Nvidia Q2 Revenue Hits $96.2B as Memory Costs and Circular Financing Concerns Loom

Key points

• Nvidia's Q2 revenue reached $96.2 billion, beating the $92.29 billion analyst estimate, with adjusted EPS of $2.22 versus the $2.09 forecast.
• Data center revenue hit $89.0 billion, up 117% year-over-year, driven by sustained demand for Blackwell and the ramp of Vera Rubin platforms.
• CFO Colette Kress warned of 'extreme' memory pricing conditions, with supplier commitments rising from $119 billion to $279 billion to secure components through fiscal 2028.
• Nvidia provided its first full-year revenue guidance, projecting 70% growth for fiscal 2028, a figure significantly higher than the 45% analysts had expected.
• The company announced an expansion of its partnership with AWS, adding 2 million GPUs, and confirmed strategic financing deals with Apollo, BlackRock, and others to mobilize over $500 billion in third-party capital.
• Nvidia excluded China from its Q3 outlook due to geopolitical uncertainty, noting less than 1% of Hopper 200 revenue came from Chinese customers in the prior quarter.

Background

Nvidia has faced scrutiny over its financing models, with recent archive coverage highlighting concerns about circular financing and memory cost pressures impacting hyperscaler capex. Previous reports noted Amazon's capex rising to $54.2 billion due to memory inflation, a trend Nvidia now addresses through supplier commitments and revenue-share deals. The company's push into CPU markets and agentic AI infrastructure marks a shift from pure GPU dominance to a broader platform strategy.

How outlets are covering it

Business Insider and CNN emphasize the 'circular financing' debate, with CFO Kress defending revenue-share deals as distinct from loans, while analysts question the sustainability of hyperscaler spending. UBS and Jefferies view memory cost inflation as bullish for Nvidia, expecting margin resilience and strong Rubin ramp adoption. JPMorgan and BofA focus on competitive dynamics, noting that while OpenAI and Google develop custom chips, Nvidia's platform approach and open-model support maintain its market position. Huang dismissed open-source models as a threat, arguing they drive sales on Nvidia hardware, though Z.ai's GLM-5.3-Flash exception on Chinese chips was noted.

Why it matters

Nvidia's guidance signals that AI infrastructure demand remains robust despite memory bottlenecks and competitive pressures. The company's shift to full-year guidance and expanded financing partnerships indicates a maturing AI market where capital allocation and supply chain security are critical. Investors are watching whether hyperscalers can sustain capex amid negative free cash flows, a key risk for Nvidia's long-term growth trajectory.

What to watch

Nvidia will monitor memory supply constraints through fiscal 2028 and track the adoption of Vera Rubin and CPU products. The company expects to see returns on its financing investments as compute platforms are redeployed. Analysts will assess whether Q3 revenue hits the $108 billion target and if gross margins hold at 74% despite memory cost pressures.