Salesforce and NVIDIA are significantly expanding their strategic collaboration, pushing enterprise AI beyond digital interfaces into the tangible physical world. This pivotal partnership introduces what they term "Physical AI," transforming autonomous robots into intelligent, actionable agents capable of perceiving, reasoning, and taking action within real-world environments. The initiative aims to bridge the long-standing gap between enterprise workflows and physical operations, heralding a new era of automated physical tasks and fundamentally reshaping how businesses interact with their physical spaces.
The concept of Physical AI represents a critical evolution in artificial intelligence, marking a definitive move beyond predictive analytics and generative models into the "Third Wave" of autonomous robotic agents. For too long, physical spaces and digital systems have operated as disconnected silos, relying heavily on human workers to manually observe problems, interpret situations, and then log tickets into digital systems. This fragmented workflow has consistently led to inefficiencies, delayed resolution times, and a predominantly reactive approach to critical issues. Salesforce and NVIDIA are now directly addressing this fundamental disconnect, proposing a unified system where robotic perception, powered by advanced AI, seamlessly integrates with enterprise-level action, enabling proactive problem identification and resolution. This paradigm shift promises to unlock unprecedented operational agility and responsiveness.
The synergy between Salesforce's Agentforce platform and NVIDIA's cutting-edge visual AI infrastructure is not merely complementary; it is foundational to the realization of Physical AI. NVIDIA's Metropolis Blueprint for video search and summarization (VSS), combined with its Cosmos Reason vision language model, provides the sophisticated "eyes" and initial "System 1" perception for these robotic agents. This technology translates vast streams of unstructured video data into structured, actionable intelligence in real-time, filtering out noise to identify true anomalies. Salesforce's Agentforce then layers on the crucial "business brain," delivering "System 2 and 3" intelligence for high-level work coordination, applying enterprise context, and orchestrating autonomous actions across multiple systems and human personnel. This clear division of labor—NVIDIA handling complex real-time perception and simulation, while Salesforce provides the business logic and workflow automation—is precisely what enables the seamless integration of physical robots into existing enterprise operations.