SanDisk Corporation (SNDK) saw its stock surge 17% in pre-market trading on Friday after reporting fiscal second-quarter results that dramatically exceeded Wall Street expectations, fueled by explosive demand for flash storage from artificial intelligence data centers.
The data storage firm reported revenue of $3.03 billion for the quarter, significantly beating analyst estimates of $2.69 billion. Profit performance was even more staggering, with net income reaching $803 million, a 672% year-over-year increase. Earnings per share came in at $6.20, far above the $3.49 consensus estimate.
The primary driver of this growth was a 76% year-over-year increase in data-center revenue, as AI companies build specialized infrastructure for inference operations. CEO David Goeckeler told Reuters that "customers are prioritizing supply over price," highlighting a tight market where demand for flash storage is outpacing available inventory. This supply-demand imbalance is expected to persist through 2026, granting SanDisk sustained pricing power.
To secure its long-term capacity, SanDisk extended its flash chip supply agreement with Japanese partner Kioxia Corp through 2034, adding nine years to a deal previously set to expire in 2029. This locks in production capacity from joint venture facilities in Japan.
The company's forward guidance stunned investors. SanDisk forecast third-quarter revenue with a midpoint of $4.6 billion, nearly double Wall Street's $2.9 billion estimate. Adjusted earnings per share guidance of $12 to $14 also crushed the analyst consensus of $5.11.
Analyst Christopher Muse of Cantor Fitzgerald described the situation as a "perfect storm" of strong demand and constrained supply. He expects average selling prices for NAND flash to rise another 50% quarter-over-quarter, potentially pushing SanDisk's gross margin to a "staggering" 67%. Muse raised his price target for SNDK stock to $800 by the end of 2026, implying over 30% upside from current levels, citing the company's role as a "critical enabler" for hyperscale AI infrastructure.
