Nvidia holds a lead over AMD in AI GPUs as demand accelerates, UBS says

Nvidia is maintaining its advantage over Advanced Micro Devices in artificial intelligence GPU (graphics processing unit) shipments, according to an assessment cited by UBS, as AI-related demand continues to rise. The comparison centers on who is supplying the GPUs used to train and run AI models, a market that has become a primary driver of semiconductor spending over the past two years.

The UBS view, as reported in a Yahoo Finance post, suggests Nvidia’s position is still stronger than AMD’s on shipping volumes. The note frames the AI chip race as an execution contest, where ramped production, customer adoption, and product availability determine which suppliers win the next waves of orders.

UBS also linked the demand outlook to a shift in how companies plan to use AI. Instead of only deploying AI for existing workloads, UBS described accelerating “agentic” demand, where software agents that can perform tasks more autonomously increase compute requirements. More compute translates into more GPUs, faster refresh cycles for data center systems, and additional demand for AI infrastructure components across the supply chain.

The same reported UBS perspective highlights Nvidia’s product cadence in AI hardware. In that framing, Nvidia’s Blackwell platform is positioned as a leading contributor to demand momentum, while AMD is characterized as working through its own ramp efforts with next-generation accelerators. Competitors remain active, but UBS’s bottom line in the report is that Nvidia remains ahead in AI GPU shipments.

UBS’s broader supplier lens, again as summarized in the Yahoo Finance-linked coverage, extends beyond Nvidia and AMD. It also referenced other parts of the AI buildout, including arm-based ecosystem wins and additional semiconductor exposure tied to the AI infrastructure cycle. The takeaway is that customers are spending in layers, not just on GPUs, which can keep the entire hardware stack relevant even as individual chip suppliers compete for share.

For NVIDIA, the competitive narrative matters because its business is increasingly concentrated in data center AI systems, where GPU supply and platform integration drive revenue. In practice, companies buy not only chips but also the full systems and software stack needed to run large-scale workloads efficiently. That makes it harder for customers to switch suppliers quickly, especially when training and inference performance, deployment tooling, and availability are all tightly coupled.

Still, much of the detailed evidence behind the UBS claim is not disclosed in the publicly visible portion of the Yahoo Finance-linked coverage. The report does not specify the time period covered by the shipment comparison, does not publish exact share or unit figures, and does not quantify how much of the demand increase is attributable to agentic deployments versus broader AI capex spending. Investors and industry watchers will likely look for clearer, more granular disclosures from Nvidia and its rivals in upcoming earnings and product updates.

The next question for the market is whether Nvidia’s lead persists as AMD scales its next-generation offerings and as customers continue expanding AI capacity into late 2026 and beyond. Watch for more explicit datapoints around AI accelerator shipments, supplier mix at major cloud and enterprise buyers, and whether system-level bottlenecks, such as memory and interconnect capacity, constrain or amplify whichever vendor wins the most orders.