The sessions, scheduled for September 9, 2026, at the Marriott Marquis, will anchor the four-day event by addressing the transition from AI model training to widespread deployment. Dr. Basu’s opening keynote, "Designing AI Factories for Inference," will center on token economics—the cost and efficiency metrics he identifies as the primary drivers for next-generation digital infrastructure. He argues that as organizations scale, the industry must fundamentally rethink data center design and networking to support high-performance inference.

Following the keynote, Basu will join Pete Coticchia, NVIDIA’s Global Head of Business Development for Oracle Cloud, for a discussion moderated by IG Group CEO Vinay Nagpal. The conversation will trace the evolution of NVIDIA’s computing platforms, spanning from the Hopper and Blackwell architectures to the upcoming Rubin and Feynman generations. The participants intend to examine the practical challenges of scaling AI, specifically regarding energy efficiency, cooling, and the need for low-latency interconnects across cloud regions. By bringing together these perspectives, the conference aims to provide enterprise decision-makers with a roadmap for the hardware and connectivity investments required to sustain the ongoing AI expansion.