As artificial intelligence (AI) evolves, a seismic shift is reshaping the semiconductor landscape. Morgan Stanley’s recent report reveals that the next phase of AI, termed “agentic AI,” is moving economic value from graphics processing units (GPUs) to central processing units (CPUs) and memory. This transition is projected to create an incremental total addressable market (TAM) of up to $60 billion for CPUs by 2030.

Agentic AI signifies a departure from single-task generation to autonomous, multi-step systems. This evolution necessitates a more complex infrastructure, with CPUs and memory emerging as critical components. Morgan Stanley emphasizes that as AI systems become more sophisticated, the demand for CPUs will increase significantly, marking a structural change in how AI workloads are processed.

Transitioning from GPUs to CPUs: A Paradigm Shift

For years, GPUs have dominated the AI landscape, primarily due to their ability to handle massive parallel processing tasks. However, as AI applications evolve, the reliance on GPUs alone is no longer sufficient. The deployment of agentic AI requires a different computational profile, one that heavily relies on the versatile capabilities of CPUs.

According to Morgan Stanley, the transition to agentic AI will fundamentally reconfigure data centers. While GPUs remain essential for training large language models, the orchestration of autonomous agents necessitates a shift towards CPU-centric designs. This transition will not only enhance processing capabilities but also improve latency and efficiency in AI operations.

Morgan Stanley estimates that CPU-side orchestration could account for 50-90% of total workload latency in agentic systems.

As the demand for agentic AI grows, so does the need for robust CPU infrastructure. Morgan Stanley estimates that CPU-side orchestration could account for 50-90% of total workload latency in agentic systems. This highlights the importance of CPUs in ensuring smooth and efficient AI operations. The report further elaborates that as companies increasingly deploy agents capable of planning, reasoning, and interacting with external tools, the demand for low-latency interconnects and high-bandwidth memory will become paramount.

You may also like

Education & University Insights

Women-Led Venture Funds Transforming Startup Ecosystems

Women-led venture funds are reshaping the startup landscape, enhancing capital access and diversity in entrepreneurship. Learn how these funds are changing the game.

Read More →

Wider Implications for the Semiconductor Industry

The implications of the shift towards CPUs and memory are profound for the semiconductor industry. As the demand for agentic AI increases, the market for CPUs is expected to expand significantly. Morgan Stanley predicts that orchestration CPUs could carve out a data center market worth between $82.5 billion and $110 billion by 2030.

Moreover, the report notes that agentic workloads could drive an additional demand for 15-45 exabytes of DRAM by 2030, which is equivalent to 26-77% of the expected annual DRAM supply in 2027. This surge in demand for memory highlights the critical role that memory will play in supporting the next generation of AI systems. The need for advanced memory solutions will be crucial, as traditional memory architectures may struggle to keep pace with the requirements of agentic AI.

As the semiconductor landscape evolves, companies that manufacture CPUs, memory, and related components are likely to see significant growth. The report suggests that supply-constrained enablers, such as foundries and advanced packaging providers, will capture outsized economics as the complexity of AI systems increases. This shift also indicates a rebalancing within the semiconductor industry. While companies like Nvidia will continue to play a vital role in the AI ecosystem, the incremental revenue pool is shifting towards CPU manufacturers and memory suppliers. This diversification of the AI investment landscape presents new opportunities for companies across the supply chain.

Skills, Hiring, and Opportunity Outlook

Looking ahead, the transition to agentic AI is still in its early stages, but the infrastructure buildout required will be more complex and capital-intensive than previous waves of AI development. If execution keeps pace, the next five years could see CPUs and memory rival GPUs as the primary drivers of AI-related semiconductor revenue growth.

Skills, Hiring, and Opportunity Outlook Looking ahead, the transition to agentic AI is still in its early stages, but the infrastructure buildout required will be more complex and capital-intensive than previous waves of AI development.

The shift towards agentic AI is not just a technological evolution; it represents a fundamental change in how businesses will operate in the future. Companies that can adapt to these changes and invest in the necessary infrastructure will be well-positioned to thrive in the new AI landscape. As enterprises deploy agents capable of planning, reasoning, and interacting with external tools, the demand for low-latency interconnects and high-bandwidth memory will become paramount. This creates a pressing need for innovation and investment in these areas.

You may also like

Career Guidance

India’s New Surrogacy Bill Leaves Parents in Limbo

India’s new altruistic-only surrogacy law has halted a once-affordable avenue for many hopeful parents, pushing them toward costly overseas options and sparking a debate over…

Read More →

The question remains: how will companies navigate this transition, and who will emerge as the leaders in this new AI-driven economy? The next few years will be crucial in determining the future dynamics of the semiconductor market and the broader implications for the tech industry. Understanding these shifts is essential for stakeholders looking to capitalize on the evolving landscape of AI technology.