HONG KONG, 20 April 2026 – The next wave of artificial intelligence is poised to reshape the semiconductor landscape, with “agentic AI” expected to expand chip demand well beyond the graphics processors that have dominated the current boom.
According to Morgan Stanley, increasingly autonomous AI systems, capable of planning and executing multi-step tasks will drive a shift in computing demand toward CPUs and memory, fundamentally changing how data centres are built.
From Generative AI to “Agentic AI”
The AI boom so far has been heavily centred on GPUs, led by firms like Nvidia, which power large language models and generative AI systems.
However, Morgan Stanley highlights a transition:
- From AI that generates content
- To AI that acts autonomously and executes workflows
This new category, known as agentic AI, requires systems that can coordinate tasks, manage processes, and interact with multiple tools, rather than simply respond to prompts.
CPUs and Memory Take Centre Stage
As AI becomes more autonomous, the computing bottleneck is shifting.
Morgan Stanley notes that:
- CPUs are becoming the “control layer” for AI systems managing multi-step processes
- Memory demand is rising sharply as AI systems retain more context during operations
- Overall compute intensity is increasing across general-purpose infrastructure
The bank estimates that agentic AI could add US$32.5 billion to US$60 billion in incremental demand to the global data-centre CPU market by 2030, a market already projected to exceed US$100 billion.
The AI Trade Is Broadening
This shift signals a major evolution in the AI investment narrative.
Instead of being concentrated in GPU leaders, spending is expected to spread across the semiconductor value chain, including:
- Compute players: Advanced Micro Devices, Intel, Arm Holdings
- Memory leaders: Micron Technology, Samsung Electronics, SK hynix
- Manufacturing and equipment: TSMC, ASML
The implication: the AI boom is no longer a single-company story, it is becoming an ecosystem-wide growth cycle.
Data Centres Enter a New Phase
The rise of agentic AI is also expected to reshape data centre architecture.
Future infrastructure will likely:
- Balance GPUs with CPUs and memory
- Prioritise coordination and orchestration capabilities
- Increase overall compute density and energy demand
This marks a shift from GPU-heavy clusters to more distributed and integrated computing systems.
The Ledger Asia Insights
1. The AI Trade Is Expanding Beyond Nvidia
While GPUs remain critical, growth is spreading to CPUs, memory, and chip equipment, broadening investment opportunities.
2. Agentic AI Changes Infrastructure Economics
Autonomous AI systems require more persistent compute and memory, increasing total spending per workload.
3. Supply Chain Bottlenecks Could Create Winners
Companies in constrained segments especially memory and advanced manufacturing, may gain pricing power.
4. Asia Is Deeply Positioned in the Value Chain
With leaders like TSMC, Samsung, and SK hynix, Asia stands to benefit significantly from this broader semiconductor cycle.
A New Phase in the AI Boom
Morgan Stanley’s outlook signals a turning point:
AI is evolving from a compute-intensive technology into a coordination-intensive system.
As this transition unfolds, the semiconductor industry is set to move from a GPU-led boom to a multi-layered expansion across the entire chip ecosystem.
For investors, the message is clear: the next phase of AI will not just be bigger, it will be broader.
