A single destination for timely, editor-curated robotics news from around the world.
Ant Group's robotics division, Robbyant, has made available the model weights and inference code for LingBot-Touch, an innovative AI model aimed at enhancing the tactile capabilities of robot hands. This model was developed using nearly 5,000 hours of data collected from three different types of tac...
Noitom Robotics has unveiled the HiPHI-Hand dataset on Hugging Face, consisting of 11 hours of egocentric stereo video. This dataset includes marker-based hand ground truth data, which NVIDIA utilized as a key resource for training its EgoHand model. The release of the HiPHI-Hand dataset is signifi...
Ant Group has announced the open-sourcing of its Robbyant tactile AI model, which enables robots to effectively process touch signals from various sensor types through a unified representation. This development is significant as it enhances the ability of robots to interpret tactile feedback, which...
The Huisi Kaiwu platform has announced a major update with the official release of the Hi-WBC full body control model. This advancement is expected to enhance real robot operations significantly, pushing the boundaries of robotics applications in various sectors. This release is crucial as it repre...
Researchers at MIT and the University of British Columbia have developed WorldParticle, a unified deep-learning system designed to simulate various physical phenomena, such as sliding sand and protein folding. This innovative model allows engineers to replace multiple specialized simulators with a s...
Baidu has announced a strategic pivot towards developing embodied excavators, leveraging its veteran model. This shift comes as the company seeks to expand its footprint in the robotics sector, particularly in construction automation. The move is significant as it highlights Baidu's commitment to in...
Baidu has announced that its Apollo Go service operates in 28 cities and has completed over 23 million rides as of June 2026. This significant milestone highlights the rapid growth and adoption of autonomous vehicle technology in urban areas. The expansion of Apollo Go is crucial as it demonstrates...
Chris Alexiuk, a senior product research engineer at NVIDIA, emphasizes the importance of open source AI, advocating for transparency in large language models (LLMs). He actively shares knowledge on platforms like YouTube, teaching developers how to utilize and implement these models effectively. Al...
LiberAI has unveiled Liber-0, a world action model pretrained on over 2.2 million hours of human experience, designed to facilitate zero-shot robot manipulation. In a test involving 60 tasks proposed by GPT-6, Liber-0 achieved an overall success rate of 36.0%, demonstrating its capability to perform...
On November 3, Baidu launched Apollo World, a world model designed to simulate interactive road environments for training and testing autonomous vehicles. This innovative system is currently utilized to enhance the development of Apollo Go's global robotaxi services, generating up to 100,000 kilomet...
Stanford University has initiated a project utilizing a 7-Tesla MRI to create a predictive model of the human brain. This advanced imaging technology allows for unprecedented detail in brain mapping, which could enhance our understanding of neural functions and disorders. The significance of this p...
South Korea's President Lee Jae Myung announced a historic budget for 2027, increasing total spending by 12.8% to 820.9 trillion won. A significant portion of this budget, 4.7 trillion won, is earmarked for developing a sovereign AI model aimed at achieving the world's highest performance. Additiona...
OpenAI has upgraded its four-layer monitoring system to better detect misaligned behavior during model training. This enhanced system employs lightweight classifiers and in-depth investigations to identify suspicious activities and evaluate potential security threats. Notably, OpenAI's improved moni...
AGIBOT introduced Worken, a new metric for assessing productive robot work, during its Innovation Day 2026 event in Shanghai on October 29. This initiative includes announcements related to humanoid hardware, AI models, industrial robots, and an open-source operating system expected by year-end. Wor...
NVIDIA has unveiled new models of its RTX Spark laptops, designed for developers, creators, and gamers. These laptops feature a powerful NVIDIA Blackwell RTX GPU combined with an NVIDIA Grace CPU, offering up to 128GB of unified memory and 1 petaflop of FP4 AI performance, making them suitable for v...
iFLYTEK has introduced the SparkAudio 1.0 speech foundation model, which integrates a 0.65 billion parameter audio encoder with a 30 billion parameter A3B mixture of experts model, specifically trained on domestic chips. Additionally, the company launched the Robot Super Brain Platform 3.0, which in...
