Nvidia Up 20% Since May 5: $1 Trillion Market Cap Gains Into Earnings Next Week

Nvidia's explosive rally since May 5 has compressed months' worth of optimism into a two-week window, turning next Wednesday's earnings call into a high-stakes event that will either validate the capex super-cycle or trigger a brutal valuation reset. The stock added roughly $1 trillion of market capitalization in just ten trading days, a velocity that even seasoned AI bulls find breathtaking. The proximate catalyst was US approval to export the H200 chip to ten Chinese companies, a geopolitical signal that semiconductor supply-chain fragmentation fears may ease and that chipmakers can still serve China's AI infrastructure build-out within export rules.

Jensen Huang, Nvidia CEO, has been messaging aggressively about humanity's power needs: his claim that energy demand will increase 1,000-fold due to AI compute rounds out a narrative in which Nvidia's chips are not optional but mandatory for the next decade of infrastructure spending. The bar for earnings, however, has moved dramatically. Goldman Sachs, Morgan Stanley, and other research shops have all raised targets, but the stock is pricing in perfection at nearly every level of guidanceCompany-issued forecasts of future financial performance. and growth. If Nvidia guides conservatively, or if management acknowledges any slowdown in hyperscaler capex spending, the unwind could be swift and severe.

The rally is reshaping sector leadership: semiconductor peers like AMD and Broadcom have lagged, and while Nvidia's dominance in AI training is uncontested, the narrative has begun to shift toward supply-chain bottlenecks in DRAM and memory that could constrain AI infrastructure growth. The Roundhill Memory ETFExchange-Traded Fund - a basket of securities trading like a single stock. (DRAM) has surged to $10 billion in assets at record velocity, signaling traders' view that memory chips are the 'biggest bottleneck in the AI buildup.' This parallel narrative suggests the market is already rotating away from pure Nvidia euphoria toward a more granular capex thesis.

Historical parallels are unavoidable: critics draw comparisons to the dot-com bubble and the 1700s French stock mania by one measure, Nvidia's stretch rivals or exceeds both. Yet believers counter that the compute demands of generative AI are genuinely novel and that enterprise capex cycles are only beginning. The debate hinges on whether AI training capex peaks this year or continues accelerating through 2027. Earnings next week will be the first chance to test that assumption directly.