Nvidia Just Slid 5% Post-Earnings. Again. CEO Jensen Huang Insists "This Time Is Different." Is the Stock Still a Buy?

Nvidia reported Q2 revenue of $96.2B (up more than 100% YoY) and adjusted EPS of $2.22 (+120% YoY), beating estimates ($96.2B vs consensus; $2.22 vs $2.09). The stock surged intraday (up ~9.9%, topping $230.39) after projecting 70% revenue growth for 2027 versus analysts’ 44% forecast, but then fell 5.5% by Friday, leaving shares near pre-earnings levels—signaling investor skepticism about sustained AI spend.

Analysis

The tape is telling us this is no longer a simple beat/raise story; the market is now trading NVDA as a sentiment proxy for whether AI capex is a one-cycle air pocket or a multi-year infrastructure build. That creates a weird asymmetry: strong execution can still underperform if investors think spend is peaking, but any evidence of budget re-acceleration should force rapid multiple repair because the company is converting demand into cash flow at a scale few large caps can match.

The bigger second-order winner is the infrastructure chain, not just the chipmaker. If AI spend is truly industrialized, the marginal dollar tends to migrate into power, cooling, networking, advanced packaging, and memory, which is why ANET, VRT, MU, TSM, and AMAT matter more than the broad "AI software" basket over the next 3-6 months. By contrast, high-multiple software names that monetize AI mostly through optionality may see valuation pressure even if end-user adoption keeps rising, because capital is still chasing the toll booths.

Contrarian takeaway: consensus may be overweighting the risk of a near-term digestion pause and underweighting the duration of the spending cycle, but the stock can still tread water if investors need another quarter of proof. The main falsifier is a softer hyperscaler capex guide or a failure to hold the post-earnings gap into the next 4-8 weeks; absent that, the pullback looks more like positioning exhaustion than fundamental deterioration.

Key Decisions for Investors

• Buy NVDA on an additional 3-5% weakness via a 1-3 month call spread; upside is a re-rating back through the post-earnings resistance if next capex commentary confirms no slowdown, with defined premium risk.
• Express the quality-vs-narrative view with long NVDA / short a basket of high-multiple AI software names over the next 1-3 months; if AI spend is real, market should favor earnings visibility over story stock beta.
• Rotate into infrastructure beneficiaries on a 3-6 month horizon: long ANET/VRT/MU/TSM over broader software or consumer internet proxies; this captures the second-order spend pool if the AI buildout remains industrial.
• Set an alert on the next hyperscaler capex updates; if forward spend or procurement cadence softens, cut NVDA longs and infrastructure exposure quickly, because that is the cleanest falsifier of the thesis.