Morgan Stanley predicts a massive production increase for NVIDIA’s Blackwell GPUs, expecting up to 800,000 units to be produced by Q1 2025. This follows a 3x surge in output from Q4 2024, solidifying Blackwell’s critical role in NVIDIA’s revenue growth.
Morgan Stanley Expects NVIDIA to Ship 450,000 Blackwell GPUs in Q4, Generating $10 Billion
In recent weeks, speculation has dominated tech news regarding the production cadence of NVIDIA’s latest GPU architecture, Blackwell. While NVIDIA addressed some concerns during a recent earnings call—confirming that minor design flaws had been rectified via photomask changes—the accurate scale of the company’s production ramp-up is becoming more apparent. Morgan Stanley predicts a significant increase in production, with Blackwell volume expected to triple between Q4 2024 and Q1 2025.
In September, Morgan Stanley issued a critical note on NVIDIA, forecasting the shipment of 450,000 Blackwell GPU units in the December-ending quarter, potentially generating $10 billion in revenue. At the time, the investment bank acknowledged that NVIDIA was still ironing out a few “technical challenges” related to its GB200 server racks, noting that such issues are part of the “normal debugging process” for new product launches.
Morgan Stanley Predicts NVIDIA Blackwell GPU Production to Triple by Q1 2025, Surpassing Hopper Revenue
Morgan Stanley has released an updated investment note, confirming its previous estimate of between $5 billion and $10 billion in Blackwell-derived revenue for Q4 2024, following its channel checks. The firm also provided insights into NVIDIA’s production outlook for early 2025. Specifically, Morgan Stanley predicts that NVIDIA will produce between 700,000 and 800,000 units of Blackwell GPUs in Q1 2025, representing nearly a 3x increase in production volume compared to Q4 2024.
The investment note highlights:
“And into 1Q25, we understand that Blackwell chip volume may reach 750k-800k units, up almost 3x from 4Q24.”
Additionally, Morgan Stanley offered projections for NVIDIA’s Hopper architecture. The bank expects Hopper’s production volume—including the H200 and H20 chips—to reach around 1.5 million units in Q4 2024 and gradually decline to 1 million units in Q1 2025.
Crucially, given that the B200 chip is priced at a 60-70% premium compared to the H200, Morgan Stanley anticipates that Blackwell-derived revenues will surpass those generated by the Hopper architecture in Q1 2025, solidifying Blackwell’s importance in NVIDIA’s future growth trajectory.


OpenAI Targets $600B Compute Spend as IPO Valuation Could Reach $1 Trillion
xAI’s Grok Secures Pentagon Deal for Classified Military AI Systems Amid Anthropic Dispute
Microsoft Gaming Leadership Shake-Up: Phil Spencer Retires, Asha Sharma Named New Xbox CEO
Pentagon Weighs Supply Chain Risk Designation for Anthropic Over Claude AI Use
Samsung Electronics Stock Poised for $1 Trillion Valuation Amid AI and Memory Boom
BlueScope Steel Shares Drop After Rejecting Revised A$15 Billion Takeover Bid
Samsung and SK Hynix Shares Hit Record Highs as Nvidia Earnings Boost AI Chip Demand
OpenAI Hires Former Meta and Apple AI Leader Ruomin Pang Amid Intensifying AI Talent War
Lynas Rare Earths Shares Surge on Strong Half-Year Earnings and Rising Global Demand
Middle East Airspace Shutdown Disrupts Global Flights After U.S.-Israel Strikes on Iran
OpenAI Pentagon AI Contract Adds Safeguards Amid Anthropic Dispute
Hyundai Motor Group to Invest $6.26 Billion in AI Data Center, Robotics and Renewable Energy Projects in South Korea
Nintendo Share Sale: MUFG and Bank of Kyoto to Sell Stakes in Strategic Unwinding
Anthropic Refuses Pentagon Request to Remove AI Safeguards Amid Defense Contract Dispute
Nvidia Earnings Preview: AI Chip Demand, Data Center Growth and Blackwell Shipments in Focus
APEX Tech Acquisition Inc. Raises $111.97 Million in NYSE IPO Under Ticker TRADU
Amazon’s $50B OpenAI Investment Tied to AGI Milestone and IPO Plans 



