Sandisk can produce contradictory valuation signals at the same time.
Near-term earnings can make the expectations look modest.
It looks expensive against almost any memory cycle investors have seen before.
The difference is management's claim that this is not the old cycle.
At $1,772.10, buyers are paying for that claim to become durable cash flow.
What Sandisk actually controls
Sandisk supplies NAND flash storage. In the AI stack, that means the persistent storage surrounding training, inference, checkpoints, and enormous data sets.
GPUs process the data. They do not make the need to store it disappear.
In fiscal Q4 2026, revenue reached $8.97 billion, up 51% sequentially. Fiscal 2026 revenue reached $20.25 billion, up 175%. Data-center revenue increased 437%.
GAAP net income was $6.90 billion in Q4 and $11.43 billion for the year.
Why does that matter?
Because Sandisk is not waiting for the AI storage thesis. It is already monetizing it.
The contracts are the real investment thesis
Sandisk has described long-term New Business Model agreements that cover about half of expected fiscal 2027 bits and roughly two-thirds of fiscal 2028 bits. At Investor Day, the company disclosed tens of billions of dollars of remaining commitments at floor pricing and a weighted-average term exceeding four years.
Management's fiscal 2028 through fiscal 2030 framework targets mid-to-high-teens revenue growth, roughly 75% non-GAAP operating margin, and about 50% adjusted free-cash-flow margin.
Those are extraordinary targets for NAND.
The agreements may improve visibility and reduce the industry's habit of destroying pricing through undisciplined supply. They do not repeal technology transitions, customer bargaining power, or the possibility of weaker demand.
What $1,772.10 requires
I use fiscal 2030 because it is the end of management's disclosed target window. At a 12% required annual return and a 15 times terminal earnings multiple, today's price requires approximately $186 of fiscal 2030 EPS.
Management did not guide to $186 of EPS.
My translation of its mid-to-high-teens revenue-growth and 75% operating-margin framework produces about $154 of fiscal 2030 EPS before assuming future buybacks.
The current stock price therefore requires more than the midpoint framework, a lower required return, meaningful share reduction, or a higher terminal multiple.
Three possible outcomes
Conservative: about $818 today
Sandisk earns $107 per share in fiscal 2030 and receives a 12 times terminal multiple.
This assumes healthy AI storage demand but meaningful cycle normalization and a continuing memory discount.
Management-framework case: about $1,469 today
Sandisk earns $154 per share in fiscal 2030 and receives a 15 times terminal multiple.
The earnings estimate uses 17% annual revenue growth from fiscal 2026, a 75% operating margin, a 15% analytical tax rate, and roughly 157 million diluted shares.
Upside case: about $1,886 today
Sandisk earns $165 per share in fiscal 2030 and receives an 18 times terminal multiple.
This requires high-teens growth, durable margins, strong contract execution, and a structural re-rating.
At $1,772.10, SNDK trades about 21% above my management-framework case and only 6% below the upside case.
The stock is not waiting for the new model to be proven.
Capital allocation can help, but it cannot rescue weak economics
Sandisk ended fiscal 2026 with roughly $4.76 billion of cash, $1.78 billion of marketable equity securities, and no debt.
That is a major strength.
Management also has room to repurchase shares. Buybacks would improve per-share value if they retire stock below intrinsic value.
But buying back an expensive stock cannot substitute for delivering the margin and cash-flow framework. The first job is to make the contracts produce cash. The second is to allocate that cash well.
What to watch out for
- Quarterly data-center revenue, pricing, and gross margin - because today's economics are exceptional.
- Contracted bits, floor pricing, and remaining obligations - because the durability thesis lives inside those agreements.
- Adjusted free cash flow after capital needs - because margin targets matter only if cash reaches shareholders.
- Kioxia manufacturing execution and technology transitions - because supply economics depend on the production partnership.
- Diluted shares and repurchases - because buybacks are part of the upside case.
What could change the thesis
The thesis would strengthen if the long-term agreements protect pricing through weaker industry conditions, data-center demand stays broad, and buybacks materially reduce diluted shares.
The thesis would weaken if margins fall far below the long-term model, contract economics prove less protective than advertised, or customer commitments become concentrated in a few deployments.
The decisive test is not another spectacular quarter. It is what happens when NAND supply and demand stop being perfectly favorable.
Bottom line
Sandisk's operating transformation is real.
Business quality: strengthening rapidly.
Balance sheet: excellent.
Contract visibility: improved, but not cycle-tested.
Valuation: above the management-framework case and close to the upside case.
At $1,772.10, the stock needs approximately $186 of fiscal 2030 EPS at a 15 times multiple to deliver a 12% annual return. That is more than my direct translation of management's own framework.
The practical principle: when management offers a long-term model, translate it into per-share value before paying a premium to it.
Sources
- Sandisk fiscal Q4 2026 results
- Sandisk fiscal 2026 Form 10-K
- Sandisk Investor Day financial-model release
- Federal Reserve H.15 interest rates
Scenario values are analytical estimates, not company guidance or price targets. Educational only - not financial advice.