Nasdaq Composite Slides as AI Infrastructure Sells Down To Start H2

The Nasdaq Composite fell 0.66% to close at 26,040.03 today, the first trading session of the second half of 2026, as a sharp selloff in semiconductor and AI infrastructure stocks overwhelmed gains in previously lagging software and platform names, signalling a meaningful shift in leadership within the technology complex.

The Nasdaq 100 bore the brunt of the decline, dropping 1.54% as profit-taking accelerated across the index’s highest-flying constituents from the past year. The divergence between winners and losers revealed a market in the midst of an internal rotation, with money flowing out of crowded AI hardware positions and into underperformers that had suffered during the first-half rally.

Today’s Movers on the Nasdaq: AI Infrastructure Takes Heavy Losses

The day’s damage was concentrated in semiconductor capital equipment, memory, and AI compute infrastructure stocks.

Micron Technology plunged 10.57% to $1,032.28 after a meteoric 808% gain over the trailing twelve months. Sandisk tumbled 10.62% to $2,032.22 despite posting an extraordinary 4,334% one-year return. Applied Materials dropped 9.97% to $650.91, Lam Research fell 9.71% to $391.26, and KLA declined 11.77% to $266.19.

The rout extended to recently added Nasdaq 100 constituents.

CoreWeave, which joined the index in the June 22 rebalance, crashed 13.92% to $85.69. Nebius, another recent addition, plummeted 17.01% to $229.18. Astera Labs, the third AI infrastructure name added to the index last month, dropped 10.80% to $430.86. Teradyne fell 11.68% to $427.34, rounding out a brutal session for the AI hardware ecosystem.

Traditional semiconductor leaders also suffered.

AMD declined 6.89% to $540.88, Intel dropped 9.03% to $127.02, and ASML fell 7.36% to $1,843.04. Marvell Technology lost 8.67% to $272.05, while Western Digital and Lumentum Holdings each shed more than 6%.

Beaten-Down Names Stage Sharp Reversals

Against this backdrop, a cohort of heavily sold technology names mounted impressive rallies.

AppLovin surged 9.58% to $564.61, recovering some ground after a 31.73% decline over the prior six months. Meta Platforms jumped 8.81% to $612.91, driven by a strategic announcement that sent shockwaves through the AI infrastructure space. PDD Holdings climbed 8.18% to $82.52, Palantir Technologies rose 7.77% to $125.73, and Strategy gained 7.43% to $93.39.

Software and enterprise technology names also participated in the rotation.

Shopify advanced 6.52% to $121.63, while Workday increased 6.41% to $130.27. The moves came despite mixed longer-term performance for many of these names, suggesting markets were responding to positioning dynamics and valuation resets rather than fundamental catalysts.

Meta’s Cloud Ambitions

The session’s most significant development came from Meta Platforms, which reportedly plans to launch a comprehensive AI cloud offering dubbed “Meta Compute.” Rhe initiative will provide external access to Meta’s vast AI infrastructure through three channels: hosted model access similar to Amazon Web Services’ Bedrock platform, raw GPU capacity competing directly with independent providers, and direct access to Meta’s proprietary data centers and AI models.

The announcement positioned Meta as a direct competitor to standalone GPU cloud providers that have been among the market’s hottest performers. The strategic pivot reflects Meta’s desire to monetize underutilized compute capacity built for internal AI development, transforming a cost center into a potential revenue stream while leveraging the company’s massive capital investments in data center infrastructure.

The competitive threat to pure-play AI infrastructure providers was immediate and severe. CoreWeave and Nebius, both of which offer GPU-as-a-service to AI developers and enterprises, saw their stocks collapse as markets reassessed their competitive positioning and unit economics. The selloff extended a multi-day decline that had already raised questions about profitability timelines and capital intensity for the neo-cloud sector.

Profit-Taking Follows Record First Half

Before today’s pullback, memory and storage stocks had delivered triple-digit percentage returns in many cases, with AI-linked names benefiting from surging data center demand and capacity expansion by hyperscale cloud providers.

The June Nasdaq 100 rebalance had further concentrated AI hardware exposure by adding Astera Labs, CoreWeave, Nebius, and Teradyne to the index. The moves triggered substantial passive inflows but also created crowded positioning that became vulnerable to profit-taking. With many AI infrastructure names trading at elevated valuations despite negative cash flows and heavy capital requirements, markets appeared to be shifting from a “growth at any price” mentality to greater scrutiny of unit economics and paths to profitability.

The session’s price action marked a potential inflection point in how markets approach AI-related investments. The first half of 2026 rewarded pure-play exposure to AI infrastructure buildout, with semiconductor equipment, memory, and GPU cloud providers delivering outsized returns. The rotation into software, platforms, and previously lagging names suggested that markets were beginning to differentiate more carefully within the AI ecosystem.

For bulls, the key argument remains that AI represents a multi-year secular growth story with substantial runway ahead. Hyperscaler capital expenditure continues to accelerate, enterprise AI adoption is in early innings, and the infrastructure layer requires ongoing investment to support model scaling and inference deployment. From this perspective, Tuesday’s decline represents a healthy consolidation and position reset rather than the end of the AI investment cycle.

Bears counter that valuation had become disconnected from near-term fundamentals for many AI infrastructure names, particularly those with negative cash flows and uncertain paths to profitability. Meta’s entry into the GPU cloud market validates the opportunity but also introduces a formidable competitor with deeper pockets and existing customer relationships.

For now, markets appear to be navigating a transition from indiscriminate buying of AI exposure to more selective positioning based on competitive dynamics, unit economics, and valuation. The rotation into beaten-down software and platform names suggests that capital remains committed to the technology sector, but investors are becoming more discerning about where to deploy it. Whether this proves to be a temporary speed bump or a more significant shift in the AI trade’s trajectory will likely become clearer as earnings season unfolds and companies provide updated views on the pace and profitability of AI infrastructure buildout.