October 14, 2025 News
Coco Robotics Establishes Physical AI Research Lab with UCLA Professor to Leverage Five Years of Delivery Robot Data
Coco Robotics, a last-mile delivery robot startup, has appointed UCLA professor Bolei Zhou as chief AI scientist to lead a new physical AI research lab. The lab will leverage millions of miles of data collected by Coco's delivery robots over five years to develop autonomous navigation systems and reduce delivery costs. This initiative is separate from Coco's existing collaboration with OpenAI and focuses on improving the company's own automation capabilities.
Skynet Chance (+0.01%): The development of autonomous physical AI systems with real-world learning capabilities represents incremental progress in AI operating independently in physical environments, though the application is limited to commercial delivery robots with constrained objectives and operational domains.
Skynet Date (+0 days): The accumulation of large-scale real-world robotics data and establishment of dedicated physical AI research modestly accelerates the development of embodied AI systems that can learn and operate autonomously in complex environments.
AGI Progress (+0.01%): This represents meaningful progress in physical AI and embodied intelligence by combining large-scale real-world data collection with advanced research in computer vision, robot navigation, and reinforcement learning, which are key components for developing general-purpose intelligent systems.
AGI Date (+0 days): The establishment of a dedicated physical AI lab with substantial real-world data and top research talent modestly accelerates progress toward embodied AGI by addressing the critical challenge of learning from physical world interactions at scale.
OpenAI Partners with Broadcom for Custom AI Accelerator Hardware in Multi-Billion Dollar Deal
OpenAI announced a partnership with Broadcom to develop 10 gigawatts of custom AI accelerator hardware to be deployed between 2026 and 2029, potentially costing $350-500 billion. This follows recent major infrastructure deals with AMD, Nvidia, and Oracle, signaling OpenAI's massive scaling efforts. The custom chips will be designed to optimize OpenAI's frontier AI models directly at the hardware level.
Skynet Chance (+0.04%): Massive compute scaling and custom hardware optimized for frontier AI models could accelerate development of more capable and potentially harder-to-control systems. However, infrastructure improvements alone don't directly address alignment or control mechanisms.
Skynet Date (-1 days): The unprecedented scale of compute investment ($350-500B) and deployment timeline (2026-2029) significantly accelerates the pace at which OpenAI can develop and scale powerful AI systems. Custom hardware optimized for their models removes bottlenecks that would otherwise slow capability advancement.
AGI Progress (+0.04%): Custom hardware designed specifically for frontier models represents a major step toward AGI by removing compute constraints and enabling direct hardware-software co-optimization. The scale of investment (10GW+ across multiple deals) demonstrates serious commitment to reaching AGI-level capabilities.
AGI Date (-1 days): The massive compute infrastructure scaling, with custom chips arriving in 2026 and continuing through 2029, substantially accelerates the timeline to AGI by removing key bottlenecks. Combined with recent AMD, Nvidia, and Oracle deals, OpenAI is securing the computational resources needed to train significantly larger models faster than previously expected.
Google Announces $15 Billion AI Infrastructure Investment in India with 1-Gigawatt Data Center Hub
Google is investing $15 billion over five years to build a 1-gigawatt data center and AI hub in Visakhapatnam, India, marking its largest investment outside the U.S. The facility will offer Google's AI infrastructure including TPUs and Gemini models, and will be connected via subsea cable infrastructure in partnership with Indian telecom and infrastructure companies. This investment comes amid Indian government pushes for reduced reliance on U.S. tech giants and promotion of local alternatives.
Skynet Chance (+0.01%): The expansion of large-scale AI infrastructure increases global AI computational capacity and deployment reach, marginally raising the surface area for potential AI control challenges. However, this is primarily commercial infrastructure expansion rather than fundamental capability advancement.
Skynet Date (+0 days): Increased AI infrastructure deployment and geographic distribution slightly accelerates the pace at which advanced AI systems can be scaled and deployed globally. The magnitude is small as this represents capacity expansion rather than breakthrough capability development.
AGI Progress (+0.01%): The investment significantly expands computational infrastructure and AI model access in a major global market, facilitating broader AI development and deployment at scale. The introduction of TPU infrastructure and full-stack AI solutions in India represents meaningful progress in global AI capability distribution.
AGI Date (+0 days): The substantial infrastructure investment and commitment to deploying advanced AI systems (Gemini models, TPUs) in a new major hub modestly accelerates the timeline by enabling more distributed AI research and development. The five-year timeline and scaling to multiple gigawatts suggests sustained acceleration of AI computational capacity.