According to AI News on X, DeepSeek V4.1 Flash achieved 77% on the DeepSWE coding benchmark at under $0.20 per task, signaling cost efficient software automation.

According to TheRundownAI, California halted REK bouts, Boston Dynamics updated Atlas, FieldAI nears $10B, and exoskeletons surge in adoption.

According to SawyerMerritt, Boston Dynamics revealed a 13-DOF, 4-finger Atlas hand with tactile sensors and 100 lb+ payload for robust in-hand manipulation.

According to @CNBC, Chinese humanoid robots still trail human performance, limiting factory ROI and deployment speed in 2026.

According to AINewsOfficial_ on X, Atlas reveals modular head tech, Nucleus fields humanoids in 90 days, and FLUX 3 Video adds 13 features.

According to AINewsOfficial_ on X, Moonshot Kimi K3 challenges Google Gemini Flash as new robot hands from Generative Bionics and Mimic debut.

According to TheRundownAI, Toyota’s humanoid spinoff raises $1.1B as Boston Dynamics trials Spot delivery and Hyundai faces strikes over robot coworkers.

According to SawyerMerritt, Hyundai will buy SoftBank’s remaining 9.65% in Boston Dynamics for $325M, consolidating control for robotics commercialization.

According to @AINewsOfficial_ Boston Dynamics updates Atlas footwear, Bolt sprints 10 m/s, and Mobileye buys Mentybot for $900M.

According to TheRundownAI, Boston Dynamics trained Atlas via reinforcement learning on cloud GPUs to perform a Rabona and target factory work with Hyundai.

According to @AINewsOfficial_ three humanoid advances drop: LimX Luna at ~$41K, Atlas learns 4 soccer skills, and AGIBot debuts BFM2 locomotion.

According to SawyerMerritt, a humanoid robot fell during a live act but continued performing, highlighting emerging entertainment robotics reliability.

According to The Rundown AI, Figure showcases bed-making robots, Korea taps Boston Dynamics, Uber partner faces crashes, and Rocket Lab buys a Mars arm.

According to Google DeepMind on X, the team connected Gemini Robotics ER to Boston Dynamics’ Spot through a systems bridge, allowing operators to command the robot in plain English and enabling capabilities like free navigation, photo capture, and object grasping without writing complex code. As reported by Google DeepMind, the natural language interface acts as a tool-use layer that translates high-level instructions into Spot actions, paving the way for faster deployment of inspection, data collection, and pick-and-place workflows in industrial sites. According to Google DeepMind, this approach reduces integration costs and expands robot accessibility for field operations, creating opportunities in facility inspection, logistics support, and autonomous documentation with multimodal perception.

According to Google DeepMind on X (Twitter), the team integrated Gemini Robotics embodied reasoning models into Boston Dynamics’ quadruped robot Spot, enabling improved scene understanding, object identification, and execution of simple natural language commands such as tidying a room. As reported by Google DeepMind, this fusion of multimodal perception and planning boosts Spot’s on-robot reasoning to handle open-ended tasks and real‑world variability, signaling near-term applications in facilities inspection, logistics support, and on-site assistance where autonomy and safety are critical. According to Google DeepMind, the collaboration demonstrates practical embodied AI gains—translating language instructions into action plans, grounding object references, and verifying outcomes—which can shorten deployment cycles for enterprise robotics and reduce the need for bespoke rule-based pipelines.

According to GoogleDeepMind on X, Gemini Robotics-ER 1.6 improves visual and spatial reasoning so robots can plan and complete more useful tasks, including autonomously reading complex industrial gauges on Boston Dynamics’ Spot (source: GoogleDeepMind thread by @GoogleDeepMind). As reported by Demis Hassabis on X, the upgrade is designed to help robots reason about the physical world and operate more usefully in real environments, highlighting a step toward robust perception-to-action pipelines in robotics (source: @demishassabis). According to GoogleDeepMind, these capabilities target practical deployments in industrial inspections, where accurate analog gauge reading and context-aware navigation can reduce downtime and labor costs while improving safety at facilities (source: @GoogleDeepMind).

According to Google DeepMind on X (Twitter), Gemini Robotics-ER 1.6 can process complex analog dial images captured by patrol robots like Spot from Boston Dynamics, generating its own code to correct camera distortion and compute exact tick marks for precise readings. As reported by Google DeepMind, this upgrade targets industrial inspection workflows where consistent, accurate gauge interpretation is critical for safety and uptime. According to the posted demo video, the model’s code-writing approach enables on-device adaptations to varying lenses and angles, which can reduce manual calibration time and expand autonomous inspection coverage.

According to Google DeepMind (@GoogleDeepMind), the company is launching a strategic research partnership with Boston Dynamics to combine DeepMind's advanced Gemini model variants—designed for visual understanding and robotic action—with Boston Dynamics' state-of-the-art robotics hardware, including the new Atlas® humanoids. This collaboration aims to accelerate breakthroughs in robotic learning and practical AI deployment in real-world environments. By leveraging Gemini's foundational AI capabilities and Boston Dynamics' robust platforms, the partnership is expected to drive innovation in industrial automation, logistics, and advanced service robotics, offering significant business opportunities for enterprises seeking scalable AI-powered automation solutions (Source: @GoogleDeepMind, https://x.com/GoogleDeepMind/status/2008283100254494916).

According to @GoogleDeepMind, a new partnership with Boston Dynamics will integrate Gemini Robotics’ state-of-the-art AI models with Boston Dynamics' advanced Atlas humanoid hardware. This collaboration aims to accelerate the deployment of artificial general intelligence (AGI)-driven robotics in real-world environments, a key step for the future of physical AI systems. The partnership is expected to enhance the capabilities of robotics in logistics, manufacturing, and service industries by combining DeepMind’s foundational AI research with Boston Dynamics’ proven robotics platforms (Source: @GoogleDeepMind, Jan 5, 2026).

According to @GoogleDeepMind, Google DeepMind and Boston Dynamics have entered a new research partnership to integrate DeepMind's Gemini Robotics foundational AI capabilities into Boston Dynamics' next-generation Atlas humanoid robots. This collaboration aims to enhance the real-world autonomy, perception, and manipulation skills of Atlas robots, leveraging Gemini’s multi-modal reasoning and planning abilities. The partnership is expected to accelerate the deployment of advanced AI-powered robotics solutions for industrial automation, logistics, and manufacturing sectors, presenting significant business opportunities for enterprises seeking scalable and adaptive robotic systems (source: @GoogleDeepMind, Jan 5, 2026, goo.gle/49paguA).