SanDisk has set an unusually aggressive financial target for fiscal 2028 through fiscal 2030, projecting mid-to-high double-digit annual revenue growth, roughly 80% non-GAAP gross margins and about 75% operating margins as it positions flash storage for an expansion in AI inference workloads.
The outlook, presented at the company’s 2026 investor day, helped drive a sharp move in SanDisk shares. SNDK rose as much as 17.6% during Thursday’s session before closing nearly 14% higher. The rally spread across storage manufacturers, with SK Hynix and Western Digital each gaining more than 7%, Seagate Technology rising nearly 5%, and Micron Technology adding more than 4%.
SanDisk’s presentation puts the storage industry’s supply discipline at the center of its long-range model. Rather than pursuing the highest possible volume of NAND flash bits, the company said it will adjust saleable output according to profitability, an approach that could restrain supply during periods when newer technology sharply increases storage density.
The forecasts are forward-looking and depend on demand, average selling prices, competition, technology shifts, supply-chain conditions and the storage sector’s normal cyclicality, SanDisk said.
Cash returns follow spending on technology and growth
SanDisk said it intends to return 100% of cash remaining after business investments to shareholders. Its framework prioritizes investment in technology and growth, followed by sustaining free-cash-flow generation and then distributing surplus cash.
For fiscal 2028 through fiscal 2030, the company projected adjusted free cash flow equal to about 50% of revenue after taxes, capital expenditures and working-capital requirements. It also expects operating expenses to run at around 5% of revenue, while other income and expenses are not expected to materially affect the model.
The targets cover a period beginning around July 2027. SanDisk’s fiscal year ends on the Friday nearest to June 30, rather than on the calendar-year schedule used by many public companies. Fiscal 2026 ended on July 3, 2026, according to the company.
SanDisk said revenue growth during the 2028-2030 period should broadly match the growth of bit shipments. In NAND flash, bit shipments measure the total storage capacity sold, making them a central indicator of demand and manufacturing output.
Profitability takes priority over maximum NAND output
SanDisk’s emphasis on managing saleable bits addresses a major feature of NAND manufacturing: each transition to a more advanced production node can add large quantities of storage capacity even if wafer output remains unchanged.
KC Rajkumar, an analyst at Lynx Equity Strategies, said SanDisk Chief Executive David Goeckeler indicated that the company’s long-term input-bit growth target remains in the mid-to-high double digits. Saleable output will be adjusted to optimize profitability, Rajkumar said, although output growth could surpass that range in certain periods.
During NAND node transitions, SanDisk may reduce wafer production to avoid flooding the market with newly available capacity, the company said. Rajkumar said SanDisk estimates an average node change produces roughly 54% bit growth, a gain large enough to rapidly expand supply if every available bit were released for sale.
That discipline is central to the company’s margin assumptions. The flash-memory business has historically been vulnerable to periods of oversupply, when rapid capacity increases can pressure selling prices. SanDisk is signaling that it would accept lower unit growth if necessary to protect the economics of its output.
Customer contracts cover much of the planned volume
SanDisk also disclosed new business model agreements with eight customers. The contracts include committed volumes, minimum financial protections and structured pricing mechanisms, according to the company.
Those agreements cover about half of SanDisk’s expected fiscal 2027 bit shipments and approximately two-thirds of projected fiscal 2028 shipments. The expanding share of contracted volume gives SanDisk greater visibility than spot-market sales alone and could make its production planning less dependent on short-term swings in memory prices.
The company’s market projections are equally expansive. SanDisk estimates the flash-storage market could rise from a historical annual revenue base of about $60 billion to more than $300 billion in 2026 and nearly $500 billion in 2027. By 2030, it expects the enterprise data-center flash market to reach 1.2 zettabytes of addressable demand.
These figures are SanDisk’s own estimates, and they reflect its expectation that AI systems will require far more data to remain readily accessible during inference, the stage in which trained AI models generate answers, classifications or other outputs.
AI inference shapes the technology roadmap
SanDisk is developing high-bandwidth flash, or HBF, for AI inference applications. The product category is aimed at workloads that require storage to feed large volumes of data to computing systems efficiently, potentially creating a role for flash closer to performance-sensitive AI infrastructure.
The company is also pursuing what it calls a two-dimensional scaling strategy built around CMOS direct-bonded array technology, or CBA. SanDisk said the approach would support more customized products and improve capital efficiency.
Its BiCS9 QLC design combines a BiCS8 memory array with a BiCS10-based CMOS wafer, according to SanDisk. QLC, or quad-level cell flash, stores four bits of data per cell, allowing more capacity per chip than lower-bit-count designs. SanDisk said its BiCS10 QLC node provides a 60% increase in bit density compared with BiCS8.
The storage rally may draw attention from cryptocurrency traders because AI data centers, cloud providers and blockchain infrastructure can all compete for semiconductors, servers and power. SanDisk’s presentation offers a narrower conclusion than claims that storage constraints will directly limit blockchain validation or force a broad repricing of digital assets.
NAND flash is primarily used for data storage, while the computing hardware used for proof-of-work mining or transaction validation depends on the network: Bitcoin mining relies on specialized ASIC machines, while many other blockchain systems do not depend on graphics-card mining at all. Rising storage demand can raise infrastructure costs for data-heavy applications, but the effect varies substantially by network design and operator.
Explore how institutional investors navigate AI-driven opportunities in crypto—start with our comprehensive cryptocurrency guide to understand this evolving market.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
