Nvidia Q2 Earnings Set $93B Test for AI Stocks

Nvidia's August 26 earnings have Wall Street targeting $93-95B in Q2 FY2027 revenue as Blackwell demand faces AMD and hyperscaler custom silicon competition.

Lead

What Does Wall Street Expect From Q2?

Non-GAAP earnings per share consensus sits near $2.07, up from $1.73 in Q1. The data center segment, which generated $75.2 billion in Q1 - 92% year-over-year growth - is forecast to deliver approximately $87 billion in Q2, maintaining roughly 90% of total company revenue. Gross margins are expected to hold near the 73-75% range reported across recent quarters. Nvidia's own guidance issued in May called for Q2 revenue of $91 billion, plus or minus 2%, placing the consensus about $2-4 billion above the midpoint and embedding a beat expectation before results are published. Q3 guidance, to be issued alongside results, will carry significant weight; markets are watching for a midpoint above $95 billion as confirmation that the $100 billion quarterly threshold remains on track for fiscal 2028.

Why Does This Quarter Define the AI Sector's Direction?

Nvidia's quarterly revenues have grown from $22 billion in Q2 FY2024 to $81.6 billion in Q1 FY2027, a trajectory that has anchored the broader bull case for AI-infrastructure equities. SMH and SOXL, which track the semiconductor sector with standard and leveraged exposure respectively, are acutely sensitive to any revision in the demand narrative. Softer guidance, evidence of inventory buildup, or margin compression below 72% would reverberate across the Nasdaq complex and reset valuations across the entire AI capital expenditure chain. Conversely, a revenue beat combined with guidance above $95 billion would reinforce the view that hyperscaler spending on GPU clusters remains structurally elevated through at least 2027.

What Is the Competitive Threat From AMD and the Cloud Giants?

What Happened With China, and Does It Change the Picture?

The China revenue question - which produced a $4.5 billion inventory charge in Q1 FY2026 when H20 export licenses were revoked - has partially resolved. The U.S. government authorized Nvidia to resume H200 chip sales to Chinese customers in early 2026, capping annual shipments at 75,000 units under a framework that also imposes a 25% surcharge. Nvidia CEO Jensen Huang confirmed license receipt in March 2026, ending a roughly 10-month supply freeze. The arrangement restores a material revenue channel, but the volume cap and tariff-driven margin compression limit upside, and the prolonged interruption accelerated Chinese hyperscaler investment in domestic chip alternatives, creating a structural headwind that the H200 framework does not fully reverse.

Blackwell Ramp and the Rubin Architecture Pipeline

The Blackwell platform is the central execution variable. Nvidia claims up to 4x performance improvement over the prior Hopper generation for large language model training, and adoption by major cloud providers is advancing. Markets are monitoring Blackwell shipment volume closely as the leading indicator of whether the company can sustain 14-17% sequential revenue growth through fiscal 2028. Looking further ahead, Nvidia is preparing the Rubin architecture for 2027 deployment, maintaining a product cadence that has historically outpaced competitive timelines and preserved its software ecosystem advantage.

Outlook

The August 26 report will validate or challenge the foundational assumption behind much of the current AI investment cycle - that Nvidia's hardware and CUDA software ecosystem constitute a durable competitive moat against a field that now includes AMD, four of the world's largest technology companies deploying custom silicon, and a Chinese market adjusting to constrained supply. With Q1 FY2027 revenue already 85% above the prior year and four consecutive quarterly records on the books, the question is no longer whether demand is real but whether it is broad and durable enough to justify valuations near $2.3 trillion. The guidance figure for Q3 will answer that question more definitively than the Q2 revenue line itself.