NVIDIA (NVDA) Bank of America 2026 Global Technology Conference summary

Key announcements and strategic direction

• Vera Rubin platform is in full production and set to launch in Q3, with significant performance improvements and a unique co-designed CPU for agentic solutions.

• RTX Spark, developed with MediaTek, targets AI PCs and agentic workloads, expanding into new customer segments.

• Product portfolio diversification addresses hyperscalers, AI clouds, enterprises, industries, and sovereigns globally.

• New CPU offers 2x performance over x86 CPUs, available both in systems and standalone, supporting agentic computing.

• Focus on full-stack solutions, integrating hardware and software for end-to-end AI and data center needs.

Market segmentation and growth trends

• Revenue split is roughly 50/50 between hyperscalers and AI-focused clouds, with hyperscalers growing faster.

• AI clouds are rapidly expanding in the US, Europe, and Southeast Asia, serving enterprises and regional needs.

• Demand for compute remains high, with ongoing supply tightness and customers seeking both new and upgraded infrastructure.

• ROI for cloud providers and enterprises has improved, driven by agentic workloads and consumer adoption.

• Hardware is not commoditized; customers seek best-of-breed, full-system solutions rather than just chips.

Technology and competitive positioning

• Extreme co-design of systems, including compute and networking, results in over 90% networking attach rates.

• Full-stack software, including CUDA and domain-specific libraries, is a key differentiator and not easily replicated by ASICs.

• Systems are optimized for both training and inference, with Grace Blackwell leading in inference workloads.

• Continuous software innovation extends hardware useful life and enhances value for customers.

• New data centers prioritize sustainability, efficiency, and liquid cooling, with Vera Rubin as the next major upgrade.