NVIDIA and Deloitte Bringing Physical AI Coming to Cities | Deloitte at SCEWC25

Why It Matters

Physical AI bridges the gap between digital intelligence and real‑world assets, giving cities and industries a fast, scalable way to cut costs, improve safety, and generate new revenue streams.

Key Takeaways

• Nvidia and Deloitte deploy “physical AI” using camera‑based visual agents.
• Digital twins simulate, train, and deploy AI for city infrastructure.
• Visual agents turn static city assets into autonomous, cost‑saving robots.
• Two‑billion global cameras become the primary sensor for smart‑city services.
• Rapid deployment platform enables weeks‑long rollout of AI‑driven safety services.

Summary

The video showcases a partnership between Nvidia and Deloitte aimed at turning ordinary urban infrastructure into "intelligent cities" through a new paradigm called physical AI. Leveraging Deloitte’s Smart Spaces offering across Italy, Greece and Malta, the teams are deploying camera‑centric visual agents that perceive, reason and act in the physical world, supported by Nvidia’s hardware and software stack. Key insights include the three‑stage loop of simulation, training and runtime deployment, where digital twins generate billions of miles of virtual data before models are refined and installed in real‑world assets such as intersections, autonomous vehicles and manufacturing lines. Generative AI now enables these visual agents to accept natural‑language tasks—producing safety reports, monitoring road conditions, or inspecting factory defects—using the world’s two‑billion cameras as the primary sensor. Notable examples highlighted are the Polarmo pilot, where AI‑driven safety services are already live, and more experimental demos like a basketball‑game‑monitoring agent. Deloitte emphasizes a Center of Excellence that can spin up solutions in weeks, while Nvidia stresses that its platform is sensor‑agnostic, allowing customers to choose camera‑only or multimodal setups based on redundancy needs. The broader implication is a rapid shift of AI from purely IT environments into the physical economy that accounts for roughly half of global GDP. Cities and enterprises can cut operational waste, unlock new revenue streams, and address labor shortages by turning static infrastructure into autonomous, data‑rich robots, accelerating the rollout of smart‑city services worldwide.

Original Description

A new era of transformation is dawning for cities and physical industries, where AI is no longer confined to digital realms but is directly embedded within the physical world. In this session, Deloitte and NVIDIA explore how 'physical AI' enables urban centers, infrastructure, and industrial settings to perceive, reason, and act instantaneously. This innovation turns everyday spaces – from intersections to public facilities – into intelligent, responsive systems. By integrating advanced AI platforms with deep local expertise and advisory services, cities can deploy intelligent solutions more rapidly, reduce operational complexities, and generate new value for their citizens, moving smart city initiatives from experimental pilots to scalable, real-world applications.