Industry outlook points to intensifying competition for data-center AI chips, with AMD highlighted alongside Nvidia and hyperscalers

A new 2026 to 2040 market outlook on computing and AI for data centers is placing AMD in the center of a broader shift: merchant GPU suppliers are competing not only with each other, but also with large cloud providers that are designing more custom silicon for their own workloads. The report, published through a Yahoo Finance technology roundup on Aug. 17, 2026, frames the period ahead as a structural change in how data centers buy and build compute for AI and general-purpose acceleration.

The analysis starts from the size of the opportunity it expects to grow. It describes the sector as having been valued at about $215 billion in 2025, then projects expansion out through 2040. Within that growth arc, the report argues the market is evolving as demand for AI training and inference increases pressure on performance, power efficiency, supply, and software ecosystems. That sets the stage for why the document focuses on both mainstream accelerators, like GPUs, and alternatives, like custom ASICs, which stands for application-specific integrated circuits built for particular data-center tasks.

A central theme in the report is how the competitive landscape is splitting across two fronts. On one side are established merchant GPU players, including Nvidia and AMD, which sell widely used accelerators that can be deployed across customers. On the other side are hyperscalers, including companies discussed in the report such as Google and AWS, which increasingly rely on proprietary silicon to capture more control over cost and performance for their own AI stacks.

The report also highlights what it calls the evaluation of different compute architectures and strategies. In addition to GPU competition, it discusses the relative roles of custom ASICs and platform choices such as Arm versus x86, with Arm referring to a processor instruction-set architecture often used by mobile devices and some servers, and x86 referring to the dominant architecture used in most traditional servers. While the Yahoo Finance summary does not provide a detailed AMD technical roadmap or specific product commitments, it does position AMD as part of the “strategies” being assessed against Nvidia and the hyperscaler-build approach.

For AMD specifically, the Yahoo Finance summary presents the company less as an isolated play and more as a participant in a market where compute buyers may increasingly mix and match chip types. The report’s emphasis on merchant GPU leaders suggests AMD’s role is evaluated alongside Nvidia’s, while the mention of hyperscalers and custom silicon implies that AMD and other GPU vendors face both co-investment and competitive displacement depending on how customers build their data-center systems.

The summary further indicates the report examines “strategies” across multiple players rather than only forecasting unit growth. That matters because chip demand in data centers is not solely a function of spending levels. It is also influenced by procurement preferences, integration with software frameworks, and the pace at which customers adopt new models and deployment patterns. In those conditions, a vendor’s software and interoperability profile can be as important as raw performance.

What the Yahoo Finance summary does not disclose is the specific numerical forecasts tied to each participant, such as projected market share by chip type for AMD, nor any explicit milestone timeline for AMD’s next-generation accelerators in the 2026 to 2040 window. It also does not include direct quotes from AMD executives, product guidance, or financial projections from the company. As a result, the takeaway from the available material is mainly about market structure and strategic evaluation, not new AMD-specific commitments.