The rising costs of computers, servers, and cloud-based services have not gone unnoticed by millions of users worldwide, and the reason for increasing prices for most technology products is the rapid growth of artificial intelligence (AI). The focus is not about the hype surrounding AI; rather, it is the infrastructure required to support AI’s growth.
Unlike traditional software programs, AI does not exist independently of hardware; therefore, when AI runs, it utilises a large amount of specialised hardware (CPUs, GPUs, RAM, and Storage). Many of the same components used in AI are also used to build professional workstations and servers, and cloud and corporate data centres are purchasing all of these components in large volume.
Hardware Shortage Issues
This large volume of purchasing by hyperscale companies has resulted in the diminished availability of components for other users. In addition, as demand for AI continues to grow, the companies utilising these components will experience increased prices and increased wait times. For example, many companies that do not operate with AI are now experiencing increased prices and wait times for standard IT equipment (computers, printers, etc.) because of the large quantities of GPUs that hyperscale companies are purchasing.
GPUs are the primary source of pressure associated with this issue. GPUs are essential for supporting the workloads associated with AI applications; however, they also serve other purposes, including supporting video editing, engineering, and accelerating server operations. The increasing adoption of AI means that GPUs are no longer niche components; they are vital assets in a company’s technology inventory.
There are similar sources of pressure on RAM and storage. AI workloads require vast amounts of data and thus create an increased demand for faster RAM, as well as larger capacity hard drives. Even if your business will never run AI workloads locally (due to resource limitations), manufacturers of memory and storage are adjusting their pricing models for all companies as RAM and storage become increasingly limited.
Cloud computing pricing will also continue to evolve based on AI workloads. Running workloads using AI requires significant amounts of computing resources, and cloud providers will not be able to absorb the costs of running AI workloads indefinitely. In many cases, the primary distinction between traditional data centre service models and cloud computing service models is the growing reliance on AI-based technologies for delivering IT services. As more companies and services embed AI functionality into their platforms, cloud pricing structures are shifting significantly. In some cases, cloud service pricing is rising considerably; in other situations, AI functionality is simply bundled into more expensive tiers or licensing options.
The Long-Term Changes
Unlike previous temporary spikes associated with supply chain challenges, The growth of AI is clear and therefore, has caused major change not only within technology but, also, in the way all products and services are designed to work. Therefore, businesses should expect to see the demand for AI continuing to increase and ultimately lead to a long-term negative impact on budgets.
As a result of AI’s increased use, the [IT] planning function will need to evolve to adapt to the new economic reality of the IT landscape. Delaying upgrades and/or budgeting based on historical prices will almost always result in lower quality products being purchased and subsequently being replaced much sooner due to software becoming more resource-intensive.
However, for most companies, there is no need to purchase AI-grade (high-end) hardware. All companies need is the ability to plan ahead by identifying the best equipment to purchase, understanding the lifecycle cost of acquiring that equipment, and avoiding making reactive cost-driven decisions when prices increase again.
AI’s transformative effects on how people work will be accompanied by the tremendous changes ongoing in the economic geography of technology. The cost of using technology will not simply be a result of market dynamics based on market timing, but will be permanently affected by AI.
