NVDA's stronger earnings report, yet flatter stock price? Morgan Stanley's research report lays out the real reason.

First, a plain-language summary:
NVDA's stock price is stagnant, and the market is essentially asking: "It's strong now, but can it still be this strong in 2027?"
The answer is:
● Judging from actions across the supply chain to the hyperscalers, money isn't just about slogans; it's about using prepayments to lock in capacity for the next few years.
● Market share may fluctuate in the short term, but customer attention will focus on Rubin/the roadmap.
● One of the most critical upcoming public validation points is GTC: first address concerns about share/leadership, then gradually give the market more confidence in sustainability.

Detailed interpretation below:
Also part of the AI chain: memory stocks have risen 300%–900% over the past few months, while NVDA has barely moved, yet current-quarter earnings expectations have been revised up 38% over 6 months—this is the contradiction Morgan Stanley's report aims to resolve.
Report's Core Conclusions
①NVDA's short-term underperformance isn't due to poor data, but market concerns about two things: market share & growth sustainability.
②Share concerns will be addressed at the upcoming GTC: The report believes GTC will showcase a clearer leadership roadmap, putting NVIDIA ahead of the competition again.
③Sustainability is harder to prove, but the report emphasizes: based on various surveys/supply chain feedback, hyperscaler capex looks more like a multi-year upcycle, and confidence will gradually build.
④Report action: Puts NVDA back as a semiconductor Top Pick, gives an Overweight rating; also sets a $260 price target.

Key Data
A. Where does the "stock stagnant but expectations rising" contradiction come from?
NVDA's stock price is nearly "flat," but current-quarter earnings expectations are up +38% over 6 months; in contrast, memory stocks are up +300% to +900% over the same period.
B. Valuation and Earnings Framework from the Report
Report lists FY ending in Dec: EPS (12/26–12/29e) 4.77/7.93/10.14/13.11; corresponding P/E 38.5/22.4/17.6/13.7. Report's growth assumptions on the "Risk/Reward" page: 2026 revenue +63.8%, 2027 +27.8% (model assumption, not a forecast). Another key input: GAAP Revenue ($mm) 215,938/353,807/452,409/557,971 (2026–2029e); Gross Margin (MW) 71.3%→74%+ range; MW EPS 4.61/7.92/10.05/12.96.
C. Competitive Landscape
Report's revenue share estimates: NVIDIA ~85%, ASIC slightly above 10%, AMD slightly below 5%.
Logic Breakdown
Logic 1: What exactly is the market afraid of?
Report attributes stock performance to two types of concerns: share and growth sustainability.
The "downstream version" of sustainability concerns is: negative cash flow in some cloud infrastructure businesses could theoretically squeeze future investment space, leading the market to question whether spending will slow next year.
Logic 2: Why does the report dare say "the cycle is more than one year"?
It points to very hard evidence: prepayments/prepaid orders.
Summary: Hyperscalers are placing 3-year orders with memory suppliers, with extreme cases like "receiving 100% prepayment for 2028 revenue this quarter," indicating strong locked-in demand, not mere forecasting. Further, the report cites Jensen Huang's explanation on the call: these investments can generate higher earning power than market models assume. The report also mentions a market participant's intuitive take: the compute supply-demand gap is still widening.
Logic 3: Will share be eaten away?
Report expects peers may grow slightly faster than NVIDIA in CY26, with share possibly dipping 1–2 percentage points, for a simple reason—size is too large, making continued high-speed growth harder in a supply-constrained environment. But it emphasizes customers care most about Rubin; with supply still tight, competition dynamics will still revolve around Blackwell.
Another noteworthy point: Even as top customers "bet on multiple architectures" (using NVIDIA/AMD/TPU/in-house ASIC), the report's surveys still show many scenarios favor NVIDIA; and it notes that among the largest ASIC user and largest potential AMD user, "their respective business with NVIDIA could still grow 80%+ in CY26."
Logic 4: A point most overlook—bottlenecks will shift
AI processor supply tightness may ease in the coming months, but this isn't necessarily negative for NVDA, as the real bottleneck may shift to DRAM, eSSD, HDD, optical modules/optics, microprocessors, power, etc.; once GPU constraints ease, it could actually lead to share re-acceleration (the report uses CY24 experience as an analogy).
Logic 5: The role and limits of GTC
GTC may not completely convince the most skeptical about capex sustainability, but it can first pull market confidence back to "share/roadmap"; and it's expected to resemble 2024's "laying out a 4-year roadmap," emphasizing the race isn't just about chips, but also racks and ecosystems; Groq-related IP may also enter the roadmap narrative.

Highlights (the truly valuable part of this report)
Prepayments aren't a sentiment indicator, they're a hard constraint: Extreme actions like receiving 100% prepayment for 2028 revenue this quarter are used to prove multi-year supply lock-in.
The contradiction is obvious: upstream says growth for three years, downstream cash flow looks ugly—and it provides Jensen Huang's explanatory framework: these investments may earn more in the future than market models assume.
Share isn't a binary of being crushed by ASIC or crushed by AMD: Reality is more likely multi-architecture coexistence, but it also notes certain top customers' business with NVDA could still grow 80%+ in CY26.
GPU supply easing doesn't necessarily mean demand disappears: Bottlenecks may shift to DRAM/storage/optics/power, etc. If the market only focuses on CoWoS and lead times, it may misread demand.
Clearly highlights the ceiling factors for valuation upside: Market cap, index weight, 85% share, extremely high margins make it harder to "significantly raise expectations further."

Feel free to let me know what else you'd like to see. This article is solely for report analysis and information sharing and does not represent personal views. This article is solely for report analysis and information sharing and does not represent personal views.$NVIDIA(NVDA.US) $AMD(AMD.US) $Lumentum(LITE.US)