Skip to main content
The Markets by Proactive
Go to Proactive UK
Proactive UK has moved. Proactive’s coverage of London’s small caps continues on proactiveinvestors.com Go there →
Advertisement
The Markets
by Proactive
Proactive UK has moved.
Coverage of London’s small caps continues on proactiveinvestors.com
Go to Proactive UK
The Markets
by Proactive
Proactive UK has moved.
Small-cap coverage continues on .com
Go to Proactive UK
Advertisement
The Markets
by Proactive
Proactive UK has moved.
Small-cap coverage continues on .com
Go to Proactive UK

Hardware & electrical equipment

Nvidia expands AI leadership with $1T sales outlook, Bank of America says

Nvidia Corp (NASDAQ:NVDA, XETRA:NVD) is set to extend its lead in artificial intelligence (AI) hardware and software, unveiling a broad, end-to-end AI product roadmap that now gives the company visibility to more than $1 trillion in data center sales over 2025-2027, according to a note from Bank of America.

Following Nvidia’s flagship GTC conference and a post-keynote meeting with the company’s CFO, Bank of America maintained its “Buy” rating on the chipmaker, with a price target of $300, citing its top position in AI inference and customized compute solutions.

The updated outlook doubles Nvidia’s previous data center sales forecast of $0.5 trillion for 2025-2026 and reflects confidence in supply chain stability. Standalone CPU and low-power LPU offerings announced by Nvidia are expected to provide additional incremental revenue.

Bank of America highlighted key elements of Nvidia’s expanded AI pipeline, which includes the Rubin (2026), Rubin Ultra (2027), and Feynman (2028) GPU systems, as well as SRAM-based LPUs that will be launched alongside LPX racks capable of up to 35 times higher throughput for low-latency AI workloads.

The note also pointed to Nvidia’s flexibility for customers, noting that its CPO scale-up and scale-out switch offerings remain optional, allowing continued use of traditional copper connections while providing more efficient optical alternatives.

Bank of America noted that Rubin systems (building on the Blackwell 2024-25 generation, which lowered token costs by up to 35 times compared with the prior Hopper generation) could further improve efficiency by two- to 10-fold across the full performance spectrum. The firm added that for high-end, low-latency workloads, efficiency could increase 35 times from one generation to the next when paired with SRAM-based LPX racks. “Tokens are the new commodity, and compute equals revenue,” analysts wrote.

Bank of America expects around 60% of Nvidia’s projected data center spending to come from the top five hyperscalers, with the remaining 40% from enterprise, industrial, and government customers.

Advertisement
The Markets
by Proactive
Proactive UK has moved.
Small-cap coverage continues on .com
Go to Proactive UK