NVIDIA's Jensen Huang has signaled $1 trillion in order visibility through 2027 — a $500B increase since October 2025 — while Blackwell Ultra ramps and Rubin looms on the 2026 horizon. Alphabet, meanwhile, is committing $175-185B in 2026 capex on the back of Google Cloud's 48% YoY Q4 growth and a $240B backlog, yet trades at just 27.4x TTM earnings — cheaper than AAPL and close to AMZN. Both names are BULLISH; NVDA is the higher-beta conviction play, GOOGL is the stealth value story hiding in plain sight.
Let's start with NVIDIA, because the order-book revelation from GTC 2026 deserves more credit than the market is giving it. CEO Jensen Huang flagged visibility into at least $1 trillion of demand through 2027 — that's $500 billion of net new orders added since October 2025 alone. New Street Research's Pierre Ferragu picked up on this immediately, adding NVDA to their best ideas list on March 19 and noting the company is already running at a $1T+ annual revenue run rate trajectory. At $177.39 with a $4.31T market cap and 36.2x TTM P/E, NVDA isn't cheap in the traditional sense — but 19.97x P/S on $215.9B in TTM revenue against a $1T forward demand signal starts to look very different. The stock is -6.06% YTD after a massive +74.29% 52-week run, and I'd argue the YTD weakness is a digestion period, not a thesis break.
The product roadmap is what makes the valuation case holding. Blackwell Ultra is ramping quickly. Rubin is targeted for 2026 launch. NVIDIA Groq 3 LPX is slated for H2 2026. And Huang is deepening the ecosystem moat with a $2B strategic investment in Marvell focused on silicon photonics — this isn't just supplier management, it's vertical integration of next-gen interconnect technology that will matter enormously when data center scale goes from megawatt to gigawatt. JPMorgan raised their price target to $265 (from $250) maintaining Overweight. At $177.39, that's roughly 49% implied upside to the JPM target. My own read: the $265 target is probably conservative if $1T in 2027 revenue materializes. Beta of 2.375 means you'll feel every macro tremor, and the short ratio of 1.28 suggests the skeptics aren't piling on — which is a subtle positive.
Now pivot to Alphabet, which I think is one of the most mispriced large-cap AI stories heading into mid-2026. At $295.77 — down -6.08% YTD and up +96.92% over 52 weeks — GOOGL trades at 27.4x TTM earnings with a $3.58T market cap. Consider what you're getting: Google Cloud Q4 2025 revenue grew 48% YoY to $17.6B, Cloud operating income grew 30.1% to $5.3B, and the Cloud backlog hit $240B — up 55% sequentially and more than doubled year-over-year. That backlog number is extraordinary. It means Google Cloud has years of contracted revenue locked in at accelerating rates. The $175-185B capex commitment for 2026 is either the most aggressive infrastructure bet in corporate history or the necessary cost of competing for a multi-trillion-dollar AI infrastructure market. I lean toward the latter.
The valuation case for GOOGL is straightforward: at 27.4x TTM, it's cheaper than AAPL (32.4x), cheaper than TSLA (by a country mile), and trades at a modest premium to MSFT (23.4x) — but GOOGL's Cloud growth rate, Search advertising resilience, and Gemini 3.1 Pro positioning arguably warrant that premium. The EV/EBITDA of 19.46x is reasonable given the growth profile. The $402.8B TTM revenue base with a 32.8% profit margin and 31.6% operating margin shows the legacy cash generation machine is funding the AI buildout without equity dilution. The February 2026 Gemini 3.1 Pro release demonstrates Google is competitive at the frontier model level — and with $240B in Cloud backlog, enterprise customers are voting with their wallets.
Relative to the broader Mag 7, here's my current read: NVDA is the highest-conviction growth name with the clearest demand signal and the most compelling product cycle. GOOGL is the most attractively valued AI infrastructure + application story at current prices. MSFT remains solid at 23.4x but -20.85% YTD is a significant underperformance story that needs Azure Q3 numbers to recover the narrative. AAPL at 32.4x with -5.48% YTD is still pricing in a supercycle that hasn't printed in the revenue line. My stance across the pair: BULLISH on both NVDA and GOOGL, with NVDA as the higher-beta expression of AI infrastructure spend and GOOGL as the more risk-adjusted compounder for investors with lower volatility tolerance. The $1T NVDA order book and GOOGL's $240B Cloud backlog are the two most important data points in Mag 7 land right now.