Artificial intelligence (AI) has rapidly evolved, with significant investments made in training large-scale models. Now, the industry is entering a new and potentially much larger phase: AI inference.
The Shift from AI Training to Inference
Artificial intelligence (AI) has rapidly evolved, with significant investments made in training large-scale models. Now, the industry is entering a new and potentially much larger phase: AI inference. This phase is characterised by the widespread deployment and adoption of AI models across consumer and enterprise applications. While still in its early stages, inference is expected to unlock long-term potential as AI’s capabilities are scaling for everyday use by businesses and consumers globally - driving substantial demand for supporting infrastructure.
A Signal for Data Centre Growth
Recent developments across the technology sector have marked a critical turning point in the AI inference phase. The industry is witnessing a surge in demand for cloud infrastructure, with major long-term contracts being signed by leading technology companies. This trend signals the beginning of a massive build-out of data centres to support the growing needs of AI inference.
Technology industry leaders are emphasising that the inference phase represents a much larger opportunity than the initial AI training market, as organisations scale up AI through real-world deployment across both consumer and enterprise applications.
The Data Centre Opportunity: Scale and Financing
Morgan Stanley research indicates that global data centre capacity will need to grow six-fold by 2035 to meet the demands of cloud computing and AI. This translates to an estimated US$3 trillion investment in data centre infrastructure between 2025 and 2028. While much of this will be funded by the cash reserves of major technology companies, private and corporate credit markets will also play a crucial role in bridging the financing gap.
As AI adoption accelerates, more traditional workloads are expected to shift to the cloud, further fueling the need for expanded data centre capacity. Despite the scale of the investment required, the industry is well-positioned to meet these challenges, supported by strong cash flows and innovative financing solutions.
Storage and Memory: Meeting the Demands of AI
The rapid growth of AI is driving increased demand for data storage, particularly in the form of hard disk drives (HDDs) and solid-state drives (SSDs). Data is essentially the “oil” that powers AI, and the need for high-capacity, reliable storage solutions is greater than ever. Recent trends show supply shortages in HDDs, driven by the explosion of AI-generated content such as images and videos.
While HDDs remain the preferred solution for large-scale data storage due to their cost efficiency, SSDs offer lower latency and are expected to gain market share as technology advances. The underlying semiconductor technologies that enable these storage solutions are also experiencing rising demand, with memory components becoming increasingly important.
Semiconductors and Server Infrastructure: Broadening the AI Ecosystem
Semiconductors have been at the forefront of AI’s growth, particularly in the training phase. As the inference phase takes hold, the range of beneficiaries is expanding to include companies involved in memory, storage, and server infrastructure. The demand for AI servers is projected to rise significantly, with unique requirements for processing and managing vast amounts of data.
Server manufacturers are seeing strong growth in AI-related revenues, driven by orders from cloud service providers, large enterprises, and the expansion of AI workloads. This trend is expected to continue as AI becomes more deeply integrated into business operations and consumer applications.
Funding and Sustainability: The Road Ahead
The AI industry is entering a transformative phase, with inference and data infrastructure at the heart of future growth. The build-out of data centres, advancements in storage and memory, and the broadening of the semiconductor and server ecosystem all point to a dynamic and evolving landscape. As AI becomes increasingly embedded in everyday life, the opportunities, and challenges, will continue to grow.
The scale of investment required for AI infrastructure has raised questions about the sustainability of funding, especially as commitments become more interrelated across the technology ecosystem. However, with a significant portion of the required capital already backed by major technology companies and strong revenue generation prospects, the outlook remains positive.
The success of leading AI solutions will be a key factor in maintaining confidence in future funding, as will the continued support of credit markets and private investors.
