Why the Nasdaq's Artificial Intelligence (AI) Rally Could Be Just Getting Started: 2 Best Growth Stocks to Buy

The article argues that Alphabet and Amazon are the best positioned long-term AI winners, citing Alphabet’s seventh-generation TPUs, Gemini model, Google Cloud deployment, and Wiz acquisition, alongside Amazon’s custom chips and cloud scale. Amazon’s chips are described as a $20 billion run-rate business and potentially a $50 billion business including internal use, while Alphabet is framed as having a major cost advantage in AI training and inference. Overall tone is bullish on both stocks, but the piece is primarily opinion/commentary rather than new hard financial data.

Analysis

The market is starting to value AI less as a model race and more as a cost-arbitrage war. That shifts incremental economics toward the hyperscalers that can internalize inference, improve utilization, and blunt external chip pricing power; in that regime, the biggest losers are not obvious software names but the ecosystem vendors whose margins depend on customers not building in-house silicon or optimizing workloads.

The more important second-order effect is that custom silicon narrows the gap between capex and monetization. If TPU/Trainium-style deployments keep scaling, the next 12-18 months should see a widening spread between headline AI spend and reported operating leverage, because internal chips reduce both depreciation per token and cloud procurement leakage. That favors GOOGL and AMZN over stand-alone AI infrastructure trades that need continued multiple expansion to justify current prices.

The contrarian miss is that “AI winner” may be too simple a framing: these names are increasingly defensive quality-growth compounds, not pure AI beta. The market has likely underpriced how much search, ads, cloud, and e-commerce efficiency can reaccelerate once agentic workflows start driving higher query intensity and better conversion, while it may be overpricing the near-term payoff to the sell-side favorite hardware beneficiaries. The key risk is timing: the fundamental upside is a 12-36 month story, but sentiment can wobble over the next few weeks if AI capex headlines or cloud growth prints disappoint.

Another underappreciated risk is cannibalization and execution drag. As these firms push AI deeper into products, they may mask near-term margin pressure from infrastructure buildout, and investors could mistake that for deterioration rather than deliberate reinvestment. The trade works best if one is willing to own interim volatility and focus on the eventual margin capture from vertical integration.

Key Decisions for Investors

• Long GOOGL vs. a basket of AI infrastructure hardware names over 6-12 months: thesis is that internal silicon plus software distribution monetizes AI more efficiently than external chip demand; target a relative outperformance of 15-20% with downside limited to sentiment-driven multiple compression.
• Long AMZN on a 3-6 month horizon into cloud reacceleration: use pullbacks to add, with the setup driven by operating leverage from automation and inference cost control; risk/reward is attractive if AWS growth inflects while retail margins keep expanding.
• Pair trade: long GOOGL / short NVDA sized modestly, not as a bearish NVDA call but as a hedge against over-earnings multiple reliance in the hardware layer; expect the spread to work if capex intensity normalizes and buyers optimize in-house.
• Buy 6-12 month upside calls in GOOGL or AMZN instead of common if entry timing is uncertain: the names have a favorable medium-term fundamental arc, but short-term reactions to capex commentary can be noisy.
• Avoid chasing the most promotional AI infrastructure beneficiaries here; wait for a 15%+ drawdown or evidence of sustained backlog conversion before adding, because the marginal upside is now more dependent on valuation than on incremental AI adoption.