sameness. For both sectors, the next year will be critical. How will AI manifest its transformative potential? Can XR find its killer app?
At this year’s CES, the answers were unclear.
