Nvidia's data center business reaches 76% of total revenue in Q2 FY2024, driven by explosive cloud AI server demand.
NVIDIA's data center business reached US$10.32 billion in FY2Q24—a QoQ growth of 141% and YoY increase of 171%—and now accounts for over 76% of the company's revenue. This dominance represents a dramatic shift from FY4Q22, when data center revenue accounted for approximately 42.7%, trailing its gaming segment by about 2 percentage points. By FY1Q23, data center business surpassed gaming at over 45% of revenue, and starting in 2023, with major CSPs heavily investing in ChatBOTS and various AI services for public cloud infrastructures, NVIDIA reaped significant benefits. TrendForce attributes the primary driver of this robust revenue growth to NVIDIA's data center AI server-related solutions, including AI-accelerated GPUs and the AI server HGX reference architecture, which serve as foundational AI infrastructure for large data centers.
NVIDIA is pursuing a multi-tiered GPU strategy to capture both cloud and edge data center markets. For high-end applications, NVIDIA is prioritizing the H100 series to boost data-center-related revenue growth, though production remains constrained by limited CoWoS and HBM technology capacity. The L40s serves as the company's flagship mid-tier product, utilizing GDDR memory rather than CoWoS packaging, allowing rapid introduction to the mid-tier AI server market and filling the gap left by the A100 PCIe interface for enterprise customers. The L40s targets enterprises that don't require large parameter models like ChatGPT, focusing instead on compact AI training applications with parameter counts ranging from tens of billions to under a hundred billion, as well as edge AI inference and image analysis tasks. For lower-tier GPUs, NVIDIA highlights the L4 or T4 series, designed for real-time AI inference or image analysis in edge AI servers, emphasizing affordability while maintaining a high cost-performance ratio.
NVIDIA is actively promoting its HGX and MGX AI server reference architectures as main weapons for AI solutions, and ODMs including Inventec, Quanta, FII, Wistron, and Wiwynn, along with brands such as Dell, Supermicro, and Gigabyte, are encouraged to follow these reference designs. However, these companies must undergo NVIDIA's hardware and software certification process for these AI server reference architectures. Leveraging this approach, NVIDIA can bundle and offer integrated solutions like its Arm CPU Grace, NPU, and AI Cloud Foundation. TrendForce notes that NVIDIA is also partnering with entities like VMware on solutions including the Private AI Foundation, extending NVIDIA's reach into the edge enterprise AI server market and underpinning steady growth in its data center business.
For ODMs and OEMs, NVIDIA's expected significant achievements in the AI server market for CSPs from 2023 to 2024 will likely boost overall shipment volume and revenue growth. However, the introduction of standardized AI server architectures like HGX and MGX will make core product architecture more homogenized, intensifying competition among them as they vie for CSP orders. Large CSPs such as Google and AWS are leaning toward adopting in-house ASIC AI accelerator chips in the future, presenting a potential threat to a portion of NVIDIA's GPU market. This appears to be one reason NVIDIA continues rolling out GPUs with varied positioning and comprehensive solutions, aiming to further expand its AI business aggressively to Tier-2 data centers like CoreWeave and edge enterprise clients.