Vera Rubin servers
Search documents
Nvidia stock fails to rally after Huang's speech but analysts remain bullish
Invezz· 2026-03-17 16:47
Core Viewpoint - Nvidia's stock experienced a slight decline following CEO Jensen Huang's keynote at the developer conference, indicating investor skepticism about the potential for a stock rally despite positive announcements regarding AI revenue forecasts [1][6]. Stock Performance - Nvidia shares fell 0.4% to $182.42 after a modest gain of 1.7% in the previous session, continuing a trend of muted reactions to recent conferences [1][2]. - The stock has been trading within a narrow range of $180–$190 for several months, reflecting investor caution amid macroeconomic pressures [3]. Macroeconomic Context - Concerns about the sustainability of AI infrastructure spending and broader macroeconomic issues, such as geopolitical tensions and recession fears, have contributed to the stock's stagnation [3][4]. Analyst Sentiment - Despite the stock's near-term challenges, Wall Street analysts maintain a bullish outlook on Nvidia's long-term prospects, with Rosenblatt Securities highlighting a potential $1 trillion in cumulative AI-related revenue from 2025 to 2027 [5][6]. - UBS reiterated a Buy rating with a price target of $245, identifying inference, infrastructure, and robotics as key growth drivers [7]. Revenue Outlook - Nvidia announced an upgraded revenue outlook, expecting at least $1 trillion in cumulative revenue from its Blackwell and Rubin chip platforms between 2025 and 2027, a significant increase from a prior estimate of $500 billion through 2026 [8][9]. - Analysts noted that the updated forecast largely confirmed existing expectations rather than providing a substantial upside surprise [10]. AI Inference Focus - A major theme from the conference was Nvidia's emphasis on AI inference, with the introduction of a new system designed for inference workloads, integrating Vera Rubin servers with specialized chips [12]. - Nvidia plans to ship Vera Rubin systems in the second half of 2026 to compete with alternative chip architectures optimized for inference [13].