U.S. CES Opens, 14 Key Statements

Article: Ba Jiu Ling, Wu Xiaobo Channel

Las Vegas, the plane has landed. In the first week of the new year, many are heading to CES, the annual tech gala has kicked off.

The focus these days is on the people and what they say.

On January 4 and 5, the days before the official start of the exhibition, is CES Media Day, when many companies choose to release new products. As in previous years, these two days almost turn into a “Summit of Leaders” for top global AI companies: Nvidia founder Jensen Huang, AMD Chairman Su Zifeng, Intel CEO Chen Liwu, Lenovo CEO Yang Yuanqing, Qualcomm CEO Cristiano Amon… all appear and deliver speeches.

But these speeches are not just about new product launches; they are more like statements on corporate technological paths and industry layouts. Every judgment is dissected and pondered by the world.

Senior tech media personality Zhuang Minghao commented that CES is originally a consumer electronics show, but looking at Nvidia and AMD, especially Jensen Huang’s speeches, they have essentially no content targeting the “consumer market.” Instead, they focus entirely on the GPU needs of data centers under the current AI narrative, and on grand themes like the “limits” of the physical world. It seems human consumer goods are no longer as important.

Over the next four days, as the main stage lights gradually dim and the exhibition halls open one after another, the official CES exhibition period has just begun. Reports indicate that over 2.5 million square feet of exhibition space attracted more than 4,000 exhibitors and hundreds of thousands of attendees worldwide. The hottest tech products—robots, AI PCs, smart glasses—will be showcased here.

CES 2026 Opening

The presence of Chinese companies is equally dense. Old brands like Lenovo, Hisense, TCL, and newer robotics firms like Yushutec, Zhiyuan, Yunshedu, are present. In niche fields like floor scrubbers, lawnmowers, stair-climbing robots, pool cleaners, Chinese manufacturers are also active.

As Jensen Huang repeatedly emphasized in his speech: “The AI race has already begun, and everyone is striving to reach the next level.”

So, how do the leaders of major AI companies interpret industry trends and depict their future blueprints during the multiple keynote speeches on the first day? According to Du Yu, director of the Unknowable AI Institute:

Jensen Huang’s core logic is that AI must understand common sense about the physical world to truly interact with reality. From a business perspective, the ceiling of the real world is higher than that of the online world.

Jensen Huang dressed in crocodile skin this time.

Intel emphasizes hybrid AI and edge computing. Essentially, many scenarios in the real world require edge AI, such as healthcare, finance, industry, emphasizing data privacy, low latency, and zero disconnection.

Su Zifeng from AMD said that in the next few years, computing power needs to increase by 100 times, fundamentally solving the “computing power shortage.” AMD’s strategy is to use higher cost-performance computing to seize Nvidia’s data center market.

Invited by Su Zifeng, “AI godmother” Fei-Fei Li reiterated the natural limitation that “large language models are ultimately constrained by language itself.” She believes language is a tool to describe the world, not the world itself.

From physical AI, hybrid AI, edge AI, to computing power, spatial intelligence, AI agents… these keynote themes, old and new concepts alike, sketch a panoramic view of future AI development. We’ve compiled fourteen golden quotes related to future AI, leaving a small timestamp for this rapidly evolving era, to be validated in the future. We also invited experts in related fields to share their insights on these revelations.

Fourteen Golden Quotes on Future AI Development

1. “The ChatGPT moment for physical AI is imminent.” — Jensen Huang

2. “Breakthroughs in physical AI will take AI from screens into our physical world—timely, as the world is building all kinds of factories for chips, computers, life-saving drugs, and AI. With global labor shortages worsening, we need automation driven by physical AI and robotics more than ever.” — Jensen Huang

3. “Today I want to talk about a little AI pamayo—our work in autonomous driving—where we not only open-source models but also the data used to train them. Only then can you truly trust the source of the models. We open-source all models, helping you create derivatives from them.” — Jensen Huang

4. “If you look at the models of the world, OpenAI-generated tokens are more numerous than any other model, and the second largest group might be open-source models. My guess is that over time, open-source models could become number one.” — Jensen Huang

5. “We believe that as AI capabilities continue to improve, localized computing will become increasingly important. First, the higher the degree of localization, the lower the latency and the better the performance; second, more localized AI is safer and truly ‘yours’; third, each inference costs money, and centralized cloud computing faces cost and bandwidth bottlenecks. Localized computing can significantly reduce overall costs by decreasing transmission and infrastructure reliance; fourth, for enterprises, the core value of local computing is not performance but regaining control over data, intelligence, and authority.” — Sri Nivas, CEO of Perplexity

