Does the Sandisk earnings report create a buying opportunity?

👆Click on the blue text to follow us

Sandisk Financial Report

Now, let's take a look at the Sandisk financial report. Sandisk delivered a mixed report. The bad news is that all numbers this quarter exceeded expectations, but not by much, and the worse news is that next quarter's guidance is fairly flat, near the lower end of market expectations, and the growth rate has also started to slow down. However, the good news is that this time management disclosed detailed long-term contract information, further strengthening the consensus that the cycle is lengthening. So, let's skip the financial data and go straight to the most important part—the long-term contracts.

Sandisk announced its first five long-term contracts in April, and this quarter added five more, totaling ten, covering eight data center and edge computing enterprises. Of the new contracts, three are new customers with new agreements, and two are expansions with existing customers. This means new clients are willing to join and old clients are expanding their orders, at least indicating the long-term contract model is moving from trial to scale, even normalization.

As for terms and pricing, the contracts are as long as five years, and the average duration of the ten contracts is more than four years, meaning very few clients sign for just three years. Moreover, supply and demand commitments are further split by year and quarter, so customers can't just give a five-year total procurement target and then defer orders at will. They must purchase according to the agreed amount every quarter and year. This helps Sandisk better plan capacity and inventory, and makes revenue clearer. For customers, while they gain certainty in NAND supply for several years, they lose flexibility.

In terms of pricing, there are two mechanisms: fixed price and floating price. The floating price cap was not disclosed; the lower limit was described as very attractive but with no reference gross margin provided.

However, based on the lower limit, the minimum contractual revenue for the existing long-term contracts is $93.9 billion, about 4.6 times Sandisk's fiscal 2026 revenue. If clients and Sandisk both honor their commitments, the average revenue over the next five years should not be lower than this year. Based on this quarter's net profit margin and an average semiconductor valuation of 15x, just the long-term contracts alone can basically support the current share price. However, it's foreseeable that as long-term contracts gradually penetrate, price flexibility and gross margin may be affected.

In terms of supply quantity, the company said that in fiscal 2027, which is from this July to the end of June next year, about half of NAND has been locked in by long-term contracts. By fiscal 2028, this proportion will rise to about two-thirds.

The benefit is that most of the products have already been locked in, but the downside is that free capacity for distribution is declining. The company is trading some upside flexibility for revenue visibility.

As for terms, to prevent customers from defaulting, Sandisk obtained $16.5 billion in financial guarantees, equivalent to 18% of the minimum contract revenue. These guarantees are not released as procurement payments at the start but are gradually released as the agreements near completion. Thus, as time passes and performance obligations decrease, the guarantee coverage rate for the remaining contracts actually increases. This is very similar to Micron's terms.

Finally, to ensure quarterly delivery, the company needs to maintain higher inventory days and bear higher component costs. Management expects higher inventory levels will suppress full-year saleable NAND growth to the mid-to-high double digits. Additional capacity will be needed in the future to match demand.

Jason believes that, aside from the lack of explicit take-or-pay clauses, the rest is very similar to Micron and offers fairly comprehensive contractual protection. However, when the market saw growth slowing, gross margins peaking, and weakened price flexibility, it chose to exit, essentially still pricing with the old short-cycle mindset.

Although storage can't escape cycles, it's true that cycle risk has decreased compared to before. Now, with concrete terms and long-term contracts on the table, as long as key demand—mainly AI data center demand—does not slow down, and as long as there is no explosive commercial breakthrough in compression technology during this cycle, the upward trend in the storage cycle will be extended.

Therefore, I believe the future is more likely to follow a balanced path, rather than an overwhelmingly short-cycle mindset. We don't have to be as pessimistic as the market is today, but it's also best not to expect another wave of vertical surges in storage prices.