At Dell Technologies World, Michael Dell and Jensen Huang announced a massive shift toward agentic AI, describing current demand as parabolic. With global AI infrastructure spending projected to reach between three and four trillion dollars by 2030, the companies unveiled the Dell AI Factory, a full-stack solution designed to move enterprises beyond AI pilots into large-scale, secure production deployments. The new era emphasizes on-premises and edge computing, with 67% of AI workloads now running outside the public cloud to ensure data governance and security. Central to this announcement is NVIDIA's upcoming Vera and Vera Rubin platforms. The Dell PowerEdge XE9812 server, built on the NVIDIA Vera Rubin NVL72, promises to deliver agentic AI inference at one-tenth the cost per token compared to previous Blackwell generations. Meanwhile, the Dell PowerEdge XE9880 series, featuring the NVIDIA HGX Rubin NVL8 architecture, supports up to 144 GPUs per rack with 100% direct liquid cooling, offering 5.5 times the performance of the HGX B200. On the processing side, the NVIDIA Vera CPU delivers 50% faster agentic workload execution and up to three times faster data queries than traditional x86 processors, thanks to its 1.2 TB/s memory bandwidth and record-breaking single-threaded performance. To support these hardware advancements, Dell introduced the PowerRack, an integrated system combining compute, networking, and storage with optimized thermal and power management. The networking portfolio now features the Dell PowerSwitch with NVIDIA Quantum-X800 InfiniBand and Spectrum-6 Ethernet, utilizing liquid-cooled, co-packaged optics. Software integration includes the Dell AI Data Platform, which features Starburst, a new data engine capable of processing SQL analytics three times faster on Vera CPUs. Several major enterprises demonstrated the practical application of this infrastructure. Eli Lilly, Samsung, and Honeywell shared how they are leveraging the Dell AI Factory to drive innovation in life sciences, chip design, and industrial automation. Additionally, algorithmic trading firm Hudson River Trading is expanding its deployment to support AI-driven research at scale. A key focus of the event was security, highlighted by NVIDIA Confidential Computing in partnership with Fortanix and others. This technology allows organizations to run frontier models, such as those from SpaceXAI and Google Distributed Cloud with Gemini 3.0, without exposing intellectual property or sensitive data. The event also emphasized the expansion of open models and agent orchestration. Enterprises can now deploy open-source models like NVIDIA Nemotron, DeepSeek, and Mistral on the platform. For local operations, Dell Deskside Agentic AI introduces NVIDIA NemoClaw and OpenShell, enabling secure, customized agents to run on workstations and connect seamlessly with enterprise data. These tools provide a runtime for developing autonomous agents with built-in security and policy enforcement. Looking ahead, the collaboration between Dell and NVIDIA aims to streamline the path from development to production for complex, long-running AI agents. Further details on these technologies, including live demonstrations of deskside agentic AI, are expected to be revealed at GTC Taipei during COMPUTEX in June, reinforcing the companies' commitment to a secure, efficient, and scalable future for enterprise AI.