AMD expects 2027 data-center growth to be driven by AI inference demand

Advanced Micro Devices said it expects continued rapid growth in its data-center and server CPU businesses, pointing to accelerating demand for AI inference systems through 2027. In remarks attributed to Matt Ramsay, AMD linked the outlook to the shift from training AI models toward running them in production, where inference workloads can require significant compute across servers and data centers.

Inference is the step where an AI model takes input and produces an output, such as generating text, recognizing images, or planning actions. AMD characterized this phase as increasingly important for customers building large-scale AI systems, and said that demand for inference in the coming years should support higher volumes and greater platform spending.

The company also pointed to “agentic” AI, a term used for systems that can take multi-step actions toward a goal rather than answering a single prompt. AMD’s framing suggested these types of workloads increase compute needs and create additional pressure on data-center infrastructure, which in turn can lift demand for server CPUs and related hardware.

AMD’s comments were reported in a market update distributed through Yahoo Finance and published by MarketBeat on Aug. 11, 2026. The update focused on the forward-looking theme of AI inference demand rather than providing new financial guidance, segment numbers, or quantified targets for data-center revenue or market share.

While the update underscores a demand backdrop, it did not lay out specific assumptions such as forecast server shipments, customer deployments, or the expected contribution from particular product lines. It also did not specify whether AMD expects any change in product mix, pricing, or gross margin drivers tied to inference-heavy deployments.

Sector-wide, semiconductor demand tied to AI has tended to track two major use cases: training, which is concentrated in large clusters, and inference, which can be deployed broadly across services as models are integrated into applications. If inference becomes a dominant workload driver, suppliers like AMD may benefit not only from new builds but also from ongoing refresh cycles as operators add capacity to handle higher request volumes and lower latency requirements.

Still, investors and customers will likely want more detail on how quickly inference-driven build-outs translate into revenue. As of the Aug. 11 market update, AMD did not disclose specific 2027 revenue targets, backlog figures, or timeline granularity in the reported remarks.

For AMD and the broader server ecosystem, the next key questions are whether inference demand leads to sustained multi-quarter ordering, how customers are budgeting for agentic workloads, and whether competitive offerings change the mix of platforms deployed in data centers. AMD’s ability to convert inference-driven demand into durable share gains will be watched through subsequent earnings calls, product milestones, and guidance updates.

AMD expects continued growth in data centers as AI inference demand expands through 2027, according to comments attributed to Matt Ramsay.

AMD linked inference workloads, and agentic AI systems that take multi-step actions, to higher compute needs in data centers.

The reported remarks focused on demand drivers rather than providing new financial guidance or quantified targets.

The update was disseminated via Yahoo Finance and published by MarketBeat on Aug. 11, 2026.

No additional product specifications, customer orders, or backlog figures were included in the reported market note.

Sector context: inference is the production step where AI models generate outputs, often requiring substantial server capacity at scale.