CES 2026 – Dark Horses of Vegas

• The next edition of the traditional post-CES report highlights interesting things and trends that the tech media ignored or missed.

Agility Robotics – Actually making sales.

• Agility Robotics offers a robot for enterprises and factories, and while it may look like a horribly constructed Lego set, crucially it works, and it sells.
• It looks odd because the robot has been designed purely for its function, which is to perform a large range of tasks and replace a human or humans in the factory.
• The company is also unique in that it is realistic and upfront in terms of what robots are capable of here and now.
• This is a strong contrast to all the other humanoid robot makers who market themselves as being able to offer a fully functioning domestic servant in the near future.
• An Agility customer needs to train the robot for about three months to do the task and input is required from Agility to help the customer do this.
• This is the most realistic assessment of what robots can do now, which is why I suspect Agility is actually selling bots, whereas the robot that people queued 30 minutes to see (Atlas) won’t do anything before 2028 and maybe not anything then.
• Hence, Agility is effectively first to market but will need to keep innovating to fend off the hordes that are trying to follow it.

SoundHound – All the right boxes.

• SoundHound is an AI company that I have followed at a distance for many years that finally looks like it is coming into its own.
• The company offers AI agents and has a particular focus on the automotive industry.
• What is unique about SoundHound is that the products that it has created are exactly what I think an OEM needs to create a good, voice-based user experience in the vehicle.
• It has also created a developer kit that either the customer can use to create an agent or it can let SoundHound create it on its behalf.
• This is very like the AgentBuilder kit that OpenAI launched a few months ago.
• This, combined with some real-world traction in the automotive industry and an increasing number of partnerships, puts SoundHound in a position to do well.
• This is especially the case as AI in automotive is one of the use cases that RFM has identified as making the most difference to the user and so most likely to take off.

ProtoArc – Mixing travel and health.

• Buried deep in the bowels of the South Hall was ProtoArc, which is a Texas / China-based company that makes peripherals for PCs and handheld electronic devices focused on ergonomics.
• The company also has a product line aimed at travel, which includes foldable mice, keyboards, laptop stands and extra monitors, also aimed at health and ergonomics.
• As a frequent traveller, I have long watched this subsegment, and ProtoArc is the only company I have found that is addressing this market.
• Bad necks, backs and tendon issues are all complaints that crop up after spending years hunched over a keyboard, and with home working and flexible working becoming more common, I suspect that this problem is going to grow.
• ProtoArc’s approach is quite innovative, and I can see myself using some of their products in the future when they are a bit more mature.
• This is one to watch as I think the segment will grow, and there is almost nobody else addressing it.

Healthcare – Many claims. Zero delivery so far.

• From a business perspective, the holy grail of a wellness wearable is to be able to non-invasively and highly accurately measure blood pressure and blood glucose.
• Each of these on its own would represent a 10s of billions market opportunity, and many players have been trying for many years to get it to work.
• Even the big companies with almost limitless resources have tried and failed, meaning that this is a very tough nut to crack.
• 6 years ago, there were lots of start-ups trying, and in 2026, I have only found 2.
• These are Charmcare from Korea, which says it will launch the non-invasive H2-CBP in H2 2026 and Actxa from Singapore that has a smart ring that it says will be able to detect blood glucose accurately enough to be of use to a diabetic.
• I have been digging up companies that make these sorts of claims for over 10 years, and so far no joy.
• Hence, I am deeply sceptical that these will work, but at the same time hopeful.

Biggest disappointment – Food Tech.

• The food tech section of the Venetian was by far the most disappointing part of CES 2026.
• The section was tiny and populated with exactly the same companies as last year, all of whom are selling the same product that has not changed in 12 months.
• Furthermore, none of these offerings were anything that I would ever want to buy, and most of them looked like scrappy science projects.
• Companies with real food tech like Thermoworks, Ninja and Thermomix were absent, and I suspect they put their efforts into other trade shows aimed directly at food.
• Even though I like to cook, I don’t think I will bother with this section next year.