NVIDIA Earnings Preview: Morgan Stanley Expects Another "Beat-and-Raise," but Stock Reaction Remains Uncertain
Nashnova编辑部
Morgan Stanley expects Nvidia to deliver another beat-and-raise next week, but the stock has fallen the day after earnings four quarters in a row — the real test is no longer the numbers but whether margins hold, market share sticks, and Rubin proves Nvidia's edge in the next AI infrastructure cycle.
What is Morgan Stanley forecasting?
For the July quarter (F2Q27): revenue of $91.1 billion, EPS of $2.07. For the October quarter (F3Q27): $102.3 billion in revenue, EPS of $2.34 — both roughly in line with consensus.
The firm keeps its Overweight rating, $288 price target, and top-pick status on Nvidia.
This means → Morgan Stanley does not expect a miss. Its numbers nearly match Wall Street consensus; the disagreement is not about the figures themselves but about what comes after.
Is there still room for upward revisions?
During a June Asia field trip, Morgan Stanley found that 4nm wafer starts for Blackwell are already rising — suggesting current supply assumptions may be too conservative and earnings estimates could move higher.
But the market has already priced in October-quarter supply signals. The biggest near-term upside window is likely the January quarter, when Rubin — Nvidia's next-generation AI chip architecture — begins shipping at larger scale.
In plain terms = this quarter is not where the surprise lives. The beat that could genuinely move estimates is two quarters out.
Is Rubin shipping behind schedule?
Nvidia has confirmed Rubin will begin shipping in Q3. Morgan Stanley estimates roughly 150,000 units, worth about $9 billion in revenue at roughly $60,000 per chip.
KYEC — a key Nvidia test partner — recently cut its full-year guidance and pushed AI-related revenue into late 2026 and 2027, sparking concern about Rubin's ramp.
Morgan Stanley's Asia semiconductor analyst Charlie Chan sees this as a longer test cycle, not lower volumes — the firm's updated model keeps total Rubin test volume unchanged, shifting weight from Q3 to Q4. Even if Q3 shipments undershoot, the 150,000 baseline is already well below revised industry estimates, limiting the impact on Nvidia's overall revenue trajectory.
What are the three key debates?
Margins: Management is expected to reiterate "mid-70s %" gross margins for FY2027, but cost pressures from DRAM, front-end wafers, packaging, and substrates persist. Morgan Stanley is skeptical that margins will rebound in the second half of FY2028, as consensus currently assumes.
Circular financing: CEO Jensen Huang has already explained the logic behind a $500 billion partnership framework via a blog post. Morgan Stanley expects no material change in messaging on the earnings call.
Market share: ASIC — custom chip — competition is heating up, and AMD continues to push into AI accelerators. Management will likely highlight Rubin's economic advantages over Blackwell for AI factories, but the market needs more shipment data and customer feedback before judging Nvidia's real competitive position.
Is the market underestimating 2027 growth?
Morgan Stanley forecasts Nvidia's FY2027 revenue at $393 billion and FY2028 at $598.8 billion — the latter well above the consensus estimate of $562.4 billion.
At GTC, Huang said Nvidia has $1 trillion in demand visibility for Blackwell and Rubin combined over 2025–2027, including networking but excluding Groq, standalone CPU, RTX, and software.
In plain terms = consensus data-center revenue for that period totals roughly $1.054 trillion; Morgan Stanley's forecast is $1.09 trillion. Given that 2025 already includes about $30 billion in Hopper and networking revenue, the bar is not set especially high.
Will the stock rally after earnings?
Nvidia has fallen the day after earnings four consecutive quarters. This reflects a shifting market calculus — beat-and-raise is no longer a catalyst but the minimum expectation.
This means → the real test next week is not whether Nvidia beats again but three deeper questions: Can margins stabilize? Can market share hold? Can Rubin prove Nvidia's lead in the next round of AI infrastructure competition?
Put simply = acing the test no longer earns extra credit. The market now wants to know whether the straight-A streak can last.
Content is for reference only, not financial advice.