According to Zhitong Finance APP, South Korean tech giant Samsung Electronics Co. is further expanding its presence in the field of Smart Phones, Televisions, and AI on the device side by launching a new AI enhancement functionality suite under the "Vision AI" brand for its high-end Smart Tv product line, accelerating Samsung's ambitious goal of "AI integration in everything."

In an official statement during the pre-event for the "Sunday in Las Vegas" International Consumer Electronics Show (CES), the company stated that the Smart Tv screens driven by its device-side AI technology will be able to perform "online AI searches" for information related to the displayed content, such as recognizing actors or specific products, translating in real-time, and generating personalized background and main screen images.

The built-in AI processor in the top-tier Samsung Smart Tv models will also analyze TV content to achieve AI-driven enhanced colors and contrast, optimizing audio, while also being able to quickly access the pages users wish to browse, the movies they want to watch, and television gatherings based on voice commands or text-based input commands. It can also quickly set up personalized Smart Tv features as instructed by the user, similar to the latest version of the Siri voice assistant that can process user commands at high speed—namely, Apple's Siri voice assistant embedded with ChatGPT.

SW Yong, head of Samsung's visual display business, stated in an official statement from Samsung: "Samsung believes that televisions are not one-way Consumer Electronics devices for passive consumption but are interactive smart partners that can adapt to all user needs."

This South Korean tech giant has been the world's largest Smart Phone manufacturer for nearly two decades and has recently stated it will collaborate deeply with American tech giants Microsoft (MSFT.US) and Alphabet (GOOGL.US) under Google to strive to expand the basic AI suite of its smartphone AI brand "Vision AI" through updates and more powerful AI functionalities provided by partners.

This move is the latest advancement in Samsung's planned ambition of "AI everywhere." This AI blueprint aims to comprehensively embed generative AI technology into its Smart Phones, Smart Tvs, Smart Watches, and a broader range of device-side Consumer Electronics, creating an AI product that covers all Consumer Electronics, which is to realize the ultimate vision of "AI integration in everything."

It is understood that at last year's CES, South Korea's tech giant Samsung Electronics held a globally anticipated major launch event for Edge AI at CES 2024, with the theme "All for AI: Connectivity in the Era of Artificial Intelligence." The latest products unveiled by Samsung are all related to Edge AI, including AI Televisions, AI Refrigerators, AI Washing Machines, AI Vacuum Cleaners, AI Laptops, and more.

Edge AI may be the core focus of CES 2025.

With the emergence of new Consumer Electronics such as AI PCs and AI Smart Phones that integrate AI in 2024, the trend of "AI + All" is becoming unstoppable. This year, Edge AI will also be a central focus of the International Consumer Electronics Show (CES), discussing how AI can seamlessly integrate into Edge Consumer Electronics products and how AI can deeply penetrate human daily life, especially the new products and development processes in the most significant area of Edge AI—"AI Humanoid Robots," drawing attention from global investors and tech fans.

Wall Street Analysts generally expect that NVIDIA will showcase its latest humanoid robot strategy and Jetson Thor platform, designed specifically for humanoid robots, at CES 2025, which will closely relate to NVIDIA's "Embodied AI" strategy.

Cristiano Amon, the CEO of Qualcomm (QCOM.US), the world's largest Smartphone Chip Company, recently unveiled the new version of the Snapdragon processor. In a recent interview, he stated that AI Smart Phones—that is, smartphones capable of running powerful AI large models—are expected to become popular among every consumer globally within five years.

When the deployment scale of AI Chips in data center servers for processing massive amounts of parallel computing meets the baseline computing power needs of global enterprises and individuals, as well as the huge computing power support required for AI large model iterations, according to technological development trends, similar AI large models with increasingly powerful reasoning performance will ultimately integrate into application terminals such as Consumer Electronics, including Smart Phones, PCs, and humanoid robots, as well as software systems for Electric Autos and industrial production. This is known as Edge AI.

Currently, leading global Smartphone Hardware manufacturers are embracing the groundbreaking technology of AI large models and are integrating offline-capable, high-performing AI large models into a new generation of devices to create what is called "AI Smart Phones." Compared to general cloud-based AI large models that purely rely on cloud computing power, having local AI large models at the core while leveraging cloud AI allows Smart Phone users to more quickly and efficiently use products like ChatGPT and, with Edge AI, aim to realize a more personalized "Private AI Assistant" that resembles the all-encompassing AI companion in the movie "HER."

To enable AI large models to run on Smart Phones, some Smart Phone hardware manufacturers are exploring the integration of dedicated AI processors, designed specifically for efficiently handling AI tasks, striving to improve computing efficiency and energy consumption ratios. Manufacturers also aim to combine Cloud Computing and Edge Computing, whereby certain complex AI processing is completed in the cloud while real-time or sensitive tasks are processed locally. They are also exploring advanced technologies such as model pruning, quantization, knowledge distillation, and Lightweight architectures (like MobileNet and TinyBERT) to reduce the actual scale and computing needs of large models, making it possible to run powerful AI large models smoothly on Smart Phones, PCs, and other application devices with significantly limited training/inference computing resources.