A recent report by CAICT indicates that the prices of domestic large models nearly doubled in 2026. This significant increase in pricing coincided with a doubling of throughput, highlighting a growing demand for these models. Additionally, first-token latency saw an increase, while daily token usage...
Amap, in collaboration with Alibaba's ATH group, has introduced Lambda-0, a comprehensive humanoid foundation model. This model was developed using 500 hours of first-person human data, showcasing its capability to transfer knowledge zero-shot to the Unitree G1 robot. The release of Lambda-0 is sig...
Qianjue has introduced Polibrain, a predictive world model designed to anticipate task-relevant state changes rather than future video frames. This innovative model is adaptable for use across different robot body types, including quadrupeds, humanoids, and wheeled robots. The significance of Polib...
Eka Robotics has developed a new model called VFA (Vision-Force-Action) that surpasses OpenAI's VLA (Vision-Language-Action) model. Unlike VLA, which relies on extensive video and human demonstration data, VFA utilizes force feedback to enable robots to learn through tactile feedback, allowing them...
OpenAI has announced that its artificial intelligence models will now be accessible to developers via Google Cloud, marking a significant shift in the competitive landscape between the two companies. This move enables OpenAI to distribute its technology across all three major U.S. cloud platforms, i...
The embodied intelligence industry is facing significant waste due to the need for retraining models whenever a new robot is introduced. As the number of robot types could reach tens of thousands, the industry is exploring concepts like 'one brain, multiple forms' to enable shared intelligence among...
SenseTime has announced the open-source release of the SenseNova-RoboRSI evaluation harness along with a 27 billion parameter embodied AI model. This development follows a comprehensive study of the GPT-6-Astra, which was tested across 12 simulated benchmarks and 8 real robots. The significance of...
The 2026 Physics AI and World Model Conference is set to be held in Nanjing, marking a significant event as the first high-profile industry conference in China dedicated to this field. This conference aims to gather experts and innovators to discuss advancements in physics and artificial intelligenc...
OpenAI has made significant advancements in its models for writing software in older programming languages, achieving a remarkable 97.9% accuracy in migrating legacy programs. This improvement is crucial for businesses that rely on outdated software, particularly in sectors like banking where legacy...
Skyfall AI has introduced SKY-001, a world model aimed at enhancing the design of complex physical products through advanced spatial reasoning and planning. This innovative model is reported to be 100 times faster and operates at 150 times lower compute costs compared to existing models, with a comm...
ACE Robotics has announced the release of ACE-Brain-0.6, a 9 billion parameter model specifically tuned for visual grounding tasks. According to company-reported scores, ACE-Brain-0.6 outperformed GPT-6 Astra in six out of eight grounding tests after processing approximately 1 million spatial sample...
On October 14, Google DeepMind officially launched the Gemini 4 Argon, a universal physical reasoning model. This model is designed to handle tasks such as robot control, virtual environment setup, training data quality inspection, and experimental replication. A demonstration with the Apptronik Apo...
In recent years, numerous new models such as π0, GR00T, and SmolVLA have emerged, frequently changing the state-of-the-art in simulation benchmarks. However, deploying these models onto real robots presents significant challenges, including data format conversions and the need to rewrite engineering...
Hesai has introduced Kosmo, a 3D world data platform that integrates the Kosmo One spatial camera and Kosmo World cloud service to transform real-world environments into digital formats for robot training and world modeling. An operator can scan a typical 100-square-meter indoor area in approximatel...
Physical Intelligence has introduced a beta version of its Pi API, enabling robotics developers to access its foundation models for fine-tuning and inference. This platform allows developers to upload robot data and fine-tune models on company-managed compute resources, streamlining the adaptation o...
Neura Robotics has unveiled Neuron AI, a multimodal foundation model designed to enhance robots' understanding of the physical world. This model integrates vision, touch, audio, robot-state data, language, and memory, utilizing a Perception-Language-Action architecture to maintain context and track...
Physical Intelligence has introduced a beta API designed for training and running its robot models, marking a significant advancement in providing intelligence for various machines. Announced on October 13, this service allows teams to upload demonstrations, train policies, and test them on hardware...
