10th May 2026
The rapid growth of artificial intelligence, data centres and advanced computing demand is beginning to affect the wider technology market far beyond gaming consoles. Governments, councils, banks, universities, hospitals and large companies are all facing rising costs for servers, networking equipment, storage systems and high-performance computer chips.
The pressure comes from several overlapping problems:
AI companies are buying enormous quantities of advanced chips
semiconductor manufacturing capacity remains constrained
energy and cooling costs for data centres are rising
supply chains remain vulnerable to geopolitical tensions
demand for cloud computing infrastructure is surging globally
Companies such as Nintendo and Sony have already warned investors about higher component and manufacturing costs affecting pricing and margins. AI demand is increasingly competing with consumer electronics, automotive systems and enterprise computing for the same advanced semiconductors.
The biggest pressure is concentrated around chips produced by NVIDIA, AMD and TSMC, which dominate the advanced AI processing market. Data centre operators are paying extremely high prices for GPUs and high-bandwidth memory, pushing up costs across the wider technology ecosystem.
For governments and councils, the problem is particularly difficult because many already face major budget pressures while also being pushed toward digital transformation.
Local authorities across the UK are increasingly delaying or stretching technology replacement programmes. Instead of replacing desktop computers every four or five years, many councils are now trying to extend equipment lifecycles to seven years or longer. Servers are also being retained longer than originally planned.
This creates several consequences:
slower systems and reduced productivity
greater cyber-security risks
rising maintenance costs
difficulties running modern software
increased dependence on cloud services
Many councils are moving workloads away from expensive in-house data centres toward large cloud providers such as Microsoft Azure
, Amazon Web Services
and Google Cloud
because those companies can spread hardware costs across massive global infrastructure.
However, cloud computing is not necessarily becoming cheaper. AI demand is now increasing prices for cloud capacity too, especially for high-performance computing and storage.
Banks and financial institutions face a different problem. Modern banking systems increasingly depend on AI for fraud detection, risk modelling, cybersecurity and customer service automation. Large banks therefore cannot simply stop investing in computing power.
Instead, many are:
consolidating data centres
reducing office IT spending elsewhere
delaying non-essential upgrades
shifting toward hybrid cloud systems
negotiating long-term supply agreements
investing in energy-efficient hardware
Some banks are also redesigning software to use computing resources more efficiently because electricity and processing costs are rising together.
Healthcare systems face especially severe pressures because medical imaging, diagnostics and records systems require increasingly powerful computing infrastructure. NHS organisations already struggling financially may therefore postpone upgrades unless systems become critical.
Universities are also under pressure. Research institutions increasingly need expensive AI-capable computing systems to remain competitive in science, engineering and climate research. But many universities simultaneously face falling international student income and rising pension costs.
As a result, universities are increasingly sharing computing facilities through regional partnerships rather than each institution buying its own high-end infrastructure.
Governments internationally are responding in several ways.
One response is trying to rebuild domestic semiconductor production. The United States, Europe and Japan are all subsidising chip manufacturing because they increasingly see semiconductors as strategic national infrastructure rather than just commercial products.
For example, CHIPS and Science Act in the US committed tens of billions of dollars to domestic chip production.
The UK government is smaller in scale but has also introduced a national semiconductor strategy aimed at protecting supply chains and supporting specialist chip sectors.
UK National Semiconductor Strategy
Another response is delaying large digital ambitions. Some councils and public bodies are quietly slowing “smart city” programmes, AI deployments and large IT transformations because the infrastructure costs are rising faster than expected.
Energy consumption is becoming another major factor. AI-focused data centres consume vast amounts of electricity and cooling water. Governments and corporations increasingly face a difficult balance between pursuing AI capability and meeting climate targets.
Some organisations are therefore prioritising “good enough” computing rather than constantly upgrading to the latest hardware. Older systems may remain operational longer unless there is a clear business case for replacement.
There is also growing concern about technological inequality. Large global technology firms can still afford the newest AI infrastructure, but smaller councils, universities, charities and local businesses may increasingly fall behind because they cannot compete financially for advanced computing resources.
This may gradually create a two-tier economy:
organisations with access to advanced AI infrastructure
organisations forced to rely on ageing systems and outsourced services
For ordinary consumers, the effect may increasingly appear through:
more expensive electronics
longer replacement cycles
rising subscription costs for digital services
slower public sector IT modernisation
reduced local services as councils divert money into technology upkeep
higher banking and insurance operating costs
Ironically, many organisations are investing in AI partly to reduce staffing costs and offset rising wage pressures. But the infrastructure required to run AI itself is becoming extraordinarily expensive.
That means governments, councils and businesses are entering a period where technology is no longer automatically becoming cheaper every year — a major change from the assumptions that shaped public and corporate budgeting for the last three decades.