The Consumer Electronics Show (CES) in Las Vegas this week has marked a definitive shift in the artificial intelligence landscape. Just four years after ChatGPT's debut, the focus has moved decisively from pure software to "physical AI"—robots, autonomous vehicles, and smart gadgets that interact with the real world. The annual tech extravaganza saw major global companies and ambitious startups alike showcasing how AI is being embedded into hardware.
The Rise of Physical AI and Humanoid Hurdles
This year's event underscored a major industry transition. While big tech firms presented blueprints for future smart homes, numerous entrepreneurs displayed AI-powered devices for tasks ranging from haircuts to offering companionship. Interest in humanoid robots has boomed, with companies viewing them as the next frontier. Exhibits from South Korea's LG and others featured robots playing poker, crafting paper pinwheels, and even dancing.
However, the slow, deliberate pace of these robots highlighted significant development challenges. Experts point to issues with processing power, battery life, and handling unprogrammed situations. Consequently, a viable and affordable consumer humanoid is not expected to hit the market soon. Arm, the chip technology firm, recognizing this growing market, exclusively told Reuters it has reorganized to create a dedicated physical AI unit to expand in robotics, as noted by Chris Bergey, head of its client business.
AI Infuses Everything: From PCs to Pocket Pets
The push for AI integration was ubiquitous. Lenovo, the world's largest PC maker, hosted a flashy event featuring Nvidia's Jensen Huang and AMD's Lisa Su, unveiling its Qira AI voice assistant platform. Meta enhanced its Ray-Ban smart glasses and Neural Band with new features like live recording. Google launched its Gemini AI model for TVs and home devices.
The spectrum of AI applications was vast. Alongside serious tech, more whimsical gadgets appeared, including 3D avatars programmed to offer sweet talk and AI-powered pocket pets that evolve with their owner. More utilitarian showcases included a $599 AI dry cleaner that identifies fabric and cleans in three minutes, and AI-assisted hair clippers. After seeing AI-enhanced lawnmowers and mattresses, Seaport Research analyst Jay Goldberg noted many were formerly called "smart" devices, suggesting the AI label is now a powerful marketing tactic.
The Crucial Engine: Chip Advancements for On-Device AI
Behind the flashy exhibits, a critical technical shift is underway. While most AI currently runs in the cloud, the high cost is driving a move towards on-device processing. This requires chips with specialized AI functions. "AI is getting really expensive," remarked Perplexity AI CEO Aravind Srinivas at an Intel event, echoing a widespread concern. Intel's Jim Johnson also highlighted the soaring cloud computing costs for companies.
In response, chipmakers unveiled new products designed for this future. Intel launched its Panther Lake AI chip for laptops, its first using the advanced 18A manufacturing process. Rival AMD announced a new set of processors for AI PCs. These chips promise better speed, battery life, and security. However, a key question remains: will consumers buy in? Ben Bajarin, CEO of Creative Strategies, told Reuters that the average consumer is still unclear about what an AI PC can uniquely do for them, a gap the industry must bridge.
The message from CES 2024 is clear: AI is becoming tangible. The race is on to build the hardware—and the silicon brains inside—that can power this new physical world of intelligence.