6. “The era of hybrid artificial intelligence begins… Local AI safely executes tasks, keeping data on the device, while cloud AI handles global reasoning, planning, and multi-agent orchestration.” — Jim, Senior Vice President of Intel Client Computing Group

7. “You will see hundreds of edge devices of various shapes and sizes in key areas like smart cities, factories, healthcare, and automation systems, with huge and growing demand.” — Jim, Senior Vice President of Intel Client Computing Group

8. “The number of AI users has skyrocketed from the initial 1 million to over 1 billion active users today… We expect active users to grow beyond 5 billion, as AI truly integrates into all aspects of our lives, just like smartphones and the internet today.” — Su Zifeng

9. “The computing power we currently have is far from enough to support everything AI can do… To make AI ubiquitous, we need to increase global computing capacity by 100 times in the next few years, or more than tenfold in the next five years.” — Su Zifeng

10. “In the future, a country’s GDP growth will largely depend on its available computing power.” — Greg, President of OpenAI

11. “What excites me is the emergence of a new generation of AI technologies, including embodied AI and generative AI. We are finally able to endow machines with capabilities closer to human levels—spatial intelligence.” — Fei-Fei Li

12. “We are moving from ‘passively understanding the world’ to ‘helping us interact with the world.’” — Fei-Fei Li

13. “Today’s AI assistants are mostly reactive agents—you open an app and ask questions for responses. But when AI runs quickly on devices and is always on, it can proactively perform tasks for you.” — Amit, CEO of Luma AI

14. “2026 will be the year of AI agents; AI will help you accomplish more tasks, even end-to-end, rather than just handling small fragments.” — Amit, CEO of Luma AI

Main Commentaries

Hu Yanping: Professor at Shanghai University, scholar in intelligent technology industry and intelligent economy

Jensen Huang’s statement that “the ChatGPT moment for physical AI is coming” can be more specifically positioned around 2026 as the GPT-3 to GPT-3.5 era—significant progress, but not overly optimistic, especially regarding “brain development” in robots. However, the mass production of L3 autonomous driving this year is quite certain and optimistic.

Additionally, AMD and Nvidia both emphasize full-stack AI capabilities from chips to hosts and clusters, with products spanning chip computing power, development environments, and vertical applications, focusing on multiple vertical scenarios rather than single points. Especially AMD, which is catching up fast, has achieved a complete AI layout from data centers to personal devices, with significant performance improvements.

Fei-Fei Li showcased the first commercial world model under World Labs, Marble, aiming to generate persistent, navigable, and consistent 3D worlds to enhance human creativity rather than replace humans, aligning with her previous “human-centered” AI development philosophy.

Fei-Fei Li’s speech

Next, I will focus on five key points at CES 2026: embodied intelligence like robots; smart devices like smart glasses; the L3 autonomous driving industry chain; training and inference computing architectures and changes in terminal and edge AI computing power; and the impact of model capability injection based on sensing algorithms on smart health.

The “AI China Chain” in the AI industry chain has already taken shape, and the “AI China Loop” in AI technology and applications has also closed. Chinese companies, while launching a dazzling array of innovative products, are also expected to provide more solutions to the world.

Du Yu: Director of the Unknowable AI Institute, PhD in Technology Economics from the Chinese Academy of Social Sciences

After listening to several speeches, my feeling is that the “computing power arms race” has already heated up. The three giants all emphasize computing power, but their paths are completely different: Nvidia pursues absolute performance, AMD seeks cost-effectiveness, and Intel aims for widespread edge deployment. This reminds me of the recent IPOs of the four domestic GPU giants, each with their own strengths. This also means that in the next 2-3 years, chip price wars and performance battles will be fierce, which is a major boon for startups and enterprise clients—costs will drop significantly.

China cannot catch up in chip manufacturing in the short term, but is very strong in AI hardware applications. Among the robots at CES, over half are Chinese companies—Yushutec humanoid robots, Zhiyuan robots, Trifo vacuum robots, all Chinese. Robots are just one typical example. This is because China’s market is large, scenarios are diverse, and iteration is fast. My advice to Chinese startups: focus on differentiated competition and long-termism.

