Nvidia Stock: Morgan Stanley Reports Strong Demand for Hopper/Blackwell

NVIDIA (NVDA) is seeing robust demand for its Hopper and Blackwell GPUs, according to analysts from Morgan Stanley.

In a recent report, the investment bank noted that Nvidia’s Blackwell chips are now entering volume production, driven by significant demand from major customers, indicating strong growth potential for the company.

Morgan Stanley highlighted that Oracle has announced plans to construct a Zettascale AI supercluster that will be powered by 131,000 Nvidia Blackwell GPUs, capable of delivering 2.4 ZettaFLOPS of AI performance.

Oracle’s increased request for GPU supplies has positively impacted Nvidia’s outlook, as well as benefiting its semiconductor suppliers. “At a recent event, Oracle reportedly sought more GPU supplies, which is encouraging news for the AI semiconductor and systems supply chain in Asia,” the bank stated.

Morgan Stanley has also adjusted its expectations for TSMC’s CoWoS (Chip-on-Wafer-on-Substrate) capacity, forecasting an increase to between 80,000 and 90,000 wafers per month by 2025, up from a previous estimate of 70,000.

In the short term, the demand for Nvidia’s Hopper GPUs remains strong, easing concerns over potential inventory risks. The Hopper H200 chips are reportedly gaining traction among smaller cloud service providers and state-sponsored AI initiatives.

As for the Blackwell chips, 450,000 units are expected to be produced in the fourth quarter of 2024, representing a potential revenue opportunity of over $10 billion for Nvidia.

Additionally, Morgan Stanley mentioned that Nvidia is addressing some technical challenges with its GB200 server racks, viewing these challenges as a normal part of the product launch process. Hon Hai, Nvidia’s assembly partner, is on track to start shipping the GB200 server racks by late Q4 2024.

Morgan Stanley maintains a positive outlook on Nvidia’s AI supply chain, favoring it over the memory, PC, and cloud segments as the company continues to benefit from the rising demand for AI semiconductors.