Despite a wave of positive earnings from the artificial intelligence sector during this U.S. stock earnings season, investors have been selling off NVIDIA (NVDA.US) shares instead of increasing their holdings. After hitting a record closing high on April 27, NVIDIA's stock price fell for six consecutive sessions, accumulating a drop of 9%. During the same period, the Philadelphia Semiconductor Index rose over 5%, with NVIDIA ranking as the third-worst performer among the index's components.
The reason lies in the dual challenges NVIDIA's near-total dominance in the AI processor market now faces from other chipmakers and its own major customers, despite continued heavy investment in computing power equipment by tech giants. Last week, Alphabet (GOOGL.US) announced it would make its Tensor Processing Units (TPUs) available to select customers for building their own data centers. Amazon.com's (AMZN.US) custom AI chip series, Trainium, has secured orders representing over $225 billion in revenue and recently gained a multi-billion dollar procurement commitment from Meta Platforms, Inc. (META.US), which is also developing its own AI chips for deployment. Meanwhile, Intel (INTC.US) continues to benefit from AI industry growth, and Qualcomm (QCOM.US) is making steady progress in the data center chip market.
Bill Stone, Chief Investment Officer at Glenview Trust Company, stated, "When a company holds nearly 100% market share, the only direction from there is down. These new entrants now possess the capability to become strong competitors to NVIDIA." For now, there is no clear evidence that rivals are significantly eroding NVIDIA's market share. Bloomberg Intelligence research data indicates NVIDIA will maintain an 86% share of the AI accelerator market in 2025, unchanged from 2024. However, competitive concerns are raising doubts about its long-term growth prospects and enhancing the relative appeal of other chip investments.
NVIDIA's stock is up only 5% year-to-date, roughly in line with the S&P 500 but far behind gains seen by other chip companies. The Philadelphia Semiconductor Index has surged 55% this year, and NVIDIA is the worst performer in the 30-stock index for 2026. Stone, who manages approximately $18 billion in assets, noted, "If we see marginal business losses for NVIDIA and competitors gradually eating into its market share and pricing power, its profit growth momentum will come under pressure, and the stock price will likely weaken accordingly." An NVIDIA spokesperson declined to comment, citing the quiet period ahead of earnings.
The global surge in demand for AI computing equipment previously propelled NVIDIA to become the world's most valuable company, with a market capitalization once reaching $4.8 trillion. However, its leading position is now precarious, with Alphabet's market value rapidly approaching. Over the past year, driven by its Gemini large language model, cloud computing, and custom chip businesses, Alphabet's market cap has soared by over $2.5 trillion, reaching approximately $4.7 trillion at Tuesday's close.
NVIDIA's revenue growth remains strong. Revenue for the current fiscal year (ending January next year) is projected to surge 70%, significantly outpacing other tech giants and exceeding the previous year's 65% growth. However, market expectations are for growth to slow to 32% in fiscal 2028, with further deceleration in the following two years. The company is scheduled to report first-quarter earnings on May 20.
Bullish investors argue that demand for AI processors remains robust, and industry growth is sufficient to support performance gains for all players. The four major tech giants—Alphabet, Amazon.com, Meta Platforms, Inc., and Microsoft (MSFT.US)—have planned capital expenditures totaling $725 billion for this year, with significantly higher investment levels expected by 2027. According to compiled supply chain data, these four companies collectively contribute about 45% of NVIDIA's revenue.
Nevertheless, Bank of America analyst Vivek Arya pointed out in an April 29 report that these tech giants are now widely adopting heterogeneous computing deployments—utilizing a mix of NVIDIA chips and their own custom chips. Notably, Alphabet's TPU is considered one of the most viable alternatives to NVIDIA's products. This chip, specifically designed for accelerating machine learning workloads, is seen as Alphabet's "ace in the hole." The market has recognized the growth potential of this business; Citizens analyst Andrew Boone estimates that Alphabet's TPU-related infrastructure revenue will reach $3 billion in 2026 and is expected to surge to $25 billion in 2027.
Clayton Allison, a portfolio manager at Prime Capital Financial, which manages around $40 billion, said, "I wouldn't say NVIDIA's competitive position is severely threatened by these new chips, but the market's reaction to NVIDIA reflects growing questions about its market share, competitive moat, and profit margins."
On Tuesday Eastern Time, AMD (AMD.US) stated that demand is set to increase significantly as general-purpose CPUs play a growing role in running AI services. Despite the underlying competitive pressures, Wall Street analysts remain largely favorable towards NVIDIA. Among the 80 analysts tracked, only three recommend "Hold" and one suggests "Sell." Over the past quarter, earnings estimates for fiscal 2027 have been raised by 11%, with revenue expectations also revised upwards; estimates for fiscal 2028 have seen even larger increases, indicating continued institutional confidence in its long-term trajectory.
Glenview's Stone concluded, "NVIDIA was once the undisputed, safe play on AI; that halo has dimmed. Alphabet has become the market's new favorite AI darling. The new chips won't be a fatal blow to NVIDIA, but they will hinder its ability to return to peak growth. I'm not selling my position, but I'm also not buying the dip."