Nvidia (NVDA) shares closed at $216.85, leaving the stock at less than 33.3 times trailing 12-month earnings, near a five-year low. The AI chip leader reports fiscal 2027 second-quarter results after the market closes on Aug. 26, and investors are watching whether still-strong AI demand and record cloud backlogs can fuel a fresh rally.
The valuation reset is notable. Nvidia has gained about 21% this year, well behind the 63% rise in the PHLX Semiconductor Sector index. Its forward earnings multiple of roughly 25 sits slightly below the Nasdaq-100’s 26 times. On a price-to-sales basis, however, the stock still trades around 21 times trailing revenue, not cheap by traditional standards.
Demand signals remain robust. The four largest U.S. hyperscalers ended the second quarter with a combined $2.3 trillion in contracted backlog, according to Bank of America analyst Vivek Arya. Microsoft, Oracle, Amazon and Alphabet’s Google will need to keep investing aggressively in cloud computing capacity to work through that backlog. Oracle expects to convert only about 12% of its backlog into revenue over the next year; Microsoft expects roughly 30%. That points to continued heavy spending on AI infrastructure. Neocloud providers such as CoreWeave and Nebius, along with pure-play AI service companies, are also expanding capacity significantly.
Industry commentary adds to the bullish supply backdrop. AMD said on its latest earnings call that the total addressable market for AI compute can grow at a 40% compound annual rate over the long term, reaching nearly $2 trillion in revenue by 2030. AMD also lifted its long-term server CPU market revenue forecast from $60 billion to $220 billion, implying annual growth above 50% through 2030. Nvidia has already targeted that market, saying it can sell $20 billion of standalone server CPUs this year. Its Vera server CPUs reportedly begin selling this month, potentially helping Nvidia capture more agentic AI and inference workloads in AI data centers.
On the product side, the Vera Rubin architecture has begun shipping, and Nvidia’s CUDA programming platform plus its AI supercomputers create high barriers to entry. The company recently announced partnerships with Goldman Sachs and Blackstone to build complete data factories that can be rented to clients, and it is providing funding for OpenAI to lease a data center in Ohio for 20 years.
Wall Street’s expectations are high but may still prove conservative. Analysts project second-quarter revenue of almost $92 billion, up 97% year over year, slightly above the midpoint of Nvidia’s guidance of $91 billion. Earnings per share are expected to come in at $2.08, also roughly doubling from the prior year. Nvidia has guided non-GAAP gross margin to 75%, up from 72.7% a year earlier, which could support an earnings beat. For the current quarter, analysts are modeling EPS of $2.35, up 80% year over year. Vera Rubin processors are scheduled to begin shipping in the second half of 2026. Nvidia has disclosed an order book of $1 trillion for 2026 and 2027, compared with trailing 12-month revenue of $253 billion. Against that backdrop, some analysts may be underestimating Nvidia’s growth potential.
A stronger-than-expected report and guidance on Aug. 26 could give the stock a significant lift. But concerns remain about how quickly the AI landscape is changing, and the shares may not jump immediately after earnings. For long-term investors, Nvidia remains a core AI holding, but the stock needs a strong enough quarter to re-energize market confidence at its current below-33.3-times trailing earnings valuation.
