Nvidia posts record Q4 with $68 billion revenue, up 73% year-over-year, as data center revenue surges 75% to $62 billion. CEO Jensen Huang emphasized that in the AI era, compute directly equals revenues, stating that without computational throughput, companies cannot generate tokens or grow revenues. Real-time inference is now the primary driver of cloud monetization, making performance per watt critical as operators scale production workloads. The company has reached an inflection point with agentic AI workloads generating unprecedented compute demand. While hyperscalers represent just over half of data center revenue, enterprises, AI model developers, and sovereign customers are growing even faster. Nvidia's CUDA ecosystem remains a competitive advantage, compatible across GPU generations and present in every major cloud platform.
