CES 2026 Was All About Robots Doing Physical Tasks

What better way to start the year than experiencing all the amazing exhibitions at CES 2026 - to say it was mind-blowing is an understatement. The whole X timeline was filled with robots doing incredible things, from dancing to playing chess and even doing laundry. It's beginning to look like this is the year robots become the new normal in every home. Oh, and the most shocking of all, we saw an AI-powered refrigerator and guess what? A Caterpillar with AI integration. How nice!

OpenAI has just rolled out a major push into healthcare with two new offerings aimed at both everyday people and medical professionals. The company says over 230 million people globally ask health and wellness questions on ChatGPT every week, so they're building products specifically designed for this massive demand.

The first product, ChatGPT Health, is for regular users who want help understanding their medical information. You can now connect your medical records, Apple Health data, and wellness apps like MyFitnessPal directly to ChatGPT. The idea is to help you make sense of scattered health information that's usually spread across different portals, apps, and doctor's notes. OpenAI is clear that this isn't meant to diagnose or treat you, and it won't replace your doctor. Instead, it's supposed to help you understand lab results, prepare questions before appointments, or figure out what your wearable data might mean. All your health conversations happen in a separate, protected space within ChatGPT, and OpenAI promises they won't use this data to train their models.

The second product is ChatGPT for Healthcare, which is designed for doctors, nurses, and hospital staff. This version runs on OpenAI's GPT-5 models and is built to meet healthcare privacy requirements under HIPAA. Major hospitals like Boston Children's Hospital, Cedars-Sinai, Stanford Medicine Children's Health, and HCA Healthcare are already starting to use it. Medical professionals can use this tool to look up research, synthesize medical evidence, draft clinical documentation, and work with patient information securely. The responses include citations to peer-reviewed studies and clinical guidelines so doctors can quickly check sources.

The timing is significant because healthcare workers are overwhelmed with administrative tasks. Roughly 66% of physicians were already using AI in their practice in 2024, nearly double the 38% from 2023. Many doctors have been using their own tools because their organizations haven't adopted AI fast enough. OpenAI is positioning these products as a way to reduce that administrative burden and give medical staff more time with patients.

The tech world's biggest annual showcase just wrapped up in Las Vegas, and this year's Consumer Electronics Show had one clear theme running through almost everything on display: physical AI and robots.

Nvidia CEO Jensen Huang kicked things off with his usual flair, announcing the company's new Rubin computing architecture that's set to replace the current Blackwell chips starting later this year. But the real headline from Nvidia was Alpamayo, a family of open source AI models specifically built for autonomous vehicles. The company is making a big push to become what they're calling the Android of robotics, providing the underlying infrastructure that robot makers can build on top of.

AMD chair Lisa Su delivered the show's first keynote, unveiling the company's new Ryzen AI 400 Series processors aimed at bringing more AI capabilities directly into personal computers rather than relying on cloud processing.

Robotics were everywhere at the show. Boston Dynamics announced they're partnering with Google's DeepMind research lab to train their Atlas humanoid robots. Caterpillar teamed up with Nvidia to develop automated construction equipment, with a pilot program for an AI-assisted excavator demonstrated at the show. Ford also announced its new AI assistant, which will launch first in the company's mobile app before coming to vehicles in 2027.

Amazon used CES to expand Alexa's reach, launching Alexa.com so people can use the AI chatbot directly in their web browser, along with a revamped app. Ring announced updates including fire alerts, an app store for integrating third-party cameras, and new sensors.

One of the buzziest product reveals came from Clicks Technology, a startup bringing back BlackBerry vibes with its new Communicator phone. The device has a physical keyboard and sells for $499, with an additional $79 slide-out keyboard attachment that works with other phones.

Not everything at the show was practical. Razer introduced Project Motoko, which aims to work like smart glasses but without actually wearing glasses, and Project AVA, which puts an AI companion avatar on your desk. LG showcased its home robot CLOiD, though the live demonstration was underwhelming as the robot slowly moved a shirt to a dryer and placed a croissant into an oven.

Lego made its CES debut with Smart Play System, featuring bricks and tiles that can interact with each other and play sounds, with the first sets having a Star Wars theme. For creative types, the eufyMake E1 UV printer at $2,299 brings professional-grade customization technology within reach for small businesses and makers.

