Back to News
Market Impact: 0.4

Nvidia Reports Earnings This Month, and I'm Not Buying Shares. 1 AI Stock to Buy Now Instead.

NVDAAVGOAMZNGOOGLMETANFLX
Artificial IntelligenceTechnology & InnovationCorporate EarningsCorporate Guidance & OutlookCompany FundamentalsAnalyst InsightsAntitrust & Competition
Nvidia Reports Earnings This Month, and I'm Not Buying Shares. 1 AI Stock to Buy Now Instead.

Nvidia reported fiscal Q4 revenue of $68.1 billion, up 73% year over year, and guided fiscal Q1 revenue to about $78 billion, implying roughly 75% growth. However, the article argues that Broadcom-led custom chip programs from major customers such as Google, Anthropic, OpenAI, and Meta are creating credible alternatives to Nvidia, potentially limiting pricing power. Amazon is highlighted as a stronger AI-chip play, with its custom silicon business at a $20 billion annual run rate and growing triple digits year over year.

Analysis

The key market implication is not that Nvidia’s demand is weakening today, but that its future unit economics are moving from monopoly-like to oligopoly-like faster than the stock multiple assumes. The first-order beneficiary of that shift is Broadcom, because custom silicon wins are sticky, high-margin, and create a recurring design-socket model that can compound even if the broader AI capex cycle moderates. The second-order effect is a slower-than-expected mix shift in hyperscaler spend away from merchant GPUs toward in-house or semi-custom ASICs, which should gradually pressure Nvidia’s pricing power before it shows up in absolute unit volumes. Amazon looks more attractive than a simple “own the cloud leader” trade because its chip business turns internal capex into a strategic moat: every incremental Trainium deployment both lowers AWS customer acquisition cost and raises switching costs for large model developers. That creates a flywheel Broadcom does not fully capture, since AWS can subsidize silicon adoption through cloud commitments. The market is likely underestimating how much of Amazon’s AI capex is economically pre-committed over a multi-year horizon, which reduces the probability that near-term spending becomes stranded if the AI cycle slows. The contrarian risk on Nvidia is timing: this is not a collapse story, it is a marginal repricing story, and those often hurt most when growth remains strong but decelerates from extraordinary to merely excellent. If the next 1-2 quarters show any evidence that Blackwell demand is being satisfied by customer-specific alternatives, the multiple can compress quickly even without an earnings miss. Conversely, a sharp upside surprise in networking or gross margin would refute the near-term substitution thesis, but the more durable downside to Nvidia comes over 6-18 months as custom silicon deployments scale into production capacity rather than prototypes.