Morgan Stanley Raises Longsys Target Price to 673 Yuan, AI NAND Shortage Prompts Profit Center Reassessment
0xBroomberg
Morgan Stanley more than doubled its Longsys (江波龙) target from ¥300 to ¥673, arguing the profit anchor should reflect the AI NAND shortage, the TCM model, and enterprise SSD capabilities — extending the earnings cycle through 2027.
Target doubled but rating unchanged — what is Morgan Stanley thinking?
The new target of ¥673 is anchored to 2027 estimated EPS of ¥42.62, implying roughly 15.8× P/E.
The rating stays Equal-weight: Morgan Stanley sees the current share price as already discounting much of the NAND upside, and prefers memory makers over module assemblers within the supply chain.
This means → Morgan Stanley concedes that Longsys's profit story stretches longer, but believes the market has priced in most of the good news — the re-rating has entered a "high bar for delivery" phase.
How large is the AI NAND gap?
AI-related NAND demand rises from 205 EB in 2025 to an estimated 400 EB in 2026 and 609 EB in 2027 — its share of total NAND demand climbs from 18% → 32% → 41%.
Non-AI demand is actually a drag: estimated non-AI NAND demand in 2026 sits below 2025 levels. Smartphones, PCs, and consumer SSDs are not delivering a strong recovery.
In plain terms = this NAND upcycle is almost entirely AI-driven; the traditional consumer side is not the protagonist.
Enterprise SSD shipments grew 139% year-on-year. High-capacity products (61.44 TB, 122.88 TB, etc.) hit elevated demand levels, with QLC — a flash design that stores more data per cell — becoming more common at those capacities.
Can price increases pass through across the board?
Enterprise SSDs still carry roughly 18–23% quarter-on-quarter price gains into Q3, and channel checks show TLC enterprise SSDs up about 30% q/q.
Mobile NAND rises only about 5–10%; consumer products see only modest increases.
This means → pricing has already bifurcated — enterprise eats the steak, consumer gets the broth. The shortage persists, but price hikes can no longer pass through indiscriminately.
What justifies Longsys's high gross margins?
Morgan Stanley lifts estimated gross margin to 55% for 2026, 35% for 2027, and still 27% for 2028; the long-term assumption rises from roughly 18% to a 25–38% range.
The TCM model — Total Cost Management, a framework that shifts part of the inventory burden to clients and ties demand, custom specs, and supply guarantees more tightly — gives Longsys more stable profit-sharing with high-spec customers.
In-house controllers and firmware raise the value of customization, reducing margin squeeze by original memory makers in embedded storage and enterprise SSD segments.
In plain terms = Longsys is no longer just a buy-low-sell-high module assembler; it is moving toward "custom solutions for clients" — and that shift is the bottom-layer logic behind Morgan Stanley's decision to roughly double its margin assumption.
Where is the hard constraint on cash flow?
Q1 2026 revenue reached ¥9.91 billion with net profit of ¥3.86 billion, yet operating cash flow was negative.
Inventory rose to ¥17.96 billion and prepayments increased significantly — the company is stockpiling aggressively for the upcycle.
This means → high inventory converts to profit during a price upswing, but if prices weaken earlier than expected, that same inventory becomes a pressure amplifier.
Is supply discipline in 2028 the biggest variable?
Multiple banks are bullish in unison: Goldman Sachs raised Kioxia's target to ¥116,000, Citi raised Western Digital's target to $2,500 — both expect the NAND shortage to persist into 2027 or even 2028.
Supply discipline holds only if two conditions are met simultaneously: memory makers keep prioritizing capex for DRAM and HBM (High Bandwidth Memory), and Chinese capacity additions ramp slower than AI demand growth.
Morgan Stanley's 2028 scenario test: if AI SSD demand still grows 50–60% y/y, tightness persists; if growth slows to 30% and new capacity accelerates, the market could flip to oversupply quickly.
This reflects a deeper point: the ¥673 target ultimately rests on whether 2027 EPS of ¥42.62 is delivered — not on a single-year profit surge in 2026. The four indicators to watch are gross margin, enterprise-segment growth, inventory-to-cash-flow improvement, and long-term contract pricing.
Content is for reference only, not financial advice.