5.7.25
This was SanDisk’s first quarter as a standalone company, following their split from Western Digital. This separation reverses a merger done ten years ago. No investment banker left behind. Management believes demand will strengthen throughout the year. Revenue declined 10% in the quarter, to $1.695B from $1.876B. Gross margin dropped 980 basis points, to 22.7%. This is an acceleration from the 640 bps decline last quarter. For the prior seven quarters, oldest to newest, flash gross margins were (10.3%), 7.9%, 27.4%, 36.5%, 38.9%, 32.5%, and 22.7%. This compression in gross margin collapsed operating margin down to almost nothing. Earnings went negative in the quarter. The upturn for NAND lasted four quarters and now we are two quarters into the downturn. However, guidance for the current quarter, their fiscal fourth quarter, are for revenue to rise some, to $1.8B and for gross margin to expand to 25.9%. The quarter just finished is the seasonally weakest, so this forecasted recovery may be that seasonal strengthening. It may also be that the NAND industry has reached healthy inventory levels and demand has returned. It was an historically short upturn, so maybe it will be followed by an historically short downturn? I think that is unlikely, but we are in strange times.
Within the three end markets that SanDisk divided their business in to, cloud is the smallest. It also saw the largest drop in revenue: 21%. Client was down 10% and consumer declined 5%. There has been softening in demand for server SSDs recently, which is likely the reason for the cloud decline. Client and consumer revenue coming down is a combination of seasonal weakness in demand and ASP declines. Segment commentary included the prediction that AI-enabled PCs and mobile devices, as well as the Windows 10 end of life replacement cycle, will spur demand. Poppycock. In the Datacenter Segment, the company quoted hyperscaler cloud capex of $240B in 2024, rising 40% to $330B in 2025.
When the businesses were combined, WD would split out blended and like-for-like pricing. SanDisk doesn’t do this, which is a demerit for their IR department. They are reducing the amount of information available to investors. From the time trend of blended pricing, NAND ASPs bottomed in the third calendar quarter of 2023. After that, they recovered for three quarters (like-for-like rose for four quarters), reaching a peak that was ~50% above the nadir of the prior downturn. That top was the second calendar quarter of last year. Since then, NAND ASPs are down ~25%, having dropped for three quarters in a row. Because of the elasticity in the NAND market, bit shipment changes are partially decoupled from pricing changes. Following two up quarters, bit shipments were about flat this quarter. This period is the seasonally weakest of the year. For the last six quarters – the period since the upturn began in NAND – SanDisk’s bit shipments are down, in total, 5%. Said another way, the company is shipping 5% fewer NAND bits now than they were eighteen months ago.
This section is commentary from management’s conference call with analyst.
- · Down pricing was the result of “continued oversupply in the market.” This was above the prediction they gave at their analyst day. In response, they are extending their fab underutilization actions until they see sustained price recovery. In this same statement, they said they just started implementing price increases to their customers. This is a quick response in pricing to supply actions.
- · Their guidance for the current quarter assumes current tariffs remain unchanged throughout the quarter. Presently, there are not tariffs on their products. 20% of their products are shipped to the U.S., and 95% of their revenue is sourced from countries other than China. This is excellent detail.
- · Their fourth quarter guidance assumes bit shipments will be flat and ASPs will be up mid-to-high single digits.
- · The fourth quarter will be burdened with $55M to $65M of underutilization charges. That is on total COGS in the range around $1.3B.
- · In fiscal 2026, which begins in July of 2025, the company plans to match industry bit growth. They expect to see the industry balanced between supply and demand, resulting in an increase in average selling price. Executives don’t know much beyond one quarter out. The beginning of SanDisk’s fiscal 2026 is one quarter out.
- · The CEO says they see an undersupplied market through the end of next year. I think he means fiscal year here, which is a forecast out to the end of June of 2026. That is some wishful thinking. He characterized the dropping prices of the last two quarters as a mid-cycle pause, which he believes they are coming out of now. Watch closely to see if this prediction comes to pass. It would be an historic prognostication if it does. They believe the NAND market will continue to strengthen throughout the year.
- · They see supply in calendar 2025 growing mid-single digits and demand growing “kind of” low double digits.
- · Their CFO referred to the transition from the Big Five in NAND to the Big Eight. This is the first time I’ve heard this terminology. I believe it refers to the addition of YMTC to the market. I don’t know what the other two are. SanDisk’s problem of matching their supply to the overall market is significantly harder with YMTC in the mix. It was not really manageable with five in the market. Six just makes it that much worse.
- · Company bit shipments will grow high-single digits in the second half of calendar 2025, following a flat second calendar quarter.
- · They are seeing a “turn in the market” that is “across the board” to higher pricing. They see an undersupplied market through the end of calendar year 2026. I have never heard a memory industry executive make a prediction out that far. I believe that is delusional. Check back on this as the next eighteen months unfold.
In summary, it sure looks to me like the NAND market is in a downturn. SanDisk’s executives disagree with me. They are calling the pricing declines of the last three quarters a “mid-cycle pause.” They have gone as far as forecasting ASPs in the current quarter. This is something I can’t remember a memory company ever doing. Maybe their CEO had a minor stroke and nobody noticed. Or maybe he’s just that confident. Ninety days ago, this management team did predict the pricing headwinds they were seeing at the time (FQ2) would be less in FQ3. That prediction was true, by a narrow margin. SanDisk’s ASPs were down 10% in FQ2 and down 9% in FQ3, a small decrease. The company is underloading their fabs today in response to weak pricing. So this is what it comes down to. Is the recent softness a “mid-cycle pause,” as the SanDisk executives believe? Or is it just a temporary blip in the downward trend in pricing? The latter is my belief. If prices start dropping again in SanDisk’s FQ1-26, then I’m likely right. If they go up in FQ4-25 and again in FQ1-26, then I’m wrong and the SanDisk executives are right. I think the determining factor will be YMTC’s supply growth. Inventory levels across the industry appear to have been burned down to a healthy state. Now that we are there, is the underlying demand enough to support higher pricing beyond the next couple months? If YMTC continues to raise their output in 2025 the way they did in 2024, I think the NAND industry will return to an oversupplied state in the second half of calendar 2025.
– Smooth Hughes (short MU)