From October 12 to 16, 2026, the 26th China International Industry Fair took place at the National Exhibition and Convention Center in Shanghai. Elite Robots showcased its 'Yuanqi Primo' industrial-grade embodied large model at booth 8.1H-F008, emphasizing the transition of embodied intelligence fro...
OpenAI announced that the FBI arrested one of its employees after the company utilized its own models to investigate a hidden recording device found in its office. The models assisted in tracing the device's origins and identifying the employee suspected of planting it. Following the incident, OpenA...
Recent advancements in frontier AI models, particularly Astra and Opus, are significantly expediting the development processes in robotics. These models streamline labor-intensive tasks such as building simulated environments, creating tasks, and coding for testing. Demonstrations of Astra and Claud...
Anthropic has issued a warning that small AI models could soon develop the ability to replicate themselves across internet-connected devices, such as smart ovens and Android TV boxes. This capability, expected to emerge in just a few months, raises significant safety concerns as these models could u...
Anthropic, a U.S. AI company, has accused several Chinese AI developers of engaging in aggressive and malicious activities to distill models from U.S. AI systems, including Claude. A report from Anthropic revealed that companies like DeepSeek, Moonshot AI, and Alibaba extracted billions of tokens fr...
Helm.ai announced it has signed $70 million in commercial contracts for its foundation models focused on physical AI over the past year. These contracts involve global automotive OEMs, Tier 1 suppliers, and industrial automation companies, with a primary focus on the automotive sector. The signific...
Telecom operators are increasingly adopting open models to shape their AI strategies, driven by the need for trust, control, and customization across critical workloads. According to NVIDIA’s State of AI in Telecommunications report, 89% of respondents view open source models as vital to their AI in...
A recent study by marketing firm Graphite has uncovered that advanced AI models still exhibit unique writing characteristics, identifying 13,000 phrases that occur at least twice as frequently in AI-generated text compared to human writing. Researchers analyzed 10,000 articles published prior to the...
YouTube creator Felix Kjellberg, known as PewDiePie, has introduced Ajax, a compact language model designed for local use as an AI assistant. This development follows a contentious period where OpenAI suspended his account twice during Ajax's creation, citing 'distillation' as a reason for one of th...
Satlyt, a startup based in Sunnyvale, California, and Nairobi, has announced an $8 million seed funding round led by Non Sibi Ventures. The funding will support the development of software that allows AI models to operate on satellites, facilitating the sharing of substantial computing workloads in...
AIsphere has launched PixVerse R2, a groundbreaking real-time world model capable of generating explorable audiovisual environments. This innovative platform takes various inputs, including text, images, audio, and keyboard commands, while maintaining state throughout a session. The introduction of...
Runway, a prominent developer of generative AI video technology, is venturing into robotics with the introduction of Praxis-1, an open-weight AI model. This model is designed to leverage knowledge gained from video to control physical robots, allowing developers to download and adapt it independentl...
On September 18, 2026, independent testing organization Robocurve released a benchmark report titled RoboHarm. The report addresses a critical question regarding AI: can large language models effectively recognize and reject dangerous commands when controlling real robotic arms? The study tested thr...
Chinese AI models have seen a remarkable increase in global usage in 2026, particularly on platforms like OpenRouter and Vercel, where their token usage surged significantly. This rise is attributed to lower costs and improved performance, especially for coding tasks, although U.S. models still domi...
China Mobile has announced the open-source release of its Open-RAIL middleware, designed to facilitate the integration of VLA/WAM models with robotic hardware. This middleware introduces asynchronous inference, hardware abstraction, and approximately 50 to 100 lines of code for model hooks, enabling...
The CT-Unite team has won the ACM MM 2026 championship, marking their second consecutive victory in international AI competitions. This time, they achieved success with a brain-inspired, cross-modal cognitive neural network designed for robotics, which compresses a 100 TFLOPS model to operate at 38....
ZDTaichu 5.0-9B, a 9 billion parameter multimodal model, has achieved remarkable results in spatial embodied intelligence, securing first place in 8 out of 9 international benchmarks. It outperforms competitors like Qwen 3.5-9B and STEP 3-VL-10B in spatial perception and three-dimensional reasoning...
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.
