Gaming sequential growth driven mainly by strong RTX 40 series sales; Switch SoC down; gaming demand stable in other regions and showing recovery signs in China; >400 games support ray tracing/DLSS; GFN users >25M across 100+ countries.
Data center: North America cloud customers up year over year; China down year over year and sequentially due to bans; despite some cloud customers pausing plans in Q4, hyperscale revenue still grew strongly sequentially; despite weak economy, AI/GPU demand remains very strong; CSP revenue > vertical, more enterprises cloud-first; CSP TTM revenue 40% of data center; H100 second quarter of shipments, revenue already exceeded A100, A100 down sequentially.
Visualization workstations still in inventory correction, expected to end by H1; desktop workstations up sequentially, concentrated in auto, manufacturing and other verticals; Omniverse downloads >300K.
NV AI Enterprise 3.0 ($5K per unit? too expensive?) supports 50+ frameworks and pre-trained models; announced AI cloud service business; launched DGX Cloud, browser-accessed, already on Oracle Cloud and Microsoft Azure; Google GCP and others on the way.
Networking revenue up year over year but below expectations due to slowing traditional CPU server demand; InfiniBand kept growing, strong cloud/enterprise/supercomputing demand.
Q4 repurchased $1.05B, dividends $100M; $7.23B buyback authorization remaining through Dec 2023.
FY24 capex expected $1.1-1.3B, Q1 $350-400M.
Signed 10-year partnership with Microsoft to bring Xbox games to GFN; post-Activision Blizzard close, those titles will join GFN.
Guides Q1 data center and gaming up significantly sequentially; data center can grow year over year; visualization and auto up slightly sequentially; software revenue continues growing, in the hundreds of millions, details at GTC; expects FY24 data center full-year year-over-year growth, accelerating after Q1.
Auto business record high, growth includes development agreements; Mercedes' future software revenue split 50-50.
Gross inventory purchase and long-term supply obligations were $4.92 billion and prepaid supply agreements were $3.45 billion. Other purchase obligations, inclusive of $2.23 billion of multi-year cloud service agreements, were $3.14 billion.
At the AI platform software layer, they will be able to access NVIDIA AI Enterprise for training and deploying large language models or other AI workloads. And at the AI-model-as-a-service layer, NVIDIA will offer its NeMo and BioNeMo customizable AI models to enterprise customers who want to build proprietary generative AI models and services for their businesses.