Nvidia Q2 earnings target $91 billion amid hyperscaler spending scrutiny

The upcoming results will signal whether the current pace of AI infrastructure investment is sustainable for the streaming and cloud ecosystem. As companies like Meta and Amazon transition from asset-light software models to infrastructure-heavy operations, their ability to generate returns on these massive GPU investments remains unproven. For the streaming industry, rising server costs and potential 15% price hikes for AI hardware could squeeze margins for recommendation engines and automated content workflows. Watch for whether Nvidia’s data center growth continues to outpace the rising capital expenditure concerns of its eight largest cloud customers.

Nvidia's data center revenue trajectory has been fueled by an unprecedented wave of capital expenditure commitments from the largest cloud providers. In July 2025, Microsoft reported fiscal year 2025 capital expenditures of $88.2 billion, with the majority directed toward AI infrastructure buildout, signaling continued aggressive investment in GPU clusters and data center capacity. Alphabet followed with a $75 billion capex guidance for 2025, up from $52 billion the prior year, driven largely by AI compute demand. Amazon similarly raised its 2025 capex outlook to approximately $100 billion, with CEO Andy Jassy citing AI workloads as the primary driver. These figures represent the demand base that Nvidia's August 26 earnings report must validate or challenge.

The competitive landscape around Nvidia's GPU dominance is intensifying on multiple fronts. In May 2025, Huawei unveiled its Ascend 910D chip, which the company claims matches Nvidia's H100 in performance for AI training workloads, representing the most credible domestic Chinese alternative amid ongoing US export restrictions. Meanwhile, TrendForce forecast that global AI server shipments would grow 28% year-over-year in 2025, with Nvidia maintaining over 80% market share in discrete GPU-based AI accelerators. Oracle has emerged as a notable new buyer, with the company disclosing $12 billion in Nvidia GPU orders for its OCI AI infrastructure expansion in early 2025. Meta, which committed $65 billion in capital expenditure for 2025 with a focus on AI data centers, remains one of Nvidia's largest single customers.

On the technical side, Nvidia's Blackwell architecture is becoming the critical variable for next-quarter performance. TSMC reported that CoWoS advanced packaging capacity, which is essential for Blackwell GPU production, remained fully booked through Q3 2025, indicating supply constraints could cap revenue upside even if demand holds. IDC projected that global spending on AI infrastructure would reach $250 billion in 2025, with GPUs accounting for roughly 60% of total AI hardware expenditure. For streaming companies relying on GPU-accelerated recommendation engines, video transcoding pipelines, and generative AI content tools, any signal of supply tightness or pricing pressure from Nvidia's earnings could directly affect operational cost structures through 2026.