After a week of product announcements and robot demonstrations, the consensus is clear: the industry is betting heavily that AI's next big leap won't just be in software and chatbots, but in physical products that can actually do things in the real world.

Google just unveiled a major AI overhaul for Gmail that's designed to help you actually manage the chaos in your inbox instead of just storing it. The centerpiece is a new AI Inbox that gives you a personalized overview of what needs your attention and what's just informational noise.

The AI Inbox has two main sections. The first is called "Suggested to-dos" and it surfaces the emails that actually need you to do something, like reminding you that your electricity bill is due tomorrow or that you need to call your doctor's office to confirm your address for a prescription refill. The second section, "Topics to catch up on," groups updates by category so you can see at a glance that your Lululemon return is being processed, your shirt order was delivered, and your investment statement is ready, all without opening individual emails.

Google is clear that your regular inbox isn't going anywhere. This is just a new view you can toggle on and off when you want to cut through the clutter. The feature is currently rolling out to trusted testers before a broader launch in the coming months.

Gmail is also adding AI Overviews to its search function, which lets you ask questions in plain language instead of hunting through emails with keywords. For instance, you can ask "Who was the plumber that gave me a quote for the bathroom renovation last year?" and get a direct answer pulled from your emails with the key details highlighted. Google says the AI only searches your personal emails and doesn't use that data to train its models. This feature is rolling out to Google AI Pro and Ultra subscribers.

The company is also launching a new Proofread feature that works a lot like Grammarly. It analyzes your draft emails and offers one-click suggestions to improve clarity, word choice, and sentence structure. If you write something like "might inflict disturbance," Gmail will suggest changing it to "might disturb." It'll also catch instances where you've used the wrong word entirely, like "weather" instead of "whether." This is clearly Google's attempt to keep people from copying their emails into ChatGPT or relying on third-party proofreading tools.

Proofread is available to Google AI Pro and Ultra subscribers, but Google is making several previously paid features free for everyone. "Help Me Write," which can compose an email from a single prompt, AI summaries of long email threads, and suggested replies that match your tone are all now available to all Gmail users regardless of whether they pay for a subscription.

Caterpillar is partnering with Nvidia to bring AI and automation to its construction equipment, marking a notable shift for a company that's been making bulldozers and excavators for nearly a century.

The construction giant is piloting an AI system called Cat AI in its mid-size Cat 306 CR Mini Excavator, which was demonstrated at CES this week. The system runs on Nvidia's Jetson Thor physical AI platform and works like an AI assistant for machine operators. It can answer questions, pull up resources, offer safety tips, and schedule maintenance, all while the operator is actually doing the work.

Brandon Hootman, Caterpillar's vice president of data and AI, explained that this is particularly important because construction workers aren't sitting in front of laptops all day. They're in the dirt operating heavy machinery, so having AI assistance right there in the cab makes a real difference. The system can provide insights and help operators take action without having to stop what they're doing.

But the real value isn't just in answering questions. Caterpillar's machines are constantly sending data back to the company, roughly 2,000 messages every second. The company is using this massive stream of information to build digital twins of construction sites using Nvidia's Omniverse simulation tools. These virtual replicas let construction companies test different scheduling scenarios and calculate exactly how much material a project will need before breaking ground.

Caterpillar already runs fully autonomous vehicles in mining operations, and Hootman sees these AI pilot programs as a natural next step in expanding automation across the company's portfolio. The team chose to start with this particular excavator because it addresses a real challenge customers face today and because it's something they could bring to market relatively quickly. It also creates a technology foundation they can build on for other equipment.

The partnership fits squarely into Nvidia's broader push into what they call physical AI. While many people think of physical AI as just robots, Nvidia takes a much wider view. Deepu Talla, the company's vice president of robotics and edge AI, told reporters that physical AI is the next wave of AI and that everyone is essentially building robotics today, whether that's an autonomous car or a Caterpillar excavator.

During Nvidia's CES keynote earlier this week, the company laid out its plans for a full ecosystem supporting physical AI, including open models, simulation tools, and developer kits. The goal is to provide the complete infrastructure needed to train AI models, test them in simulation, and deploy them into physical machines.

Building AGI with our mission Intelligence with Everyone. Global leader in multi-modal models and AI-native products with over 150 million users.