Just this Thursday, Nvidia’s better-than-expected performance is already a sure thing, and the market is most concerned about the speed of Blackwell’s capacity ramp-up.

```

Author: Dong Jing

Source: Hard AI

This week, the global market’s attention will once again focus on the AI giant—Nvidia, which will release its Q3 earnings report this Thursday. JPMorgan stated that Nvidia is highly likely to again deliver a “beat-and-raise” performance.

On November 18, according to Hard AI, JPMorgan’s latest research report pointed out that Nvidia’s growth speed is currently determined not by demand, but by the capacity limits of its vast supply chain. Q3 revenue is expected to exceed the market’s consensus of about $55 billion, with guidance as high as $63-64 billion, markedly above the market’s expectation of $61.5 billion.

The report said that another robust “beat-and-raise” performance is already anticipated, with the upside mainly depending on Nvidia’s supply chain expansion capability over the next three months.

According to JPMorgan’s analysis, AI computing power demand continues to significantly outstrip supply, and Nvidia’s largest client groups—including hyperscale cloud providers, emerging cloud companies, and AI labs—still face compute bottlenecks. Regarding supply chain capacity, the bank expects Q3 rack shipments of Blackwell/Blackwell Ultra to realize about 50% QoQ growth, reaching about 10,000 racks. Looking ahead, this growth momentum is expected to continue into Q4.

The report highlights that, for the earnings call, market attention will focus on management’s statements about the growth trajectory for Blackwell/Blackwell Ultra products in the first half of fiscal year 2027, as well as responses to key issues such as AI spending sustainability, data center infrastructure power limitations, and the impact of component cost inflation on gross margins.

“Supply shortage” remains the main theme, beating expectations is highly probable

At the outset, the bank’s report makes clear that the current discussion about Nvidia in the market no longer centers on whether demand is healthy, but on whether its supply chain can keep pace with customers’ ambitious AI compute deployment plans.

JPMorgan states that, in the “initial stage” of a multi-year AI data center spending upcycle, demand for AI compute still “significantly outpaces supply.” Major clients—including hyperscale cloud providers, emerging cloud services, and AI research institutions—remain in a compute shortage even after more than two years of AI investment frenzy.

Against this backdrop, JPMorgan forecasts that Nvidia’s quarterly results will once again exceed Wall Street expectations. Specifically, the bank expects fiscal Q3 revenue through October 2026 to surpass the general market expectation of around $55 billion, and predicts the company will guide Q4 revenue to reach the $63-64 billion range, above the consensus of $61.5 billion.

The report emphasizes that another “beat-and-raise” performance is “in the cards,” and the size of the earnings surprise will fully depend on the scale of supply chain capacity increases over three months.

Supply chain capacity: the only “brake” on growth

Since demand is not an issue, supply is the sole constraint.

The report analyzes Nvidia’s supply chain execution in detail. Over the past 3-4 months, its partners performed strongly in Blackwell/Blackwell Ultra rack shipments, with Q3 shipments expected to realize about 50% QoQ growth, reaching about 10,000 racks.

Looking forward, this growth momentum is expected to continue into Q4. JPMorgan predicts Nvidia’s total rack shipments for fiscal 2026 will reach 28,000 to 30,000 racks.

Although additional 3nm and CoWoS chip capacity in the next twelve months will provide some upside for GPU shipments, the bank firmly believes that supply chain capacity will remain the key constraint for Nvidia’s revenue growth, persisting through 2026.

Nevertheless, the report presents an extremely optimistic long-term signal.

JPMorgan believes that Nvidia’s supply chain is capable of supporting a year-on-year doubling of rack shipments to 60,000–70,000 in fiscal 2027/calendar year 2026.

Most crucially, according to disclosures by Nvidia at the October GTC conference, its order backlog for calendar 2026 has already surpassed 70,000 racks. This means the company’s orders in hand exceed maximum annual capacity for next year, providing great future growth certainty.

Based on this, JPMorgan maintains its “Overweight” rating for Nvidia, with a target price of $215—15% above current levels.

Beyond the numbers: four long-term factors investors should watch

JPMorgan believes that, for the upcoming earnings call, stock price reaction will depend more on management’s outlook than just the numbers. Investors should closely watch management’s responses to the following four major concerns:

Blackwell/Blackwell Ultra’s ramp-up trajectory: The speed of capacity expansion going into the first half of 2026 (Nvidia’s fiscal 2027 first half).
Sustainability of AI spending: As per a recent in-depth report by JPMorgan’s global team, funding sources for AI are expected to remain ample through 2030.
Impact of power constraints: About 120 GW of new global data center electricity capacity are expected to come online in the next five years, but new natural gas turbine delivery lead time has surged to 3-4 years, and nuclear plants take over 10 years to build. Power supply has become a real bottleneck.
Impact of component cost inflation on gross margin: Especially pressures from rising prices of memory, chips, and other raw materials.

Notably, the report provides a detailed breakdown on the gross margin, which investors are particularly concerned about.

JPMorgan believes rising LPDDR memory prices pose a greater pressure than HBM memory. While HBM4 is expected to command a 30–40% higher average selling price versus HBM3e, Nvidia can proactively factor this cost into pricing for its next-gen Rubin platform.

However, for LPDDR memory, the company may have to accept market variable pricing—even long-term supply agreements only guarantee volume but do not lock in prices.

Nevertheless, JPMorgan believes Nvidia can still achieve its mid-70% gross margin target by the end of fiscal 2026, though further upside will be quite limited due to input cost inflation.

This article is from WeChat Official Account "Hard AI". For more frontier AI news, visit here

Risk Warning and DisclaimerThe market carries risk; investment requires caution. This article does not constitute personal investment advice, nor does it take into account the specific investment objectives, financial status, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular situation. Investing based on this is at your own risk. ```