Nvidia's Demand Outlook Still Supports The Bull Case

Nvidia remains rated Buy after Q2 FY27 data-center revenue surged 117% year over year to $89 billion, driven by hyperscale and ACIE customer demand for AI infrastructure. GAAP gross margin reached 75%, while management expects a temporary decline before margins stabilize at 72–73% in fiscal 2028 as pricing resets help offset higher costs. The results and outlook reinforce Nvidia's full-stack AI platform leadership and sustained AI-driven growth trajectory.

Analysis

The key underwriting issue is no longer demand visibility but the durability of incremental economics as system-level content rises. A 72-73% normalized gross-margin floor would support materially higher earnings power than the market typically assigns to hardware cycles, provided NVDA retains control of the software, networking and rack-scale integration layers that make replacement difficult. This shifts the relevant competitive comparison from GPU unit share to total AI-cluster spend captured per customer, favoring NVDA over merchant silicon alternatives such as AMD and custom-ASIC suppliers.

Near term, upside depends on whether hyperscaler capex remains broad-based rather than being financed by a small set of AI-native customers with potentially circular economics. The most consequential 1-3 month catalyst is evidence that backlog converts without rising receivables, inventory, or customer-concentration pressure; confirmation would reduce fears that revenue is being pulled forward ahead of a platform transition. Conversely, any indication of extended deployment schedules, lower networking attach, or margin guidance below the stated normalization range would likely trigger disproportionate multiple compression because expectations already embed execution near perfection.

The contrarian point is that a modest gross-margin decline is not inherently bearish if it reflects a deliberate mix shift toward integrated systems with higher dollar gross profit and deeper customer lock-in. The more underappreciated second-order beneficiary is AVGO: expanding AI clusters require high-bandwidth connectivity and custom compute, even where NVDA remains the primary platform owner. For 6-18 months, the larger risk is that hyperscaler-developed ASICs improve enough for inference workloads to become a meaningful price-discovery mechanism, limiting NVDA's ability to sustain platform-level pricing rather than merely reducing accelerator share.

Key Decisions for Investors

• Maintain/add NVDA on post-results weakness rather than chase strength; use a 3-6 month horizon and target a 10-15% drawdown entry window. Thesis is sustained earnings-power rerating if gross margin holds at or above 72% while revenue conversion remains clean; exit/reduce on a sub-70% normalized-margin outlook or a material cut to next-quarter revenue guidance.
• Express bullish exposure with an NVDA 6-month call spread, buying approximately 5% out-of-the-money calls and selling 20-25% out-of-the-money calls. This retains upside to another estimate-revision cycle while capping premium exposure in a name vulnerable to sharp valuation de-rating on even minor execution misses.
• Pair long NVDA / short AMD over the next 1-3 months only if NVDA's next disclosure confirms stable margin and broad customer demand. The spread isolates platform monetization versus accelerator competition; stop out if AMD demonstrates material AI revenue upside accompanied by improving gross margin, rather than growth driven by lower-priced share gains.
• Add AVGO as a complementary long rather than treating it as a direct NVDA hedge; hold 6-12 months for networking/custom-silicon attach. Reassess if hyperscaler capex guidance weakens or if AI-networking revenue fails to accelerate, as those would indicate cluster spending is being deferred rather than merely reallocated.