Morgan Stanley Raises GigaDevice Target Price to 888 Yuan, with Bull Case Reaching 1,542 Yuan
Taylor Wilson
Morgan Stanley lifted its GigaDevice (兆易创新) target from ¥585 to ¥888 — a 50%+ hike — arguing that as global fabs shift resources to HBM and DDR5, mature products like DDR4 and NOR Flash are tightening, and GigaDevice sits at the intersection of rising prices and China's DRAM supply leverage.
How does Morgan Stanley get to ¥888?
The bank uses a residual-income model — a valuation method focused on future excess profits — cross-checked against a 2026 P/E of 53×.
EPS estimates for 2026/2027/2028 were raised 30%, 46%, and 53% respectively. The widening revisions signal an accelerating profit trajectory, not a one-off bounce.
The ¥1,542 bull case assumes stronger Flash sales and higher gross margins — an upside ceiling, not the base target.
This means → ¥888 is Morgan Stanley's "likely achievable" price; ¥1,542 is what happens only if everything goes right.
Why is DRAM shifting from side business to profit mainline?
GigaDevice's 2026 procurement budget for foundry DRAM from CXMT (长鑫集团) is $825 million (≈¥5.7 billion) — the supply-resource ceiling the company has set aside for DRAM.
Morgan Stanley's view: niche DRAM — small-category chips for specific use cases that don't compete head-on with Samsung — is moving from incremental revenue to a core profit driver, now embedded in the three-year earnings model.
In plain terms = global fabs are moving capacity toward HBM and DDR5, so "old" products like DDR4 are suddenly undersupplied. GigaDevice is positioned to fill that gap.
What do other banks say, and where do they disagree?
Goldman Sachs previously set a target of just ¥257 at 55× P/E. Bank of America highlighted DDR4 price increases and improving custom-DRAM visibility.
The three targets diverge sharply (257 vs 888 vs BofA unspecified), yet the core variable is the same: niche DRAM supply shortfall.
This reflects a market consensus that the gap exists — the real debate is how long it lasts and how much share GigaDevice can capture.
What justifies a 149% NOR Flash price hike?
Morgan Stanley assumes GigaDevice's 2026 NOR Flash ASP rises 149%. The logic has two layers: U.S. makers are cutting mature NOR capacity to redeploy into DRAM and NAND, while AI servers, optical modules, switches, and automotive applications are driving demand for high-reliability code storage.
The report notes that Nvidia's next-gen Vera Rubin rack may use 50%+ more NOR than the Grace Blackwell rack, opening a new demand vector.
In auto, intelligent cockpits, ADAS, domain controllers, and battery management all require certified Flash. Certification barriers are high and customer stickiness is strong — this means → once inside an automaker's supply chain, competitors find it very hard to displace the incumbent.
What role do SLC NAND and MCU play?
An MLC NAND shortage is forcing enterprise HDDs onto SLC NAND. Morgan Stanley expects SLC NAND to rise 50–60% in Q3 and possibly further in Q4. But the market is small — useful as supplementary upside, not as a standalone valuation pillar.
For MCU, the bank assumes 28% sales growth in 2026 with flat ASPs, supported by shipment recovery and industrial/auto design wins that underpin the revenue floor.
In plain terms = MCU's real value is not its own profit contribution but its role as a customer gateway. Clients using GigaDevice MCUs are more likely to extend purchases to NOR, SLC NAND, and analog — this determines whether the company evolves from a single-product cyclical supplier into a multi-product platform.
Which signals should investors watch next?
Morgan Stanley flags five verification metrics: DRAM revenue share, CXMT procurement execution, blended gross margin, inventory quality, and MCU/NOR new-product ramp progress.
These five data points will determine whether the ¥888 model reflects sustained earnings upgrades or a linear extrapolation from a price peak.
This means → the target looks attractive, but if CXMT procurement falls short or margins disappoint, the model's foundation weakens. Investors should track these five variables, not just the headline number.
Content is for reference only, not financial advice.