SanDisk's BiCS10 and $42B in LTAs Signal Progress, but Valuation Remains Stretched

What happened

SanDisk's BiCS10 node delivers 59% higher bit density and is already in production, reducing execution risk. Data center revenue surged 230% sequentially, driven by AI inference, KV cache, and enterprise SSD demand. Five multi-year agreements secure approximately $42 billion of minimum revenue with over $11 billion of financial guarantees. However, the DeepValue report notes that the stock is priced for peak NAND shortage economics, with earnings driven by a 248% ASP/GB increase rather than volume growth. Competitors like Micron and Samsung are signing similar long-term agreements, limiting Sandisk's differentiation and suggesting the market may be overpaying for the 'de-cycling' narrative.

Implication

For position holders, the next key test is Q4 FY26 results: if revenue meets $7.75B-$8.25B and gross margin stays above 79%, the thesis of durable shortage economics remains intact. However, any miss would suggest the Q3 peak is behind us. New investors should wait for a better entry near $1,250, where the risk/reward improves. The LTAs provide visibility but are not unique, and the market's enthusiasm may fade as peers demonstrate similar contract coverage. Ultimately, Sandisk's ability to defend margins through a pricing downturn remains unproven, making the stock a hold rather than a buy at current levels.

Thesis delta

The news reinforces near-term demand strength and contract visibility, but it does not change the fundamental challenge: the current valuation is based on price-led earnings that may not be sustainable. The thesis shifts from 'early-cycle AI storage winner' to 'peak-cycle pricing story with limited margin of safety.' Investors should demand evidence that LTAs protect margins before treating Sandisk as a de-cycled growth compounder.