NVDA is up just +3.40% YTD and +9.80% over the past 52 weeks — a startling underperformance for the company that arguably owns the AI infrastructure stack. Meanwhile GOOGL has quietly compounded +74.37% over the trailing year and trades at just 16.9x TTM earnings with a $4.08T market cap, making it the most attractively valued hyperscaler in the group. The AI spend cycle is accelerating — Amazon just announced a $220B AI investment plan and said cloud demand is outpacing capacity — yet the market is pricing NVDA and GOOGL with surprising restraint. That gap deserves serious attention.
Let's start with NVDA, because the setup here is genuinely confusing if you're paying attention. The stock is at $195.04 — up +2.65% today on 127.3M shares — yet YTD performance is just +3.40%. For a company with $253.5B in TTM revenue, a 62.97% profit margin, and a P/E of just 30.2x, that's not the frothy speculative premium the bears keep warning about. If anything, NVDA has de-rated meaningfully as earnings have grown into the valuation. The beta of 2.21 tells you the market still treats this as a high-octane trade, but the fundamentals are increasingly telling a steady-compounder story. Amazon's Q2 print — $201B in quarterly revenue, 20% YoY growth, and a CFO who literally said capex won't be enough to meet demand — is exactly the demand signal that should be pulling NVDA higher. Something structural is holding the stock back, and I suspect it's a combination of export control overhang and the market's uncertainty about when the next product cycle truly inflects.
On GOOGL, the valuation story is almost embarrassingly clear. At 16.9x TTM P/E and 12.2x EV/EBITDA, Alphabet is trading at a meaningful discount to every other Mag 7 name except META — and META is under legitimate structural pressure with a -30.08% 52-week return. GOOGL has done the opposite: +74.37% over the past year. The cloud business is booming, Google DeepMind continues to ship, and the AI search narrative — once framed as an existential threat — is increasingly looking like a moat. The Reddit data-licensing situation is worth watching: Reddit may walk away from a ~$60M/year deal because Google's AI Overviews are consuming Reddit content without driving return traffic. That's not a bug for Google — that's the product working as intended. If anything, it's evidence that Google's AI search is successfully replacing the open web for information retrieval, which is a structural revenue opportunity, not a threat.
The connection between NVDA and GOOGL is important here and worth making explicit. Google is one of the largest buyers of NVDA silicon on earth — through its TPU program it partially self-supplies, but its AI training and inference workloads still flow through NVIDIA infrastructure at scale. When Google's cloud growth accelerates and AI product development intensifies, that's upstream demand for NVDA. Amazon's commentary about demand outpacing capacity is effectively a forward order book signal for the accelerated computing stack. This is why both stocks should be rising together — and the fact that NVDA is lagging while cloud hyperscalers are printing strong numbers creates a tension that likely resolves upward for NVDA into the next earnings cycle.
Valuation cross-check across the group: AAPL at 40.96x P/E with a $4.90T market cap on $451.4B in TTM revenue is the most expensive Mag 7 name on an earnings basis. MSFT at 23.4x P/E at $451.10 — the cloud print appears to have been decisive for sentiment. NVDA at 30.2x and GOOGL at 16.9x are the two names where valuation is most clearly in the investor's favor relative to AI growth exposure. TSLA at 286.8x P/E remains in a category of its own — that's not a fundamentals trade. META at 21.6x with -30.08% 52-week performance is a value trap candidate until Zuckerberg can demonstrate the $220B+ capex cycle translates to revenue.
Bottom line: I'm selectively bullish on NVDA and GOOGL as the best risk/reward pair in the Mag 7 right now. GOOGL offers compounding cloud revenue, AI search dominance, and a valuation that prices in almost no AI optionality premium. NVDA offers the infrastructure backbone for every hyperscaler's ambition, trades at a reasonable multiple for the growth profile, and has a near-term catalyst in its next earnings print. The AI spend cycle just got another major confirmation from Amazon. The question is no longer whether demand exists — it's which companies capture the margin. Right now, NVDA and GOOGL are my top two answers.