NVIDIA (NVDA) investor relations material

Key technology trends and inflection points

• AI is at its third major inflection point: agentic systems capable of autonomous task execution, moving beyond generative and reasoning AI.

• Token-based computing is now central, with engineers allocated token budgets to drive productivity and value creation.

• OpenClaw is positioned as the operating system for personal AI computers, with every company needing an OpenClaw strategy akin to past Linux or cloud strategies.

• The architecture supports all major AI models, including state space and hybrid models, ensuring versatility and future-proofing.

• The company’s full-stack approach, including hardware, software, and operating systems, enables rapid annual product cycles and seamless compatibility.

Market outlook and business momentum

• Strong visibility and confidence in $1 trillion+ demand for Blackwell and Rubin products through 2027, with expectations for further growth as new customers and regions are added.

• The addressable market is expanding, with additional upside from new products like Groq, CPUs, and storage, potentially increasing opportunity by 25–50%.

• Both hyperscaler (cloud) and enterprise/industrial (on-prem, edge) segments are growing rapidly, with the 40% non-hyperscaler share expected to eventually surpass 60% as physical AI adoption accelerates.

• AI token generation is becoming a core economic activity, with future IT business models shifting from software licensing to token rental and generation.

• All market segments—free, paid, enterprise, and high-end—are experiencing exponential growth, with segmentation and pricing evolving as the market matures.

Product and platform evolution

• Vera Rubin and Groq architectures address both high-throughput and low-latency inference, enabling new tiers of AI services and maximizing data center revenues.

• Groq is expected to account for 25% of inference workloads, increasing compute spend by 25% and not cannibalizing high-bandwidth memory demand.

• The platform’s modular rack architecture harmonizes power, cooling, and compute for scalable, efficient AI factories.

• Token costs will continue to decline with each generation, while token smartness and throughput per watt will rise, driving value and gross margins.

• The company’s ecosystem and CUDA compatibility attract developers and customers, reinforcing its leadership across cloud and on-prem markets.