NVIDIA GB200, B200, and NVL Rack Server Shipment Forecasts: Insights from Wall Street Analysts Ahead of Next Week’s Earnings

This article does not constitute investment advice, and the author holds no shares in any of the discussed stocks.

The Diverging Views on NVIDIA’s GB200 GPU Impact

Despite analysts on Wall Street expressing conflicting opinions regarding NVIDIA’s GB200 GPU, concerns about its slower production ramp-up and the implications for upcoming earnings are increasingly under scrutiny. While Baird analyst Tristan Gerra downplays the ongoing issues, suggesting these concerns will have minimal long-term effects, HSBC analyst Frank Lee adopts a more cautious stance, revising his forecast for NVL rack server shipments down from 35, 000 units to 30, 000 units.

Current Market Insights

Gerra has utilized insights from supply chain discussions to indicate that GB200 shipments are finally gaining traction after what has been described as a “late and slow start.”He attributes initial delays not only to the need for optimization of architectural innovations but also to restrictions in existing data center infrastructure, such as electricity and space. He emphasizes:

“Investors should not assume that these initial delays will lead customers to skip to the next-generation products. We expect GB200 to represent the majority of NVIDIA’s GB mix in 2H25 and into 2026. GB300 crossover is expected not until 2H26.”

HSBC’s Revised Projections

In contrast, Lee’s cautious approach has led him to lower the price target for NVIDIA’s stock from $185 to $175, while still retaining a ‘Buy’ rating. He expresses concerns that ongoing supply chain issues surrounding the GB200 could potentially weigh down performance leading into the first half of FY26. Notably, NVIDIA’s fiscal year ends in January, resulting in some misalignment with the calendar year ending in December.

Anticipated Earnings and Revenue Projections

Lee has further downgraded his fiscal year 2026 estimate for NVL rack server shipments and anticipates significant revenue contributions from B200 GPU shipments in the first half of FY26. This allows NVIDIA to potentially “book Blackwell revenues of approximately USD 5 billion and USD 10 billion for 4QFY25 and 1QFY26, respectively, ”leading to projections of total revenues of USD 40 billion to USD 42.2 billion for those quarters, surpassing consensus estimations of USD 38.2 billion and USD 42 billion.

Future Forecasts and Analyst Consensus

For the forthcoming earnings report, Lee predicts NVIDIA will realize approximately $40 billion in revenue for Q4’25, exceeding both the company’s guidance of $37.5 billion and Wall Street’s consensus estimate of $38.2 billion. He concludes with anticipated sales figures for the first half of FY26:

“We also expect 1QFY26/2QFY26 sales of USD 42.2 billion/USD 55.4 billion vs consensus of USD 42 billion/USD 46.2 billion, implying H1 FY26 is unlikely to disappoint the market despite ongoing supply-side issues.”

KeyBanc’s Perspective on Manufacturing Constraints

KeyBanc analyst John Vinh acknowledges that manufacturing constraints are indeed limiting GB200 NVL server rack shipments but believes other factors will more than compensate for this. He highlights that delayed initial manufacturing yields have allowed customers to defer GB200 orders while shifting to B200 servers.

Additionally, Vinh notes the rising demand for H20 GPUs from Chinese cloud service providers (CSPs) due to constraints in other products. He also points out that NVIDIA customers, particularly CSPs, are effectively financing their inventories at EMS providers, meaning that sell-in shipments from NVIDIA are recognized as revenue.

Oppenheimer’s Take on H200 and GB300 GPUs

Oppenheimer’s analyst, Rick Schafer, reports that the demand for H200 GPUs is exceeding expectations, predicting they will make up about 30 percent of NVIDIA’s sales mix this year. Regarding the upcoming GB300 GPU release, anticipated for late 2025, he comments:

“We believe GB300 (expected late CY25) likely involves a single rack (NVL72) configuration, suggesting DC cooling challenges are being addressed.”

Moreover, Schafer indicates that the increased capital expenditure of hyperscalers is projected to rise by 40 percent this year, surpassing $300 billion, compared to $220 billion in 2024.

AI ASICs and GPUs: A Future Coexistence?

Aligning with earlier discussions, Schafer emphasizes that ASICs will likely coexist with GPUs, with ASICs serving internal workloads while GPUs cater to more dynamic AI applications. He states:

“CSP custom AI ASIC projects coexist with GPUs, in our view—ASICs primarily used for internal workloads while GPUs used for broader, dynamic, higher performance AI applications. NVDA remains best positioned in AI, benefiting from full-stack AI hardware/software.”

Final Thoughts

As NVIDIA approaches its upcoming earnings announcement, analysts will be keenly observing their ability to deliver a triple beat in revenue, earnings per share (EPS), and guidance. What are your expectations for NVIDIA’s performance? Share your insights in the comments section below.

Source&Images