Nvidia delivers big once again, and investors brush aside the bigger picture about AI spending

Nvidia topped expectations with data-center revenue of $51.2 billion versus $49.09 billion expected, a 66% year-over-year increase, driving a >5% after-hours jump in its shares and rallies in peers (AMD +4%+, Broadcom ~+3%, TSMC +3%, Super Micro +5%, Oracle +3%). While the market cheered robust earnings and forward guidance, investors and analysts warned the core question remains whether the massive buildout of GPU capacity will be economically viable long-term—highlighted by calls for accelerated depreciation and Michael Burry’s charge that hyperscalers understate chip depreciation. Added competitive pressure from AMD and bespoke accelerators further muddies the outlook, leaving uncertainty over whether Nvidia can sustain its current pace of growth and market share gains.

Analysis

Nvidia reported data-center revenue of $51.2 billion versus analyst expectations of $49.09 billion, representing a 66% year-over-year increase; the beat and robust guidance drove a greater-than-5% after-hours jump in NVDA and sympathetic gains across peers (AMD +4%+, Broadcom ~+3%, TSMC +3%, Super Micro +5%, Oracle +3%). The stock had fallen about 8% earlier in the month, and several market participants framed the results as a buy-the-dip opportunity given continued insatiable demand for AI compute.

The central unresolved issue remains whether the massive GPU capacity being built today will be economically viable long term: management (Jensen Huang) dismissed bubble talk while analysts and commentators highlighted depreciation and monetization risks. Adam Crisafulli flagged potential calls for accelerated depreciation amid rapid upgrade cycles, and Michael Burry reiterated a view that hyperscalers may be understating chip depreciation, underscoring an accounting and profitability overhang.

Competitive dynamics add near-term uncertainty as AMD and custom accelerators from large tech firms begin to pressure Nvidia’s share, exemplified by the article’s note that Alphabet hit a new high with an AI model not using Nvidia’s Blackwell; these forces temper the otherwise strong near-term revenue trajectory and create a credible path for slower market-share or margin expansion.

Key Decisions for Investors

• Consider adding to NVDA on strength-in-earnings if underweight, but size entries modestly given unresolved depreciation and monetization risk and use defined position limits
• Implement downside protection or partial hedges to guard against an earnings-driven re-rating if hyperscaler depreciation disclosures or rapid GPU upgrade evidence materialize negatively
• Monitor competitor adoption (AMD, custom accelerators, signs like Alphabet’s model choices) and hyperscaler capex/depreciation language closely and reweight semiconductor peers (AMD, AVGO, TSM) based on confirmed market-share shifts