Nvidia falls 2.2% as AI investors rotate into memory chip stocks

Alphabet's higher AI spending failed to lift Nvidia, as money rotated into memory makers Micron and SK Hynix instead.

Nvidia Corp. shares fell 2.18% on Thursday, snapping a three-day winning streak, even after Alphabet Inc. raised its 2026 capital expenditure guidance for AI infrastructure.

"The market is differentiating within the AI trade," said Stacy Rasgon, senior semiconductor analyst at Bernstein. "Investors are looking past the mega-cap names toward companies that directly benefit from memory demand."

Micron Technology Inc. rose 2.5% and SK Hynix Inc. gained 2.8%, extending a week of outperformance for memory specialists. The rotation accelerated after Alphabet's updated capex outlook, which investors interpreted as a demand signal for high-bandwidth memory and data center components rather than GPU purchases alone.

The divergence suggests a broadening of the AI trade into "picks-and-shovels" semiconductor plays. Nvidia, trading at roughly 35 times forward earnings, faces headwinds from a market increasingly skeptical that higher AI spending will translate proportionally into GPU revenue growth.

Why Memory, Not GPUs

The sell-off reversed three consecutive days of gains for Nvidia. The decline came despite what would typically be a bullish catalyst: Alphabet, a Magnificent Seven member, signaling it will spend more on AI infrastructure in 2026 than previously planned.

The rotation into memory stocks has been building for weeks. Morgan Stanley analysts argued on July 20 that the recent sell-off in memory stocks "has created a strong entry point," estimating memory prices are up 25% this quarter. While that marks a slowdown from prior quarters, the bank called the deceleration necessary to prevent demand destruction.

Micron's second-quarter results underscore the demand surge. The company posted earnings growth of 1,350% on revenue that surged 345.7% year over year, according to Zacks data. SK Hynix, the world's largest supplier of high-bandwidth memory, has also benefited from the AI buildout, though its Nasdaq-listed shares have experienced volatility since debuting on the exchange.

Competitive Pressures Mount

The broader semiconductor sector has seen heightened volatility. The PHLX Semiconductor Index fell 10% the week ending July 17, as concerns about Chinese AI competition and infrastructure spending rattled investors. Chinese startup Moonshot AI's release of an open-source model rivaling leading U.S. models revived worries about tech giants' ability to recoup AI investments.

For Nvidia, the challenge is twofold. The company faces growing competition from Advanced Micro Devices Inc., which recently added Microsoft Corp. as a customer for its Helios rack-scale system, and from cloud providers developing in-house chips. At the same time, the market is beginning to question whether Nvidia's dominant position in AI training will translate into proportional gains as the focus shifts to inference workloads, where specialized memory and lower-cost alternatives may play a larger role.

The rotation comes during a pivotal week for Big Tech earnings. Alphabet and Tesla Inc. both reported after the bell Wednesday, with the Magnificent Seven group expected to post aggregate earnings growth of 28.7% in the second quarter, according to Zacks data. While that marks a deceleration from 48.7% in the first quarter, it still outpaces the broader S&P 500's expected growth of 25.3%.

Nvidia's own results, due Aug. 20, will be the next major test for the AI trade. The company's data center revenue has more than tripled year over year in each of the past three quarters, but expectations have risen in tandem. Any sign of deceleration could accelerate the rotation out of mega-cap AI names and into the semiconductor supply chain.

This article is for informational purposes only and does not constitute investment advice.