NVIDIA Corporation (NVDA.US) stock price flashes a "warning signal": valuation falls to its lowest in more than a decade, what is the market worried about?

NVIDIA Corporation (NVDA.US) stock is flashing a warning signal about whether its profit growth can be sustained. But the paradox is that this signal does not come from the company's own performanceit comes from the way the market is pricing it. As of mid-September, NVIDIA Corporation's price-to-earnings ratio based on expected profits over the next 12 months had fallen below 17 times, close to its lowest level in more than a decade. That multiple is only half of its 2025 level and far below the valuation of more than 25 times as recently as May. Given that NVIDIA Corporation has beaten revenue and profit expectations for several consecutive quarters, this valuation compression stands out even more. Eli Horton, a senior portfolio manager at TCW, said the extent of this valuation "de-frothing" reveals a considerable degree of market skepticism about the sustainability of the company's current profitability"Given the incredibly strong fundamental backdrop, the stock's performance is surprising, but it tells you that market expectations are below current analyst consensus." In fiscal 2026 (ending January 2026), NVIDIA Corporation's full-year revenue reached $215.9 billion, up 65% year over year, with GAAP net income of $120.07 billion. Fourth-quarter revenue was $68.1 billion, with the data center business contributing $62.3 billion, up 75% year over year. Entering fiscal 2027, growth accelerated furthersecond-quarter revenue reached $96.2 billion, surging 106% year over year, while net income doubled to $59.7 billion and gross margin remained at a high of 75%. For the third quarter, management provided revenue guidance of $108 billion, a jump of 18.7% from the previous quarter's guidance of $91 billion, marking the first time NVIDIA Corporation has forecast quarterly revenue above $100 billion. For the full year, the company expects fiscal 2028 revenue to grow about 70%, far exceeding the previous market consensus of 44%. As of mid-September, NVIDIA Corporation's share price gain in 2026 was only 22%, ranking second among the U.S. stock market's "Magnificent Seven," behind only Apple Inc. (AAPL.US) at 25%. However, the Philadelphia Semiconductor Index rose as much as 76% over the same period, with Micron Technology, Inc. (MU.US), Intel Corporation (INTC.US), and AMD (AMD.US) all gaining more than 180%. NVIDIA Corporation ranked fifth from the bottom in the semiconductor index. This divergence has left Jensen Huang rather displeased. At the Goldman Sachs Group, Inc. technology conference in early September, he declared that NVIDIA Corporation is "the world's first and only growth value stock" and said the company is "severely misunderstood." The logic he repeatedly emphasized is that the company is not only growing, but also capturing more value while growing. The question is, why isn't the market buying it? Gross Margin: A Variable Repriced Under the Halo of Growth Pressure on NVIDIA Corporation's profitability stems largely from rising costs of key components, especially memory chips. In its second-quarter earnings report, NVIDIA Corporation disclosed that gross margin will gradually decline from 75%, with expectations that it will bottom out at 71% to 72% in the fourth quarter before stabilizing at 72% to 73% in fiscal 2028. Huang's attitude toward this was quite candid. He said the company "decided to rip off the band-aid and reset the market's expectations for gross margin," acknowledging that it has already absorbed the pressure of rising costs while also repricing its products. Behind this downward revision to gross margin is an easily overlooked transmission chain: the AI buildout boom that NVIDIA Corporation itself is leading is driving up its own costs. According to a bill of materials previously disclosed by UBS Group AG, memory costs for NVIDIA Corporation's next-generation Vera Rubin AI platform have exploded, with memory's share of total system cost soaring from 53% in the previous-generation Grace Blackwell system to 62%, making it the most expensive component of the entire platform. DRAM contract prices rose 58% to 63% quarter over quarter in the second quarter of 2026, and pricing power is shifting from NVIDIA Corporation to memory chip suppliers. David Russell, global head of market strategy at TradeStation, is more deeply concerned. He believes that as NVIDIA Corporation's largest customers develop their own chips, competition will only intensify"Companies want to reduce their dependence on NVIDIA Corporation, so it is very conceivable