Original author: Li Dan
Source: Wall Street Journal
As AI inference demands continue to drive up storage needs, SanDisk has set extremely aggressive long-term financial targets.
On Thursday, the 13th, Eastern Time, SanDisk presented its long-term financial model at the 2026 Investor Day, outlining a series of financial targets to be achieved between fiscal year 2028 and fiscal year 2030, including mid- to high-double-digit revenue growth during this period. SanDisk also addressed market concerns regarding bit growth, stating that the amount of bits available for sale will be adjusted based on profitability optimization needs, while committing to return 100% of remaining cash to shareholders after completing business investments.
After the announcement, the market quickly voted with its stock prices. SanDisk (SNDK) surged as much as 17.6% during intraday trading on Thursday; the storage sector strengthened in tandem. By close, SanDisk rose nearly 14%, SK Hynix and Western Digital climbed over 7%, Seagate Technology advanced nearly 5%, and Micron Technology gained more than 4%.
FY2028-30 targets: gross margin of 80%, operating margin of 75%
The most anticipated aspect of this investor day was undoubtedly SanDisk's long-term financial model.
The company expects revenue to grow at a high single to double-digit rate during the fiscal years 2028 to 2030, in line with the growth in Bitcoin shipment volumes; meanwhile, the gross margin is expected to remain around 80%, and the operating margin around 75%, on a non-GAAP basis.
Special attention must be paid to the concept of the fiscal year. SanDisk’s fiscal year does not align with the calendar year; the company’s fiscal year ends on the Friday closest to June 30, typically consisting of 52 weeks. The previous fiscal year, FY2026, ended on July 3, 2026. Since early July 2026 already marks the beginning of FY2027, the FY2028–FY2030 period announced here does not refer to the calendar years 2028 through 2030, but rather to the upcoming three consecutive fiscal years beginning around July 2027.
Under this financial model, SanDisk expects operating expenses to account for approximately 5% of revenue, and other income and expenses will not have a material impact. Even after accounting for taxes, capital expenditures, and working capital required to support business growth, the company still anticipates an adjusted free cash flow margin of approximately 50%.
For the highly cyclical NAND storage industry, these targets are particularly aggressive. SanDisk is effectively sending a clear message to the market: AI-driven growth in storage demand is expected to keep the company’s revenue growth and profitability well above the historical averages of traditional storage cycles for years to come.
Don't blindly chase "hash rate": SanDisk will flexibly adjust salable output based on profitability.
Another important signal from SanDisk this time is that the company does not intend to simply pursue growth by increasing bitcoin shipment volumes, but will proactively adjust the amount of bitcoin available for sale based on profitability.
Previously, the market focused on SanDisk's bit growth guidance for FY2027: the company expected input bit growth to reach mid- to high-double digits, while salable output bit growth might be lower than this level. At Investor Day, management further clarified that this does not indicate a lack of capacity to increase output.
According to KC Rajkumar of Lynx Equity Strategies’ analysis of management statements, SanDisk CEO David Goeckeler clearly stated that the long-term input bit growth target is in the mid- to high-double digits, but the available bit volume for sale will be flexibly adjusted based on the need to optimize profitability; in certain periods, actual output bit growth may even exceed the mid- to high-double digits.
This means SanDisk places greater emphasis on “how much money can be earned per bit” rather than simply pursuing “how many bits can be sold.”
During the transition of NAND technology nodes, the company will also selectively reduce wafer output to avoid rapid increases in bit density from the new technology, which could lead to market oversupply.
Rajkumar noted that each NAND technology node transition at SanDisk has historically delivered an average of 54% bit growth. Therefore, if the company fully leveraged the additional capacity from these technological upgrades, it could easily recreate a supply glut. By proactively reducing wafer output during node transitions, SanDisk can control the volume of bits entering the market, thereby better preserving prices, profit margins, and capital efficiency.
