What did Nvidia report in its fiscal second-quarter results—including $96.22 billion in revenue, adjusted EPS of $2.46, approximately 106% year-over-year revenue growth, and beats versus analysts’ roughly $92 billion revenue and $2.09 EPS expectations—and why did its shares never

Nvidia’s fiscal Q2 2027 results were stronger than consensus, but two figures in the premise need correction: $2.46 was GAAP diluted EPS, while adjusted (non-GAAP) EPS was $2.22—not $2.46. And the latest post-release update showed the stock rising about 5%, not falling, after management forecast fiscal-2028 revenue growth of 70%. 2

• What it reported: Revenue was $96.221 billion, up 18% sequentially and 106% year over year; Data Center revenue was $89.0 billion, up 117% year over year. GAAP EPS was $2.46 and non-GAAP EPS was $2.22. 1
• Versus expectations: Revenue beat the $92.17 billion LSEG consensus by about $4.05 billion, and adjusted EPS beat the $2.10 consensus by $0.12. 2
• Outlook: Nvidia guided fiscal Q3 revenue to $108 billion, plus or minus 2%, versus the market’s roughly $104.8 billion expectation. The range encompasses approximately $105.8 billion–$110.2 billion, so it was a beat even before considering that it assumes no Data Center-compute revenue from China. 1
• Why a brief after-hours decline could still occur: With Nvidia, the market commonly trades the gap between results and unusually elevated “whisper” expectations—not merely the published consensus. Investors were looking for guidance in the $107 billion–$108 billion area, so a number at or near the lower end, cautious margin commentary, or concerns over supply/memory costs could initially prompt profit-taking even after a numerical beat. Nvidia’s supply commitments more than doubled to $279 billion, chiefly for memory procurement, which underscores both its demand confidence and its exposure to a tight supply chain. 2
• The larger investor debate: The critical question is whether customers can earn adequate returns on what may exceed $1 trillion of AI-infrastructure spending by 2026. That is a demand-quality and return-on-invested-capital issue, rather than a dispute over Nvidia’s current-quarter execution.
• Competition at inference: OpenAI’s Broadcom-developed Jalapeño is specifically designed for inference, the task of serving model responses. Broadcom said it was comparable with Nvidia Blackwell and Google TPUs, and OpenAI has claimed higher throughput and faster responses in its own testing. That is a credible long-term challenge to Nvidia’s inference economics, but it does not establish an immediate displacement of Nvidia’s broader training, networking, software, and systems position. 3 4
• “Circular financing” concern: The concern is that vendors, cloud providers, financiers, and AI labs can have reinforcing commercial and funding relationships that make hardware demand look more independent than it is. That is a valuation and credit-risk concern, but insufficient evidence supports treating it as proof that Nvidia’s reported revenue is artificial or improper.
• $500 billion financing initiative: Nvidia has partnered with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to mobilize more than $500 billion of third-party capital for AI-compute infrastructure. It can accelerate customers’ ability to build data centers—and therefore support demand—but it also increases the market’s focus on whether projects are financeable because of durable end-user cash flows rather than optimism about future AI demand. 5 6

So, the coherent reading is: the quarter and outlook were unequivocally strong, but Nvidia’s valuation remains sensitive to the sustainability, financing, and competitive structure of the AI capex cycle. In the latest update, however, the market’s ultimate immediate reaction was positive rather than negative. 2