After two years of conditioning the market to anticipate the remarkable, Nvidia now faces the fallout of those sky-high expectations. There's no dispute over its supremacy in AI accelerators, nor is there any doubt that the major cloud providers remain committed to heavy spending. The real question is whether the financial results can justify a share price that has already priced in such extraordinary performance.
What the Street is bracing for
Nvidia is slated to release its second-quarter results for fiscal 2027 after the US market closes on 26 August. Wall Street's consensus estimates point to earnings of roughly $2.09 per share on approximately $92 billion in revenue. The company's own projection was set at $91 billion, plus or minus 2%, which means a modest beat is largely anticipated and would likely do little to stir the market on its own. What investors truly want to see is a robust forecast for the upcoming quarter, as that will be the decisive factor in settling the debate over AI spending durability and any signs of digestion.
The first quarter established the benchmark. Revenue hit $81.6 billion, marking an 85% year-over-year increase, with the data centre segment alone contributing $75.2 billion and non-GAAP gross margins holding steady at 75%. Management subsequently raised guidance for the second quarter and, notably, did so without factoring in any data centre compute revenue from China. That detail is significant because it implies the upcoming figures are more straightforward than the headlines surrounding China might suggest.
Data centre remains the core narrative
The data centre business is far from an early-stage venture growing from a small base. It has already become the dominant part of the company by a wide margin, and it continues to compound at a pace that seems almost implausible given its scale. Many companies experience rapid growth when their revenue is modest, but adding tens of billions each quarter while sustaining margins in the mid-70s is an entirely different challenge, and it is the fundamental reason to hold the stock.
Two primary factors are fueling this momentum. The Blackwell ramp-up is still being integrated across the major cloud providers, and beneath that lies the broader expansion involving sovereign AI, enterprise adoption, AI clouds, and inference workloads. Jensen Huang has pegged the opportunity for Blackwell and Vera Rubin at over $1 trillion through the end of 2027. While that figure warrants some skepticism, it offers insight into how management views the trajectory of the current cycle.
Vera Rubin is now beginning to ship, alongside the new Vera CPU ramps, and spot rental prices for GPU capacity have climbed to record highs. This is a genuinely telling indicator. Customers scrambling for compute power at escalating prices is not the hallmark of a market on the verge of a downturn.
Margins take centre stage
Revenue narratives are relatively easy to come by these days. What sets Nvidia apart is its ability to scale at an unprecedented rate while still generating exceptional margins, and it is this combination that justifies its premium valuation.
On the day of the earnings release, the margin figures may reveal more than the revenue numbers. A strong top-line result paired with declining margins would immediately raise concerns about competitive pressures, product mix shifts, or rising system costs. If the company can hold gross margins near 75%, the valuation argument becomes significantly easier to defend.
The most obvious risk is memory costs. Memory now accounts for 40% to 50% of the cost of building Nvidia's systems, a sharp increase from the 15% to 20% range not long ago, and prices are escalating across the industry. Bank of America still projects margins will settle only modestly lower over time, around 73% to 74%, supported by long-term supply agreements including the partnership with SK Hynix. The impact on the upcoming Vera Rubin racks appears manageable. The more expensive pod-level systems will likely remain a small portion of the overall mix.
What analysts are saying
Bank of America is setting its sights higher than the consensus, modelling $94 billion to $95 billion in revenue and a guide toward $107 billion to $108 billion for the next quarter, well above the roughly $104 billion currently expected by the Street.
According to BofA analyst Vivek Arya, Nvidia Corp. shares could be trading at a discount of as much as 50%, as investors overstate the risks associated with the leader in artificial intelligence chips. The bank's conceptual sum-of-parts analysis on Nvidia's free cash flow suggests the stock is trading at a 34% to 50% discount, even after accounting for financing risks. The firm maintains a buy rating with a $350 price target on the stock.
Stifel's Ruben Roy reiterated a price target of $282 on Nvidia shares in a research note this week, backing the company to surpass consensus expectations for adjusted earnings of $2.09 per share on revenue of $91.96 billion. His target is based on a price-to-earnings multiple of 22 times his projection for the company's fiscal 2028 earnings.
"Earnings season has consistently reinforced the demand case…as CSP [cloud service provider] capex was raised meaningfully," Roy wrote. "We think that the other primary debates, memory cost and inference competition are likely to be expressed in gross margin rather than demand and, we believe, are partly reflected in the multiple."
Meanwhile, Oppenheimer's Rick Schafer reaffirmed a price target of $265 in a Wednesday research note, highlighting that the company trades at just 16 times his forecast for the company's 2027 earnings per share, compared with an average of more than 30 times for artificial-intelligence chip peers. "Nvidia's best-in-class margins and AI-driven structural growth make it the undisputed AI infrastructure leader. Annual upgrade cadence keeps Nvidia generations ahead of competitors. We remain long-term buyers," Schafer noted.
