NVIDIA flags supply constraints and cost pressure as bullish AI demand outlook continues

NVIDIA’s outlook discussion during its most recent quarterly earnings call centered on a familiar tension for the AI industry: demand remains very strong, but the pace of fulfillment can be constrained by upstream bottlenecks and component costs. In market coverage of the call, the company’s forward view emphasized that supply availability and memory-related costs are important factors that could affect near-term results even as demand for AI compute keeps accelerating.

According to the reporting of the call, NVIDIA expects fiscal 2028 revenue to rise by roughly 70%. That projection frames the company’s longer-term revenue growth case, but it is paired with explicit warnings that physical constraints in the supply chain can limit how quickly systems can be built and delivered.

The coverage also highlighted that supply limits were a central part of the company’s near-to-intermediate outlook. In other words, the gating factor may not be customers’ willingness to buy AI infrastructure, but rather the ability to obtain enough of the relevant components and capacity to meet those orders.

Another pressure point mentioned in the call coverage is higher memory costs. For AI servers and data center systems, memory and related packaging and sourcing constraints can influence both bill of materials and availability, which in turn can affect margins and the speed at which revenue can be recognized.

While the company’s longer-range growth trajectory appears intact in the discussion, the emphasis on supply constraints suggests NVIDIA is trying to manage expectations about how quickly incremental demand turns into shipments and financial results. Investors and customers have increasingly focused on whether AI infrastructure can scale faster than the industry can produce and assemble the necessary hardware.

NVIDIA is one of the main suppliers of accelerated computing hardware used to train and run AI models, selling GPUs (graphics processing units) and related platforms into data centers. When AI spending rises, the company’s results often reflect both the number of systems that can be shipped and the pricing and margin dynamics that depend on component supply, including memory.

A key caveat is that the available record here reflects market coverage rather than a verbatim transcript or NVIDIA’s full investor materials. The company’s precise language on how supply limits and memory costs would flow through to specific financial line items, time periods, or customer commitments was not included in the information provided for this story.

Looking ahead, the items to watch are whether NVIDIA later quantifies the supply situation (for example, by describing expected availability timelines or capacity increases) and whether it provides more detail on how memory cost pressure is expected to evolve. That guidance would be important for understanding how much upside from sustained AI demand can be realized beyond the constraints.

If NVIDIA’s delivery cadence improves faster than the market expects, revenue growth could track closer to the upper end of its longer-term expectations. If constraints persist or costs remain elevated, the company may continue to describe demand as strong while still tempering how quickly shipments and margins can respond.