Datacenters 2027: The Year of Acceleration, +50GW in a Single Year 60GW Under Construction, SemiAnalysis Leading Indicators, 1.3T of Lease Contracts Signed YTD, The Shortage Could Move to Chips

$SNDK Datacenter supplied 88% of Sandisk’s sequential revenue increase in fiscal Q1 2027. That makes the durability of AI storage spending central to the next stage of its growth. 1. Where the growth came from Revenue rose 23% sequentially to $11.031B. Datacenter contributed $4.802B, up 61%, lifting its revenue share from 33.2% to 43.5%. I calculate that datacenter supplied $1.825B of the company’s $2.066B net revenue increase. Revenue growth combines volume, pricing and mix; it should not be read as shipment growth. 2. The next-quarter test Fiscal Q2 guidance calls for $12.2B–$12.6B revenue, 84.5%–85.5% non-GAAP gross margin and $55–$57 non-GAAP diluted EPS, all EPS pre-split. The revenue midpoint implies 12.4% sequential growth. The margin midpoint is slightly below Q1’s 85.1%. Management is forecasting further expansion with broadly stable margins, rather than another step-up in profitability per sales dollar. 3. Keep the weaker business in view Consumer revenue fell 29% sequentially. Datacenter’s larger contribution makes its purchasing cycle more important. This quarter establishes growth; it cannot establish how long today’s pricing and margins will last. 4. What the split changes The 8-for-1 split is expected to begin adjusted trading November 20, following the November 9 record date and November 19 distribution. The same adjustment applies to EPS. Comparing a post-split share price with pre-split earnings would understate the valuation multiple by eight times. The split itself creates no additional earnings or ownership. 5. What I would track next Datacenter demand, realized pricing, manufacturing costs and cash generation will determine whether this earnings level is durable. Supply expansion and customer spending changes are the risks to test. My read: the business evidence supports datacenter-led growth. Turning that into an investment return still requires a purchase price that leaves room for a less favorable memory cycle.

Memory: Datacenter HBM unlikely to move to 4-Hi 🔑 We share TrendForce’s view that the chance of de-spec to 4-Hi is low, vs worries over 4-Hi which could deliver same bandwidth with fewer dies, cut inference costs & stretch limited DRAM supply • Two key reasons 4-Hi unlikely fly in datacenter AI:
– Per-stack capacity is too small (~16GB for HBM5/HBM4E) → hyperscalers would need more chips, racks & scale-up networking
– Higher volume still increases base-die demand on tight advanced logic nodes (esp. N2 for HBM5) → little to no supply relief • Supply-chain feasibility low: mainstream production is focused on 8-Hi & 12-Hi, vs development cost and low-volume 4-Hi flows • Cost per Gb is higher on 4-Hi vs 8-Hi, with similar total cube cost based on current price quote. • Nvidia’s Rubin Ultra dropping to 8-Hi is well expected (our note on Aug 6) • For Rubin Ultra we expect parallel development of HBM4 8-Hi 192GB / HBM4E 8-Hi 256GB / HBM4E 12-Hi 384GB, with higher chance of HBM4 8-Hi first, then HBM4E 8-Hi as the mainstream SKU (consistent with our earlier AI Thermal report of maintain 2600W) • Reiterate our positive view on the DRAM/HBM sector #HBM# #DRAM# #MU#