Nvidia is too cheap versus chip peers, Bank of America argues
Bank of America's call reframes Nvidia's underperformance as a valuation anomaly rather than a fundamental warning sign, arguing the stock's next-12-month price-to-earnings ratio near an 11-year low leaves room for a re-rating once August earnings land. The note pushes back directly on two of the market's biggest overhangs, rising high-bandwidth memory costs and custom-chip competition from hyperscalers, framing both as manageable given Nvidia's pricing power and scale. If Arya's mid-70% gross margin outlook holds through the Rubin platform transition, it would undercut the bear case that memory inflation is compressing Nvidia's economics. The relative valuation gap to Microsoft and Apple, despite comparable AI exposure, is likely to keep drawing attention from investors positioning ahead of the print.
---
BofA says the market is punishing Nvidia far more than its fundamentals justify.
Summary:
- Nvidia shares have risen only about 3% this year through Monday, versus an over 80% gain for the PHLX Semiconductor Index
- BofA's Vivek Arya calls the underperformance an enhanced buying opportunity for Nvidia stock
- Rising high-bandwidth memory costs have worried investors about gross margins, but Arya expects pricing power on the Rubin platform to keep margins in the mid-70% range
- Concerns over custom chips such as Google's TPUs are overstated, Arya argues, noting Nvidia's GPU revenue has risen 700 times over the past decade despite TPUs existing that whole time
- Nvidia trades at roughly half its 10-year average forward P/E ratio, near an 11-year low, and at a discount to Microsoft and Apple
- Arya expects Nvidia's August earnings report to reaffirm its moats in products, pricing and supply chain
Nvidia's stock has badly lagged the broader chip sector this year, and Bank of America says that gap has created what it calls an enhanced buying opportunity rather than a red flag, according to MarketWatch /Dow Jones (may be gated). Despite dominating the market for artificial-intelligence chips under CEO Jensen Huang, Nvidia shares rose only about 3% in the year through Monday, while the PHLX Semiconductor Index climbed more than 80% over the same stretch.
BofA analyst Vivek Arya acknowledged the concerns weighing on the stock but argued the market's reaction has been overly harsh. Rising costs for high-bandwidth memory components have fed worries about Nvidia's gross margins, but Arya believes investors are overstating that pressure while underestimating Nvidia's pricing power, scale and roughly 119 billion dollars in supply-chain commitments. He expects Nvidia can offset higher HBM costs per rack because pricing on its next-generation Rubin platform could run well above that of the current Blackwell lineup, keeping gross margin in the mid-70% range.
A second source of investor anxiety, competition from custom chips such as the tensor processing units Google co-designs with Broadcom, also looks overdone to Arya. He pointed out that Google's TPUs have existed for more than a decade, a period in which Nvidia's GPU revenue has grown 700-fold, and expects Nvidia to hold between 65% and 70% of hyperscaler AI infrastructure spending over the long term.
Valuation adds to the case. Nvidia's forward price-to-earnings ratio sits at 18.69, nearly half its 10-year average of 36.90 and close to an 11-year low, leaving it trading at a discount to peers like Microsoft and Apple despite comparable AI exposure and similar memory-cost pressures. Arya expects Nvidia's August earnings report to demonstrate that its moats in products, pricing and supply chain remain intact.
This article was written by fl6553e4b45d84486a91658a8b3f02bf22 at investinglive.com.提供 MainLink：Investinglive RSS Breaking News Feed
FX初心者には必須 無料のうちにGET！
