This AI Stock Has Been Called Overvalued for 3 Years. The Bears Keep Losing.

Nvidia reported fiscal Q1 2027 revenue of $81.6 billion, up 85% year over year, with data center revenue rising 92% to $75.2 billion and gross margin at 74.9%. The company guided to about $91 billion in fiscal Q2 2027 revenue and reiterated long-term AI demand, while analysts remain bullish with 52 of 54 ratings at buy and an average price target of $299, about 35% above the current price. The article argues Nvidia is more likely undervalued than overvalued given its growth and a 25x forward P/E, versus 61x for AMD and 37x for Broadcom.

Analysis

The key second-order takeaway is not just that NVDA is still compounding, but that the AI supply chain is becoming more concentrated around a single pricing node. If Nvidia can keep expanding revenue at this scale while preserving outsized margins, it effectively retains bargaining power over hyperscalers, OEMs, and software buyers, which tends to delay any meaningful commoditization of accelerator spend. That usually means the near-term beneficiaries are still the platform leaders, while weaker competitors are forced into margin-sacrificing bundle strategies.

The market is likely underappreciating the duration of the capex cycle rather than the absolute level of demand. When management talks about a multi-quarter, multi-product ramp, the more important signal is that customers are still building for shortage, not for equilibrium; that supports not only NVDA but also advanced packaging, HBM memory, and high-speed interconnect vendors. The flip side is that this kind of demand concentration is fragile if a small number of customers pull forward purchases, so the next 1-2 quarters matter more than the three-year narrative.

Contrarian risk: the consensus may be too focused on multiple comparison and too little on terminal growth risk. If Blackwell/Vera Rubin adoption is front-loaded, then the market could already be discounting peak scarcity economics, leaving the stock vulnerable to any deceleration in growth rate even if absolute revenue remains enormous. For AMD, the article’s framing is a reminder that the relative bull case is not share gains alone, but whether it can force a change in pricing discipline; absent that, AMD remains a lower-quality second derivative on the same AI spend.

The cleanest read-through is that this is bullish for the AI infrastructure basket, but not necessarily for broad semis. Investors should expect NVDA to continue pulling capital toward the winners with the tightest supply chains and highest attach rates, while late-cycle hardware laggards struggle to translate AI narrative into durable earnings power. In other words, the trade is quality-plus-scarcity, not beta to semiconductors as a whole.

Key Decisions for Investors

• Maintain long NVDA versus a basket of lower-quality semis; prefer a 3-6 month relative-value long NVDA / short SOXX-style basket if entry can be timed on post-earnings consolidation, because the stock still has better pricing power and backlog visibility
• Add exposure to AI supply-chain beneficiaries on weakness — especially HBM/packaging/interconnect names — with a 3-6 month horizon, since the next leg of spend is more likely to accrue there than to generic semiconductor exposure
• Short dated downside hedges on AMD via put spreads into any AI hype-driven rallies; risk/reward is favorable if the market starts to question share gains or pricing power over the next 1-2 quarters
• If already long AVGO, keep it as a slower-moving way to own AI infrastructure, but size below NVDA because the upside is more capped unless networking and custom silicon spend reaccelerates over the next 2-3 quarters
• For event-driven accounts: buy NVDA pullbacks after guidance-driven volatility rather than chasing strength; the setup remains structurally bullish, but the main risk is multiple compression if growth decelerates even modestly