CES 2025 has set the stage for deep discussions and numerous demos in the Edge AI area. With a focus on efficiency, privacy, and real-time processing, Edge AI redefines how we interact with smart devices and the broader IoT ecosystem.

By bringing processing closer to the data source, Edge AI minimizes latency, reduces bandwidth usage, and improves data privacy. It powers applications ranging from autonomous vehicles to smart home systems, wearable health tech, and industrial automation.

Thanks to Edge AI, robotics is experiencing growth. At CES 2025, companies showcased robots and machinery capable of detecting anomalies, predicting maintenance needs, and improving safety – all through local AI processing. Nvidia’s Project DIGITS was a standout example, introducing a compact AI supercomputer for researchers and data scientists. Powered by the GB10 Grace Blackwell Superchip, it enables sophisticated AI models to run locally without cloud reliance, offering real-time processing.

Such advancements in AI processors have accelerated new developments in robotics beyond traditional industrial uses. Humanoid robots and cobots equipped with Edge AI were a highlight at CES 2025, demonstrating human-machine interaction. Their ability to process and respond to data instantly makes them faster, more reliable, and safer in environments requiring quick interactions. We saw robots capable of real-time facial expression analysis, speech recognition, and natural language processing, all processed locally. There were AI-equipped autonomous systems for warehouse logistics, agriculture, housekeeping, delivery, and digital health. Visitors lined up to interact with bartender and cooking robots, along with various robot companions for emotional support.

AI-powered wearables stole the show this year. Their functionality ranges from entertainment and human-machine communication to real-time health monitoring and diagnostics.

Nearly a couple of dozen smart glasses were showcased, providing not only visually augmented reality but also AI assistants, live speech translation, and improved audio that dampens noise and amplifies the voice of a person one is talking with. Some models connect to smartphones and smartwatches, offering a hands-free experience; others are sports-tuned. Among them, Halliday Smart Glasses stand out with their rich AI functionality for advanced business users. Smart glasses have already become so-called composite AI devices using both predictive and generative AI. According to Gartner, by 2029, composite AI Edge devices will account for at least 60 percent of edge computing deployments.

This year, the focus was on human well-being and comfort, including improved product design that suits the end-user’s everyday life. PieX showcased its groundbreaking AI Emotion Tracking Pendant, which can analyze and interpret the wearer’s emotional state in real time. This wearable processes data on-device, ensuring user privacy while providing immediate feedback.

Several OTC hearing aids were disguised as attractive glasses to make users feel less conspicuous.  In addition to enhanced sound clarity, some novel solutions offered real-time translation across multiple languages and customizable tinnitus (ringing ears) relief.  

Some products we saw this year at CES were impressively innovative:

– Whispp’s Live Conversation AI technology converts whispered and vocal cord-impaired speech of people with severe voice disabilities into clear, natural speech without noticeable delay based on voice tract articulation only.

– Tone smartbuds by NextSense is a pair of brain-sensing smart earbuds equipped with EEG sensors that monitor brain activity in real time, providing precise assessments of sleep stages and depth. The earbuds utilize an AI-driven audio system that responds to the user’s brainwaves, delivering personalized audio cues to facilitate falling asleep and maintaining sleep throughout the night.

The Smart Home experience is transformed by Edge AI. Devices like AI-driven thermostats, security cameras, and voice assistants are now equipped with local processing capabilities. This enhances response times and ensures that sensitive conversations or video feeds never leave the home network.

Edge AI-based precise voice recognition allows users to control various smart home systems and access information through natural speech commands.

Gesture control is another area experiencing revitalization.  Neuranics showcased its novel magnetic sensor, which detects minute magnetic signals generated by muscle activity. This sensor facilitates real-time gesture recognition and seamless human-machine interaction.

Mudra Link neural gesture-control wristband by Wearable Devices received widespread acclaim at CES 2025.

Samsung’s SmartThings Ambient Sensing intuitively understands the user’s environment and situational context by analyzing human movements and ambient sounds through connected devices within the home. This enables those devices to intelligently and seamlessly respond to and adapt to daily routines.

This year’s CES provided further evidence that AI is beginning to become more fully integrated into end-user devices. AI-enabled gardening assistants,  dog collars, spice dispensers, bird feeders, and cat litter receptacles were displayed in Las Vegas.

All those innovations support enhanced privacy and security by keeping sensitive data local through consumer-focused applications like privacy-first smartphones with operating systems that aim to be more secure alternatives to Apple’s iOS and Google’s Android. The hardware solutions that process data on-device ensure users maintain complete control over their information.

Challenges and Opportunities

While the benefits of Edge AI are compelling, challenges remain. Processing AI workloads locally requires advanced chip designs with exceptional power efficiency. Also, interoperability between diverse and sometimes legacy devices and ecosystems is essential for widespread adoption.

Diversity in use cases slows down the market maturation for Edge AI solutions. Intel’s Dan Rodriguez, Corporate VP and GM of Edge Computing Group called it a use case fragmentation in the “AI at the Edge: Transforming Industries and the Workplace” discussion at CES. Proximity to customers with various power and application needs makes generalizing hardware and software requirements difficult. Edge AI, therefore, is all about application-specific hardware. That’s what we at POLYN believe.

The opportunities, however, are immense. CES 2025 made it clear Edge AI is no longer a niche technology. Its applications are rapidly expanding across consumer electronics, healthcare, automotive, and industrial sectors. As devices become more intelligent and more autonomous, the shift to localized processing will continue to gain momentum.