AI Chip Stocks Stumble Following Record $2 Trillion Second-Quarter Surge

The meteoric rise of semiconductor stocks that defined the first half of 2026 is facing a rigorous reality test as the third quarter begins. While memory makers Micron and Intel led a second-quarter rally that added $2 trillion in combined market value to the sector—with Micron soaring over 240%—the tide turned sharply in early July. The Philadelphia Stock Exchange Semiconductor Index, which recently notched its best quarter on record with an 88% advance, suffered its worst two-day selloff in months, falling as much as 6% as investors grow wary of the sustainability of current spending levels.

Market analysts, including Jim Cramer and Bloomberg's Mike Shepard, suggest the 'AI trade' is undergoing a fundamental shift. While the second quarter rewarded a broad swath of chipmakers including AMD, Marvell Technology, and Sandisk, the third quarter has opened with a 'dud' as Micron wiped out nearly $200 billion in market capitalization in a single session. Critics like Radio Free Mobile founder Richard Windsor argue that the market may have mistaken cyclical demand for permanent compute growth, signaling that the AI rally has ticked many of the boxes defining a classic financial bubble.

Despite the volatility, some industry heavyweights continue to demonstrate immense fundamental strength. Samsung Electronics projected a 19-fold surge in second-quarter operating profit, driven by robust demand for high-bandwidth memory chips. However, even this record-breaking guidance failed to insulate Asian equities from a broader pullback, as the Magnificent 7 stocks collectively shed $2.3 trillion in value amid jitters over the long-term return on AI infrastructure investment.

The divergence between cooling stock prices and massive corporate commitments remains a key theme. Even as valuations contract, firms like SK Hynix are planning U.S. ADR debuts to fund expansion, and companies are increasingly focusing on managing 'AI token spend'—using techniques from the cloud computing era to control the escalating costs of running AI agents. For now, the market appears caught between the spectacular earnings of producers and a growing fear that the infrastructure build-out may peak sooner than anticipated.