SanDisk -15% Pullback|BofA and Bernstein Both Just Raised Their Targets

Chapter 1: The Dip That Divided Wall Street

SanDisk fell more than 15% over two sessions last week on a broad AI trade selloff after Broadcom earnings. For a stock up 550% year to date, the retreat looks like the first crack in a parabolic run. But Bank of America and Bernstein both used the dip to raise targets — not trim them. BofA's Wamsi Mohan lifted his target to $2,100. Bernstein's Mark Newman raised his to $3,000. The stock fell on macro sentiment. Two high-conviction analysts called it a buying opportunity. The tension resolves not on price, but on the contract structure underneath. SanDisk's New Business Models — NBMs — are multiyear supply agreements with fixed pricing floors and upfront buyer commitments. Five deals have been signed: more than $42 billion in committed revenue, $11 billion in guarantees. More than one-third of fiscal 2027 bit shipments are already locked in. That is the variable that determines whether the pullback is a trap or an entry.

Chapter 2: What the NBM Floor Actually Guarantees

The conventional memory cycle: strong pricing invited new supply, which crashed prices and crushed margins. Investors who bought at the top spent years recovering. Older long-term agreements were written in the customer's favor. The new generation locks in fixed prices with upfront buyer commitments. Bernstein's Newman found that as customers draw down contracts, the remaining guarantee covers a larger share — the downside protection strengthens over time. He then ran the stress test on a price collapse worse than 2010. With 60% of volume under NBM coverage, SanDisk's fiscal 2030 EPS would still come in at $214 per share. Without those contracts, the same downturn cuts earnings to $81 per share. That gap — $214 versus $81 — is the number the bears are not pricing. Morgan Stanley's Joseph Moore called the shift a fundamental repricing of the business after meeting management in late June. The buried assumption the market is making: that these contracts carry the same risk as prior-cycle deals. Bernstein puts the floor price at roughly 29 cents per gigabyte — in line with current market pricing. One set of articles explicitly states "the catch is these are cyclical-peak profits." BofA and Bernstein argue the floor is already contractually set. Two named desks, same event, directly opposed conclusions.

Chapter 3: The Supply Constraint That Makes the Floor More Credible

The NBM floor thesis requires the demand side to hold — and the supply picture reinforces it. NAND wafer starts are declining approximately 5% in fiscal 2026, with no significant new supply expected until 2028. Apple CEO Tim Cook called the current shortage a "hundred-year flood" and raised MacBook and iPad prices by $100 to $300. He urged policymakers to ease restrictions on Chinese memory suppliers — a signal that Apple sees no near-term relief. Goldman Sachs argues SK Hynix, Samsung, and Micron are all expanding capacity arriving in 2027 and 2028. That supply addition is real. But SanDisk's NBM contracts lock pricing above what Goldman's bear case models. The question is not whether supply normalizes — it is whether the floor in the contracts holds through the drawdown. The earliest verification point is Q4 FY2026 earnings: SanDisk guided EPS to $30 to $33 per quarter. If that guide lands with margins intact, the NBM thesis gains its first concrete evidence.

Chapter 4: What Holders and Watchers Each Watch Before Acting

The pullback created two separate decisions requiring different monitoring variables. For a holder who rode SNDK through the 550% run, the pullback is a thesis test, not a reversal signal. The counter-evidence that breaks the read is a customer pulling back from an NBM commitment — termination, renegotiation, or guarantee breach. No such evidence appeared. The pullback was macro-driven, not contract-driven. The bear case — cyclical-peak profits — remains an assertion about the future, not an observed event. That asymmetry is why BofA and Bernstein moved louder into the selloff. But the watch-list candidate faces a different problem: entering a stock up 550% after a 15% dip is not a mispricing catch. The entry setup confirms if Q4 EPS lands in the $30 to $33 guided range. The trap confirms if Q4 margins compress before new supply arrives — signaling the spot floor is not holding inside the contract period. Both conditions tie to one variable: whether Bernstein's 29-cents-per-gigabyte floor functions as the contractual minimum when tested. That is the metric to watch — not the next analyst price target.