Agentic AI workloads are shifting compute value from GPUs to CPUs and memory, creating a $60bn incremental CPU market opportunity by 2030.
According to Morgan Stanley's latest research, artificial intelligence is undergoing a structural pivot from raw compute to orchestration, with the economic value of the AI infrastructure stack shifting beyond accelerators. As AI moves from single-task generation to autonomous, multi-step "agentic" systems, CPUs and memory are emerging as the new bottlenecks alongside continued GPU demand. The report notes that while each model call now requires more coordination, persistent memory and system-level processing, "Agentic AI widens the trade beyond GPUs, with CPUs becoming the control plane for multi-step workflows and system orchestration."
The infrastructure transformation carries quantifiable implications. Morgan Stanley estimates that CPU-side orchestration can account for 50–90% of total workload latency in agentic systems, materially raising general-purpose compute intensity. The brokerage projects an incremental CPU total addressable market of $32.5–60 billion by 2030, within a total server CPU TAM of more than $100 billion. On the memory side, agentic workloads could drive 15–45 exabytes of additional DRAM demand by 2030, equivalent to 26–77% of 2027 annual DRAM supply.
The architecture of AI clusters is fundamentally changing to support this shift. Agentic workloads rely on CPU-centric or hybrid designs to manage reasoning, tool execution and memory orchestration, pushing the CPU-to-GPU ratio higher at the cluster level. Memory is no longer passive storage but an active system component supporting persistent context and continuous learning. This drives content growth across CPUs, DRAM and the broader infrastructure stack, including foundry capacity, ABF substrates and interconnect layers.
The investment implications span the full stack globally. "Supply-constrained enablers (foundry, ABF, BMC, interconnect) should capture outsized economics as system complexity rises," Morgan Stanley stated, with beneficiaries including CPU vendors, memory suppliers, storage companies, advanced packaging and substrate providers, foundries, equipment makers and server manufacturers. The brokerage expects orchestration CPUs to carve out an $82.5–110 billion data center market by 2030, with agentic workloads contributing the bulk of incremental growth. For investors, the report emphasizes that "The beneficiaries of this shift are global and full-stack," pointing to a wider and more diversified AI investment landscape.
As the transition from generative to agentic AI unfolds over the next five years, the infrastructure buildout will be more complex and capital-intensive than the first wave. Data center architectures will require re-optimization for coordination rather than just peak compute, placing greater emphasis on low-latency interconnects, high-bandwidth memory and resilient foundry supply chains. The winners will be those that can deliver end-to-end system efficiency rather than point solutions, and if execution keeps pace, CPUs and memory could rival GPUs as the primary drivers of AI-related semiconductor revenue growth.