Zhang Xiaorong: Director of the Deep Technology Research Institute

Jensen Huang’s speech reflects not just technological upgrades but a paradigm shift. Previously, AI was “keyboard-mouse interaction”; now, it’s “visual and language interaction.” The “physical AI” Huang defines is “action interaction.” Through Cosmos (learning physical laws from videos) and Newton engine (real-time physics computation), he aims to solve AI’s “hallucination” problem—making machines understand that “water is fluid, glass is fragile.” This is to enable AI to work safely in factories and homes, not just write poetry or draw pictures.

Both AMD and Intel avoid direct competition with Nvidia’s GPUs, choosing different approaches:

Su Zifeng highlighted AMD’s position in the AI industry chain. She emphasized the Helios system and Ryzen AI 400 series for PCs, signaling clearly: the competition for computing power is a long-term battle. AMD is ready to compete and focuses more on edge and cost-effectiveness. AMD’s strategy is more like a “pragmatic idealist,” emphasizing “breaking through the bottleneck of computing power,” essentially lowering the barrier to AI use. If costs don’t come down, small and medium enterprises can’t afford it, and the AI ecosystem will wither.

Su Zifeng introduces AMD Instinct MI455X GPU

Intel emphasizes “local computing,” which is actually a way to find a differentiated survival path for AI under Nvidia’s strong cloud monopoly. Intel is indeed fighting a “defensive battle,” but also a “battle to win”—they see a pain point: not all AI tasks need to go to the cloud. Privacy issues (like home surveillance), latency (like gaming response), and costs all demand compute to sink to the edge. Intel deploys NPU (Neural Processing Units) on PCs and edge devices, building the “capillaries” of AI in the era. If AI only exists in the cloud with giants, it’s unhealthy; Intel is trying to make AI truly “everywhere.”

The core concepts from these speeches—physical AI, local computing, breaking bottlenecks, spatial…—outline a complete path from “cloud brain” to “physical world.”

Based on CES 2026’s trends, I summarize future AI with three words: “Intelligent Agents, Embodied Intelligence, Technical Solutions.”

1. AI will emerge from “dialogue boxes”: future AI won’t just be Copilot (co-pilot), but Co-worker (colleague). We look forward to AI agents helping us book tickets, operate software, or even clean rooms with robots.

2. Explosion of hardware forms: 2026 will be the year of humanoid robots and AI-defined cars. We will see more forms of AI physical entities on factories and roads.

3. Cost reduction: With Nvidia’s Rubin architecture and AMD/Intel solutions, inference costs will plummet. This means more affordable, user-friendly AI applications will appear, not just expensive luxury items.

Liu Xingliang: Renowned digital economist, member of the Ministry of Industry and Information Technology’s Information and Communications Economy Expert Committee

In these speeches, physical AI, hybrid AI, spatial intelligence are not isolated but part of a co-evolving ecosystem. AMD and Nvidia provide nearly unlimited cloud computing power, driving models to become incredibly powerful. These capabilities are then distributed to our devices via Intel’s hybrid architecture and multimodal models envisioned by Fei-Fei Li, ultimately creating value in the real world through physical AI and robotics.

In this grand picture, Chinese companies are not bystanders but vital participants, expected to play key roles in:

Application innovation battlefield: China has the largest and most diverse application scenarios and market demands worldwide. In e-commerce, social media, mobile payments, smart cities, manufacturing, Chinese companies can combine leading global AI models and hardware with deep local insights to create world-class AI applications. For example, in supply chain AI, personalized recommendations, industrial automation, China has huge advantages.

Key part of the hardware industry chain: China holds a core position in global electronics manufacturing and supply chains. From server manufacturing, AI terminal devices (PCs, smartphones, robots) to data center construction, Chinese companies are indispensable in transforming advanced AI tech into tangible products and large-scale delivery.

Domain-specific tech innovators: In AI chip design (Huawei Ascend, Cambricon, etc.), autonomous driving (Baidu Apollo, Xiaopeng, Huawei Inside), robotics, Chinese firms have accumulated deep technical strength. Facing international competition, they can achieve differentiated breakthroughs in specific verticals and leverage the domestic market to scale.

Active contributors to open-source ecosystems: More Chinese tech companies embrace open source, contributing code, models (like DeepSeek), and datasets globally. This enhances China’s influence in the global tech community and benefits from collaboration.

The future depicted by CES 2026 is clear and exciting. Chinese companies should leverage their advantages in markets, supply chains, and application innovation—actively integrate into the global tech ecosystem while tackling core technologies. The future AI world will be a multipolar, collaborative, competitive stage, with Chinese companies destined to be key players on the stage.