NVIDIA’s CUDA moat and rack-scale AI platforms support exceptional growth potential, but Jensen Huang’s departure, concentrated customers, and margin and financing risks make execution the critical test.
Overview
NVIDIA is the foundational platform provider for accelerated computing, combining GPUs with CUDA software, networking, CPUs, and rack-scale systems. In Q2 FY2027, revenue rose 106% year over year to $96.22 billion; Data Center contributed $89.02 billion, or 92.5% of sales, while Graphics generated $7.20 billion. The company beat revenue consensus by 4.5%, and its $2.22 non-GAAP EPS exceeded estimates by about 6%. Management guided Q3 revenue to $108.0 billion ±2% and initially indicated approximately 70% FY2028 revenue growth, describing the outlook as supply-constrained. The competitive case rests on CUDA’s switching costs and the increasing value captured through integrated racks: up to $40 billion per gigawatt for Vera Rubin versus $18 billion for Hopper. **The near-term setup is complicated by Jensen Huang’s departure, a roughly 30% share-price correction, and memory-driven margin pressure.** Shares at about $162.85 trade at approximately 20.4x trailing and 18.0x forward earnings, below the 10-year historical P/E average of 54.85x. The report’s probability-weighted five-year target is $750.66, with catalysts including Q3 results, the Vera Rubin ramp, enterprise and sovereign AI deployments, and margin recovery.