Artificial Intelligence (AI) As the Semiconductor Growth Engine: AI White Paper Volume 1 – Memory & Computing

This market research report was originally published at the Yole Group’s website. It is reprinted here with the permission of the Yole Group.

Yole Group Launches its First AI White Paper.

KEY TAKEAWAYS

  • Compute at the limit: Training and inference workloads now demand gigawatts of electricity, raising urgent questions on sustainability, cost, and infrastructure capacity.

  • NVIDIA at the core: with more than 90% GPU market share, NVIDIA remains the dominant player but faces rising competition from AI ASICs, hyperscalers, and geopolitical headwinds.

  • Memory market rebound driven by AI and HPC, Combined DRAM and NAND revenue reached $170 billion in 2024, with HBM revenue set to double to $34 billion in 2025.

  • Storage evolution: enterprise SSDs and all-flash arrays are critical to sustaining AI training efficiency, ensuring GPUs remain fully utilized.

  • Long-term impact: by 2030, AI processor shipments will reach 16.8 billion units, representing a $300 billion market.

Yole Group today announces the release of its first Artificial Intelligence White Paper, an in-depth analysis of the memory & computing technologies, markets, and strategies driving the AI revolution. This volume of a new series dedicated to AI explores how generative, agentic, and physical AI are transforming the semiconductor industry, from data centers to edge devices. This White Paper is part of the 2025 White Paper Collection powered by Yole Group.

AI redefining the semiconductor roadmap

The emergence of generative AI in 2022 with the launch of OpenAI’s ChatGPT marked a pivotal moment in the development of the technology and a new source of semiconductor demand, relying as it does on data center-heavy computing.

With new applications and use cases emerging every month, data centers are rushing to scale up capacity. However, despite the increase in performance from the latest generations of processors from the likes of market leader NVIDIA, Yole Group expects the industry will continue to need more computing power.

This is set to strongly drive semiconductor revenue higher, as the requirements are growing faster than the increase in technical capabilities.

Memory and storage requirements are also expanding rapidly, as training AI models on vast datasets is driving a surge in capital expenditure. These trends are expected to drive continuous market growth throughout the next five years.

As memory technologies encounter technical constraints in providing the bandwidth and capacity improvements necessary to cope with demanding workloads, particularly from data center applications, new solutions are needed for long-term scaling, particularly in the DRAM business.

2025 marks a turning point: AI is no longer just a workload. It is the growth engine for the semiconductor industry. With massive investments in generative AI exceeding $200 billion in 2024, the pressure on computing, memory, packaging, and storage has never been greater. Yole Group’s white paper provides clarity on where value is shifting across the ecosystem and what technologies will define the next era.

Yole Group’s AI White Paper, Vol.1: Memory & Computing, delivers a comprehensive and independent analysis of the technologies and market forces redefining the future of AI infrastructure. Centered on computing and memory, it reflects the group’s deep expertise in semiconductor innovation, strategy, and supply chain dynamics. It provides critical insights for industry leaders navigating one of the most transformative shifts in the history of technology.

Yole Group’s insights

“The longer the machine thinks, the smarter it gets – at a cost. The challenge is how to sustain this exponential compute demand without breaking energy and capital limits.”
Adrien Sanchez
Senior Technology and Market Analyst, Computing, Yole Group

“Semiconductors are the backbone behind AI, transforming algorithms into real-world impact. But tomorrow’s breakthroughs will come from memory bandwidth, packaging, and system-level integration, not logic scaling alone.”
Ying-Wu Liu
Senior Technology and Cost Analyst, Computing, Yole Group

“With no true DRAM alternative, the industry must innovate within its limits, scaling the fastest, most cost-effective memory we have, in lockstep with AI giants like NVIDIA.”
John Lorenz
Director, Memory and Computing, Yole Group

This publication is part of the Yole Group 2025 White Papers Collection, which will soon feature additional volumes, including semiconductors in power electronics and photonics. This full white paper is available now. Download your copy of AI White Paper Vol.1 – Memory & Computing Focus to explore Yole Group’s exclusive insights, data, and forecasts.

Stay tuned on yolegroup.com!

Here you’ll find a wealth of practical technical insights and expert advice to help you bring AI and visual intelligence into your products without flying blind.

Contact

Address

Berkeley Design Technology, Inc.
PO Box #4446
Walnut Creek, CA 94596

Phone
Phone: +1 (925) 954-1411
Scroll to Top