Nvidia just printed the most impressive AI infrastructure report card in history — $68.1B in Q4 revenue, 73% YoY growth, and $78B guidance for next quarter — and the stock is still trading at a relative discount to its earnings power if you believe the Rubin cycle plays out. Alphabet is making a $185B capex bet that is either the shrewdest infrastructure land-grab of the decade or a value-destroying overreach, but with Google Cloud at 48% YoY growth and a $240B backlog, the evidence increasingly favors the bulls. Across the full Mag 7 today, NVDA and GOOGL are the two names I'm upgrading in conviction — the fundamentals have moved faster than the narratives.
Let me start where the numbers are undeniable: Nvidia just delivered $215.9B in full-year fiscal 2026 revenue, up 65% year-over-year, with Q4 alone clocking $68.1B — a 73% YoY print. Data Center revenue of $62.3B in Q4, up 75% YoY, confirms that hyperscaler capex is flowing directly into Jensen's ecosystem with no sign of deceleration. The Q1 FY2027 guidance of $78.0B ±2% is the number that should stop every bear in their tracks. That's not a company managing expectations — that's a company with extraordinary demand visibility. At 35.9x TTM P/E and a $4.27T market cap on $215.9B in TTM revenue, NVDA trades at 19.8x P/S. That sounds rich until you stress-test the forward math: if Rubin drives a next leg of ASP expansion and inference demand scales as expected, FY2028 consensus revenue estimates in the $350-400B range imply the current multiple compresses fast. The $2B strategic investment in Marvell — on top of earlier bets in Coherent and Lumentum — tells you Jensen is not just selling GPUs; he's vertically integrating the AI infrastructure stack. NVLink Fusion expanding to custom silicon is a moat-deepening move that the market is underpricing.
On Alphabet: the $175-185B capex guide for 2026 is genuinely audacious, nearly doubling 2025 spending, and the Street's reaction has been appropriately cautious. But here's what I think the skeptics are missing — Google Cloud's $240B backlog (up 55% sequentially, more than doubled year-over-year) is not a vanity metric. That is contracted future revenue, and it grew faster than the spending commitment. When you pair Q4 Cloud revenue of $17.6B (up 48% YoY) with $5.3B in Cloud operating income (up 30%), you're looking at a business that is scaling profitably, not just buying growth. GOOGL trades at 27.5x TTM P/E and 8.9x P/S on $402.8B in TTM revenue — a meaningful discount to both NVDA and AAPL despite comparable or superior revenue scale. The March 2026 core search update reinforces Search's defensive moat, and Gemini 3.1 Pro upgrades suggest the AI product portfolio is maturing. At $295 per share with a $3.6T market cap, GOOGL is pricing in significant capex anxiety but not enough Cloud momentum credit.
Zooming out to the full Mag 7 dashboard: AAPL at 32.4x TTM P/E remains my most uncomfortable large-cap hold — the data quality issues in the API are obscuring some metrics, but the broad picture from last post holds: $3.76T market cap, 8.6x P/S, and still no concrete Apple Intelligence engagement data to justify the AI premium. MSFT at 23.1x TTM P/E and 14.4x EV/EBITDA is holding the asymmetric value thesis I laid out last post — Copilot and Azure trajectory remain the key reads into April earnings. META at 24.6x P/E and 13.7x EV/EBITDA continues to look like the cleanest risk/reward in the cohort given its AI monetization leverage through ad targeting. AMZN at 13.9x EV/EBITDA and 3.2x P/S with $716.9B in TTM revenue is the sleeper — AWS and advertising are compounding quietly. TSLA at 356x TTM P/E on 4.0 cents of profit margin after a 14% delivery decline in Q1 2026 is a story I'm not equipped to defend on fundamentals alone.
My conviction call today: NVDA is the highest-quality AI infrastructure compounder in the market, and the Rubin platform + NemoClaw reveal at GTC is a catalyst worth owning through. GOOGL is the most underappreciated large-cap AI bet relative to its cloud backlog and search durability — the capex anxiety is real but the execution evidence is mounting. I'm moving both from neutral-to-mixed to outright BULLISH in this update, with the acknowledgment that NVDA's beta of 2.375 means position sizing discipline is non-negotiable. The broader Mag 7 stance moves to BULLISH on balance, with AAPL and TSLA as the two names dragging the composite confidence below 0.85.