Sandisk’s new annual report captures one of the sharpest reversals in the semiconductor-memory cycle: fiscal 2026 revenue nearly tripled, gross margin surged above 70%, operating cash flow exploded and the company returned billions of dollars to shareholders. The same filing also explains why investors should not mistake a cyclical peak for a permanently de-risked earnings model. Sandisk has converted a market recovery into exceptional profitability, but it remains deeply exposed to pricing, supply commitments and execution with its manufacturing partner.
The company’s fiscal 2026 10-K, filed August 17 for the year ended July 3, reported revenue of $20.25 billion, up 175% from $7.36 billion in fiscal 2025. Gross profit rose to $14.47 billion from $2.21 billion, taking gross margin to 71.5% from 30.1%. GAAP operating income was $12.39 billion, reversing a $1.38 billion operating loss a year earlier, while GAAP net income reached $11.43 billion versus a $1.64 billion net loss.
The quality of the revenue change matters. Datacenter revenue climbed from $960 million to $5.15 billion, Edge revenue rose from $4.13 billion to $12.16 billion and Consumer revenue increased from $2.27 billion to $2.94 billion. The filing attributes the gains to a combination of higher sales and sharply improved pricing in Datacenter and Edge. This is the bull case in concentrated form: flash storage is no longer merely a low-margin component business when enterprise and AI-related data demand collide with constrained supply.
Cash flow validates that the earnings improvement was not solely an accounting event. Sandisk generated $11.67 billion of operating cash flow in fiscal 2026, compared with just $84 million the prior year. It used $4.50 billion for share repurchases and $1.90 billion to repay and settle its term loan. That balance-sheet strengthening gives management flexibility to invest in new process nodes and return capital at the same time.
Yet the annual report contains the warning label. Sandisk procures all of its flash-memory wafers from three joint ventures with Kioxia, in which Sandisk owns 49.9%. It is generally entitled to half the output, but it must fund half of fixed costs even if it chooses to buy fewer wafers. The ventures have been extended through December 31, 2034, and Sandisk has committed to pay Kioxia $1.2 billion between 2026 and 2029 for manufacturing services and continued supply availability.
That structure has genuine strategic benefits. It secures access to leading-edge wafers, shares development cost and gives Sandisk more manufacturing control than a pure merchant buyer would have. But it also limits flexibility in a downturn. Memory pricing can change rapidly; fixed-cost and supply commitments become most visible precisely when demand weakens and inventory needs to be managed carefully.
There is a second reason to temper extrapolation. Fiscal 2026 margins were powered by higher pricing as well as volume. The company reported Datacenter revenue per gigabyte up almost 150% and Edge revenue per gigabyte up almost 180%. Such changes create extraordinary operating leverage on the way up, but the same mechanism works in reverse if pricing normalizes. Sandisk itself expects increased fiscal 2027 capital investment as it transitions to newer technology nodes, a sensible strategic move that will require sustained cash discipline.
The market has already recognized much of the turnaround: SNDK closed near $1,786.85 on August 17 after a powerful rally. The next test is whether management can convert cycle-driven profitability into a more durable model through its long-term customer frameworks, data-center mix and supply planning. The 10-K supports confidence in the company’s execution; it does not remove the memory cycle.
Verdict on Sandisk (SNDK): HOLD. The company’s fiscal 2026 transformation and datacenter exposure are compelling, but current pricing leaves little room for a sharp memory-market reversal or margin normalization.
Price Target: $2,000.00
Disclaimer: The information provided in this article is for informational and educational purposes only and should not be construed as financial advice. Readers should do their own due diligence before making any investment decisions. Past performance does not equal future results.
