Morgan Stanley maintained its "Overweight" rating and $300 price target for Nvidia on November 11, approximately 77.6% higher than the closing price of $168.89 on November 10. The bank's core rationale for its bullish outlook is the continued strong and sustained demand for AI infrastructure, with training and inference computing power demand consistently exceeding supply. The next-generation Rubin platform is expected to become the mainstream choice for generative AI workloads, and Rubin Ultra is also expected to maintain its performance lead, driving further upward revisions to earnings forecasts.
More importantly, the bank believes that the bottleneck in data center construction is expanding from chip supply to land, power, and building infrastructure (LPS), which could present an opportunity for Nvidia to expand its market share. With tightening power supplies, customers are increasingly focused on the computing power generated per unit of electricity. Nvidia's system-level advantages are expected to improve the output per unit of power in data centers. In August, the company disclosed that its Vera Rubin platform offers 30 times higher throughput per megawatt and 35 times lower token costs (company-reported) compared to Grace Blackwell Ultra.
Furthermore, Nvidia's competitive advantage has expanded from GPUs to a complete AI computing platform comprised of CPUs, high-speed interconnects, networking devices, and software, capable of covering training, inference, and agent workloads. Morgan Stanley believes that strong demand, product iteration, and system-level advantages collectively support its long-term growth, and its current valuation is approximately 20 times its expected 2027 earnings per share, representing a discount compared to some of its computing semiconductor peers.
