BNP Paribas Rates Nvidia 'Outperform' With $285 Target, Sees 36% Upside on Data Centre Demand

BNP Paribas has issued an "Outperform" rating on Nvidia (NASDAQ: NVDA) with a price target of $285, implying an upside of approximately 36% to the stock, according to a brokerage assessment published by Economic Times Markets on 27 August 2026.

The brokerage placed Nvidia at the heart of what it described as the data centre revolution, pointing to the industry-wide shift toward accelerated computing — the use of specialised processors such as graphics processing units (GPUs) to run demanding artificial intelligence and high-performance workloads, rather than relying on general-purpose CPUs alone. That shift has coincided with multi-billion-dollar capital expenditures by the world's largest cloud providers — including Microsoft, Amazon, Alphabet and Meta — to expand AI computing capacity, spending that has run alongside surging demand for AI accelerators.

Nvidia, headquartered in Santa Clara, California, is a leading designer of the GPUs and AI accelerators that underpin modern data centre infrastructure, and its platforms are widely used by cloud providers and enterprises building out AI computing capacity. Its data centre business has grown to become the company's largest source of revenue. On brokerage rating scales, an "Outperform" rating generally denotes an expectation that a stock will deliver returns ahead of the analyst's benchmark or coverage universe.

In support of its stance, BNP Paribas also cited Nvidia's strong second-quarter (Q2) financial results, and noted robust demand for the company's Blackwell and Vera Rubin platforms. Blackwell is Nvidia's current-generation data centre GPU architecture used for AI training and inference, while Vera Rubin is the follow-on platform on the company's publicly disclosed roadmap, continuing Nvidia's practice of naming its architectures after pioneering scientists. Nvidia is not the only supplier chasing this demand: AMD's Instinct accelerators and cloud providers' in-house chips, such as Google's tensor processing units (TPUs) and Amazon's Trainium, compete for the same AI workloads, a competitive landscape that features prominently in analyst assessments of the company.

The brokerage observed that demand for these platforms has remained robust despite ongoing supply chain constraints. Constraints of this kind have been a recurring feature of the AI chip industry, with tight availability of high-bandwidth memory (HBM) and advanced packaging capacity widely reported as limiting factors for accelerator output. For those following the story, Nvidia's subsequent quarterly reports and the capital-expenditure guidance issued by its largest cloud customers are among the indicators analysts routinely track to gauge how the data centre build-out is progressing.

Source: Economic Times Markets