This line of thinking also explains why SanDisk dares to set a long-term gross margin target of approximately 80%: improvements in bit density driven by technological advances do not necessarily translate entirely into increased supply; the company can proactively “apply the brakes” to convert part of its technological gains into profitability.
More of the "profits earned" returned to shareholders: 100% buyback or dividend of excess cash
In addition to revenue and profit margin targets, SanDisk has provided very clear commitments on capital returns.
SanDisk's Chief Financial Officer (CFO), Luis Visoso, said that after completing the investments needed to support business growth, the company expects to return 100% of excess cash to shareholders.
This means that SanDisk’s future capital allocation framework will be centered on three key pillars: first, investing in businesses and technologies that drive growth; second, maintaining strong free cash flow generation; and third, returning as much remaining cash as possible to shareholders.
An adjusted free cash flow margin of approximately 50% suggests that if the long-term financial model is realized, SanDisk will have very strong cash generation capabilities, which is one of the key reasons the market is willing to assign it a higher valuation.
Eight major clients have signed long-term agreements, covering approximately two-thirds of Bitcoin shipment volume for FY2028.
A key reason SanDisk has strong confidence in the above long-term financial model is that the company is transforming the traditional NAND industry business model.
The company has disclosed that it has signed new business model (NBM) agreements with eight clients. These agreements include committed purchase volumes, binding contractual frameworks, minimum financial safeguards, and structured pricing mechanisms that align customer demand with the company’s capacity planning and reduce the impact of cyclical fluctuations in the traditional storage industry.
More importantly, the scale covered by these agreements is already substantial: currently, the signed NBM agreements cover approximately 50% of Bit's shipment volume for FY2027 and about two-thirds of Bit's shipment volume for FY2028.
SanDisk believes this model can lead to more predictable revenue, higher cash flow visibility, and more sustainable profit growth.
In other words, SanDisk is not only betting on the growing storage demand driven by AI, but also attempting to partially transform the cyclical nature of its traditional NAND business into more stable and predictable revenue and cash flow through long-term agreements.
AI reasoning is driving a larger storage market, with the enterprise SSD TAM projected to reach 1.2 ZB by 2030.
Another major driver of SanDisk's high growth over the next three years is the new demand for storage infrastructure as AI expands from training to inference.
SanDisk estimates that the flash memory market could grow from a historical annual revenue base of $60 billion to over $300 billion by 2026 and approach $500 billion in 2027.
The company stated that AI inference workloads are driving rapid growth in token usage, and KV Cache is reshaping the memory hierarchy in data centers. As AI inference scales, AI data centers will become increasingly reliant on storage; SanDisk expects the total available market (TAM) for enterprise data center flash to reach 1.2 zettabytes (ZB) by 2030.
On the technical level, SanDisk is advancing a two-dimensional scaling strategy based on CMOS Bonded Array (CBA) to more flexibly develop customized products that meet diverse market demands while improving capital efficiency.
The company's latest BiCS9 QLC technology is the first example of this strategy. This technology combines BiCS8 arrays with CMOS wafers based on BiCS10; meanwhile, the new BiCS10 QLC node achieves a 60% increase in bit density compared to BiCS8.
HBF bets on AI inference, leading to a collective surge in the storage sector
SanDisk is advancing new high-bandwidth flash (HBF) technology for AI inference. The company states that HBF is emerging as a critical storage solution for the AI inference era, and the related industry ecosystem is currently taking shape.
To the market, this means SanDisk's AI storage logic has moved beyond the notion that "AI data centers need more SSDs" and has further expanded to encompass the AI inference architecture's demand for higher performance, lower power consumption, and greater storage density.
This expectation also quickly spread to the entire storage sector. SanDisk's own rally was even more pronounced: after announcing its long-term financial targets, its stock surged nearly 18% during trading. Year to date, SanDisk's stock has climbed more than 530%.
However, SanDisk also emphasized that the aforementioned long-term financial targets are forward-looking statements based on a series of estimates and assumptions, and actual results may still be affected by factors such as demand, average selling prices, competition, technological advancements, supply chain, and the storage industry cycle.