that its market position weakens over time, which means gross margin is more likely to decline than improve, and that is a big problem for investors," Russell said. "Valuation multiples expand when companies are in a favorable position and have the potential to get better, and NVIDIA Corporation does not have that." What Russell described is not hypothetical. Alphabet Inc. Class C's (GOOGL.US) TPU has already gone from an internal project to a genuine commercial weapon. Thomas Kurian, head of Alphabet Inc. Class C's cloud business, revealed at the Goldman Sachs Group, Inc. conference that the TPU business is already more than twice the size of the comparable business of the second-largest hyperscale cloud provider, and that servers equipped with its self-developed chips have a payback period of less than one year. Alphabet Inc. Class C has begun delivering TPU systems directly to customer data centers and has signed customers such as Anthropic. At Meta (META.US), the self-developed AI chip code-named "Iris" is scheduled for mass production in September, and testing found no major issues in just six weeks. Meta's goal is to increase overall AI compute to 14 gigawatts next year. These self-developed chips will not overturn NVIDIA Corporation's roughly 90% share of the AI accelerator market in the short term. But they are changing the bargaining structure in a perceptible way. JPMorgan analyst Harlan Sur expects that over the next few years, market share between NVIDIA Corporation GPUs and custom chips such as ASICs and XPUs will gradually converge. The "Trees Don't Grow to the Sky" Problem of Capital Expenditure NVIDIA Corporation's current growth story is essentially tied to one assumption: that hyperscale cloud service providers' capital expenditure will continue to climb. That assumption still appears valid in the short term. The four major hyperscalersAmazon.com, Inc., Alphabet, Microsoft Corporation, and Metahave combined capital expenditure plans of about $750 billion for 2026, up roughly 70% from 2025. Wedbush analysis has pointed out that about 60% of this round of investment flows to NVIDIA Corporation's GPUs and supporting hardware, while large technology companies are increasingly becoming among the world's largest issuers of corporate debt. But that is precisely where the problem lies. TCW's Horton said: "The prudent approach is to step back and think about whether all this spending is sustainable, because trees don't grow to the sky." He further analyzed that NVIDIA Corporation's current valuation actually already implies expectations of a slowdown in AI capital expenditurewhether from hyperscalers voluntarily scaling back investment or from regulatory frameworks delaying or halting projects. However, Horton believes neither scenario is very likely at present, which makes NVIDIA Corporation's valuation look relatively attractive. This mismatch"valuation implies pessimistic expectations, but reality may not deliver them"is at the core of the current bull-bear divide. Research from Morgan Stanley and BofA Securities both noted that NVIDIA Corporation's forward P/E of around 18 times implies an assumption that the company will have almost no growth after 2027. Yet analysts' actual expectations are for fiscal 2027 revenue of about $394 billion, up 82% year over year, and further growth to about $561 billion in fiscal 2028. BofA maintained its "Buy" rating on NVIDIA Corporation with a price target of $350, based on 26 times calendar 2027 earnings per share. The firm noted that NVIDIA Corporation's sales compound annual growth rate from 2025 to 2028 could reach 48%, with EPS CAGR of 52%, while non-NVIDIA Corporation large tech peers are expected to see sales and EPS CAGRs of only 16% and 15%, respectively, over the same period. On that basis, NVIDIA Corporation's 2027 PEG is only 0.3 times, far below the peer average of 1.6 times. The situation NVIDIA Corporation currently faces is less a deterioration in fundamentals than a market repricing of the long-term risks in the AI trade. Rising memory costs eroding gross margin, customer self-developed chips shaking bargaining power, and the unprovable sustainability of capital expenditurethese are all real risks, but whether they are enough to justify a P/E of less than 17 times is clearly a matter on which the market is divided. Horton said no one knows how the stock will move, "but the current risk-reward setup is very attractive, and I like this probability distribution. If the question is whether the valuation will repair upward or continue to drift lower, I would definitely choose the former. As an entry point, this P/E level looks quite favorable."