CES首发|获全球700+外媒重磅报道!原力无限以“具身大脑”定义下一代具身智能
原力无限

全球霸屏!

原力无限获700+外媒报道,触达全球近2亿受众!

拉斯维加斯当地时间1月9日,随着原力无限(INFIFORCE)在CES核心Panel的亮相,关于“具身大脑”的讨论迅速席卷全球科技圈。

据美通社官方最新数据显示:截止发稿前,此次传播触达全球近2亿受众,核心文章深度阅读量突破6.7万+,包括Yahoo Finance、BusinessToday、Morningstar等在内的700+全球主流媒体对原力无限进行了重磅报道。

*部分媒体报道截取

CES 2026,在官方核心论坛 "Not Quite Human" 的聚光灯下,原力无限(INFIFORCE)创始人、CEO白惠源(Isabella Bai)博士作为全场唯一中国具身智能领域企业家,向全球传递了一个信号:在具身智能的深水区,中国企业正跨越“制造”的边界,开始掌舵“定义”的航向。

原力无限给出的答案是:机器人的未来,不在于更像人的躯壳,而在于理解物理世界的“大脑”。

01

共振:AI下半场的“物理时刻”

2026年的拉斯维加斯,空气中弥漫着“硅基生命”觉醒的气息。

NVIDIA创始人黄仁勋在演讲中,将“Physical AI”定义为AI的下一波浪潮:“AI正在从屏幕中的数字助手,进化为物理世界中的机器人。” 与此同时,AMD、Tesla等科技巨头纷纷加码机器人算力与硬件,具身智能(Embodied AI)已毫无悬念地成为全球科技的新高地。

这与原力无限(INFIFORCE)的战略判断不谋而合。

“具身智能是AI的下半场。” 原力无限创始人兼CEO白惠源(Isabella Bai)在CES现场指出。大模型已经征服了虚拟世界的文本与图像,而接下来的关键跨越,是让AI进入真实的物理世界,去感知、去决策、去执行。

这不仅是技术的延伸,更是维度的跃迁。

 

02

“躯体”已就绪,“大脑”仍缺席

在这一跨越中,行业面临着显著的非对称发展。

纵观本届CES,人形机器人硬件层出不穷,关节电机、精密减速器等产业链已高度成熟。然而,白惠源在Panel上的一席话,道出了行业的隐忧:“The industry is 'Body-Ready' but 'Brain-Weak'.”

缺乏对物理规律深层理解的机器人,即便躯体再逼真,也难以应对非结构化的复杂环境。只有突破“大脑”的瓶颈,让机器人足够聪明,具身智能产业才能迎来真正的爆发。

基于此,原力无限确立了“一脑多身多场景”的战略核心,致力于构建一个通用的“具身大脑”,让智能成为驱动所有形态机器人的核心引擎。

 

03

Hyper-VLA与因果世界模型的融合

在CES的全球舞台上,原力无限展示了支撑这一战略的底层逻辑——全栈自研的持续学习Hyper-VLA端到端多模态模型与因果世界模型(Causal World Model)。

这是一条少有人走、却通向本质的技术路径:

Hyper-VLA架构赋予了机器人多模态的感知与持续学习(Continuous Learning)能力;

因果世界模型则让机器人理解了物理世界的“Why”——理解力的作用、物体的属性与行为的后果。

不同于其他主流的纯End-to-End(端到端)路线,原力无限的技术方案有效解决了AI在物理世界中的“幻觉”难题,确保了机器人在复杂环境中的鲁棒性与泛化能力——这一架构的成功验证,标志着中国科技力量在具身智能领域,已具备了底层技术路线的定义权。

CES官方核心Panel的席位,向来是全球科技风向的标尺。

原力无限之所以能与全球顶尖科技领袖同台,并非偶然。这是国际视野对原力无限“技术前瞻性”与“商业落地能力”的双重确认。 凭借“一脑多身”的策略,我们已斩获超5亿元的商业订单,证明了“具身大脑”不仅是实验室的构想,更是产业变革的现实推力。

正如白惠源在Panel结束时所言:“Robots handle the Logic, We handle the Magic.”

