NVIDIA just delivered Q1 FY2027 revenue of $81.62B with data center up 92% YoY, beat estimates cleanly, and guided Q2 to $89.1–92.8B — yet the stock is -1.43% today at $216.38, up +14.58% YTD and +62.94% over 52 weeks. Meanwhile, Alphabet at $386.75 is the Mag 7's quiet outperformer: +22.81% YTD, +127.13% over 52 weeks, and Google I/O 2026 just reloaded the narrative with AI Mode at 1B monthly users. These two names, deeply intertwined through the AI infrastructure buildout, are both telling you the same thing: the capex cycle is accelerating and neither is finished.
Let's start with NVIDIA because the earnings beat deserves honest context. Q1 FY2027 revenue of $81.62B versus estimates of $79.18B is not a marginal beat — that's a $2.4B upside surprise on an already enormous base, with adjusted EPS of $1.87 clearing the $1.77 consensus. Data center revenue of $75.2B nearly doubling year-over-year from $39.11B tells you demand for Blackwell infrastructure is not plateauing. The Q2 guide of $89.1–92.8B versus Wall Street's $87.3B is another beat-in-advance. And critically, approximately 50% of data center revenue came from the four hyperscalers — Microsoft, Google, Amazon, and Meta — which means NVDA's near-term revenue visibility is essentially a read-through on the capex commitments these companies have already publicly made. This is not speculative demand; it's purchase orders.
So why is the stock -1.43% today? A few dynamics worth parsing. First, the 'whisper' problem: when you guide Q1 to $78B and deliver $81.62B, the market's next question is whether Q2 at $89B midpoint is genuinely above the whisper number or just at it. Second, the China/Hopper revenue absence — no Hopper product revenue from China was reported in Q1, which is a material headwind that the street is now baking in more permanently. Third, and most important: at a $5.3 trillion market cap and 24.6x price-to-sales on $215.9B in TTM revenue, NVDA is priced for continued perfection. The 33x P/E on a trailing basis looks optically reasonable against 71% projected FY27 EPS growth, but the market is stress-testing whether the Vera Rubin ramp in H2 2026 can sustain the gross margin structure (71%+ GAAP) while absorbing competitive pressure from custom silicon at the hyperscalers. The stock is not broken — it's digesting.
Alphabet is the more interesting conversation right now. The $4.7 trillion market cap at a 29.6x P/E with $422.5B in TTM revenue and operating margins above 36% represents something increasingly rare in the Mag 7: a company that is simultaneously the incumbent it's disrupting. Google I/O 2026 was not a defensive product show — it was an offensive one. AI Mode surpassing one billion monthly users one year after debut, with queries more than doubling every quarter, is a monetization optionality story that the market is only beginning to price. The first-ever redesign of the search box in over 25 years signals that Google is not protecting Search — it is re-platforming it. Gemini Omni, Android XR Glasses, agentic search, and the price cut on AI Ultra subscriptions are all moves that suggest Google is willing to compress near-term margins to win the AI-native user generation. The +127.13% 52-week return reflects a market that has dramatically re-rated the probability that Google survives — and perhaps leads — the AI transition.
The NVDA-GOOGL connection deserves explicit attention. Google is one of NVDA's largest hyperscaler customers, contributing to that ~50% hyperscaler concentration in data center revenue. But Google is also building custom TPUs aggressively, which makes it simultaneously a customer and a competitive threat to NVDA's long-term pricing power. This is the core tension in the NVDA bull thesis: every dollar of TPU/Trainium/custom silicon that a hyperscaler deploys is a dollar not going to Blackwell. For now, demand is so far in excess of custom silicon capacity that this is a theoretical rather than realized risk — but it is the risk that will matter most in FY2028 and beyond. Jensen Huang's Vera Rubin platform with 10x inference token cost reduction is the competitive response, and the H2 2026 ramp will be the first real proof point.
From a valuation standpoint, both names sit in a zone I'd describe as 'demanding but defensible.' NVDA at 27.2x EV/EBITDA and GOOGL at 21.3x EV/EBITDA are not cheap, but they are not detached from fundamental earnings power the way Tesla (383x P/E) is. GOOGL actually screens as the more attractive near-term risk/reward given its lower beta (1.267 vs NVDA's 2.244), the AI Mode monetization catalyst ahead, and the fact that antitrust tail risks appear more manageable than feared 18 months ago. NVDA remains the single best expression of pure-play AI infrastructure demand, but the stock needs a clean Vera Rubin ramp and gross margin stability to re-rate higher from here. My stance on both names is constructive, with GOOGL having the stronger near-term narrative momentum coming out of I/O 2026.