当“大脑”足够强大,机器人将不再是工具,而是人类探索物理世界的智能伙伴。

而这,也正是原力无限在CES 2026带来的未来图景。

以下为本次外媒报道原文内容

INFIFORCE's Isabella Bai at CES 2026: China is Moving from Manufacturing to Defining Next-Gen Embodied Intelligence

LAS VEGAS, Jan. 9, 2026 /PRNewswire/ -- Isabella Bai, Founder and CEO of INFIFORCE, has pointed out the core bottleneck of AI industry at the "Not Quite Human: How Humanoids are Changing Work and Home Life" panel during CES 2026, stating that now the hardware is ready but the "humanoid forms" are missing the brain and intelligence, and Chinese tech companies are transitioning from "manufacturing" to "cognitive definition" in the AI era.
INFIFORCE has showcased its proprietary "Causal World Model" and "Hyper-VLA (Vision-Language-Action) Foundation Model" at CES 2026, which tackles the challenges of giving robots the intelligent brain by establishing a complete closed loop of "Perception-Cognition-Decision-Execution," breaking the boundaries between the virtual and the real to empower robots with a "Super Brain" and the true ability to understand the world, engage in continuous learning, and act autonomously. 
This strategy of deploying one unified "Brain" into diverse bodies has secured the company CNY500 million (USD71.55 million) worth of commercial orders across automotive and healthcare sectors.
"At CES, we see incredible mechanical engineering, but a robot without a causal understanding of the world is just a sophisticated puppet," said Isabella Bai, Founder and CEO of INFIFORCE. "Our 'Hyper-VLA' architecture gives robots the ability to think before they act. By separating the evolution of the 'Brain' from the 'Body,' we have enabled our FORCE and AstroDroid robots to handle complex tasks today, not ten years from now. The CNY500 million in orders we hold is proof that the market craves intelligence, not just novelty."
The "Causal World Model" allows robots to predict outcomes and understand physical laws, a missing link to solving the "hallucination" and safety issues in the current AI robotics development. At CES 2026, INFIFORCE also showcased its product matrix:
AstroDroid: a general-purpose humanoid robot series powered by Jetson AGX Orin for robust computational performance and fully integrated with the Hyper-VLA. It not only possesses dexterous upper limbs comparable to humans but also excels in efficient mobile operational capabilities. The flagship model AD-01 stands 170cm tall with 43 degrees of freedom, and the series is expanding to cover more sizes and models to meet needs for various general application scenarios.
FORCE: fully autonomous mobile charging robot, that introduces innovative active charging paradigm of "robot finds vehicle," revolutionizing the traditional "user finds charger" model. It's equipped with L4-level autonomous driving and multi-robot collaborative scheduling capabilities and is already deployed at scale by leading enterprises such as Sinopec, Wanda, and Intime.
Hyper-Components: core execution unit and perception module for embodied intelligence, features HyperArm, the lightweight high-degree-of-freedom bionic arm. INFIFORCE's self-developed core components ensure high-precision execution of "brain" commands in the physical world, achieving ultimate performance with integrated hardware and software.
Founded in 2023, INFIFORCE is a leading AI-driven innovation enterprise dedicated to Embodied AI. Headquartered in Hangzhou with R&D centers in Shenzhen and Beijing, the company brings together executives from Global 500 firms and a top-tier scientific team, with over 75 percent holding master's or PhD degrees.
As of now, INFIFORCE has accumulated more than 35 invention and utility model patents, along with 20 software copyrights. Additionally, over 30 core technology patents are currently under application, with planned patent exceeding 100 in the future.
"The biggest barrier we need to break is Data, not internet text, but real-world physical data. At INFIFORCE, we built a Data-Model-Body-Scenario Growth Flywheel, with more Scenarios give us more data, which makes the Model smarter, and that then creates better Data. We have the Brain, we have the Scenes, so we can break the barrier," noted Isabella Bai. "Robots handle the Logic. We handle the Magic."