17th October 2026
Artificial intelligence is attracting billions of pounds in investment, transforming expectations for business and raising hopes of a new wave of productivity. But as spending on computer chips, data centres and AI services continues to grow, investors are asking an increasingly important question: will the profits eventually justify the enormous cost?
The question became more pressing this week as technology shares came under renewed pressure on international stock markets. On Friday, 16 October, US technology stocks fell sharply amid concerns that companies building the infrastructure for artificial intelligence may struggle to generate sufficient profits to repay the money they have borrowed. Shares in data-centre operator CoreWeave fell by around 12% that day, while its bonds also came under pressure.
The concern is not that artificial intelligence has no commercial value. Businesses are already using it to assist with customer enquiries, software development, research, document processing and a growing range of routine tasks. The difficulty is establishing whether the financial benefits will arrive quickly enough to justify the enormous investment now being made.
The cost of building the AI economy
The AI boom requires much more than clever software. It depends on specialised computer chips, enormous data centres, high-capacity networks, reliable electricity supplies and expensive cooling systems. These facilities require substantial investment before they can generate revenue.
For some of the biggest technology companies, that spending can be financed from existing profits. Other businesses rely more heavily on borrowing and outside investors. That creates a potential problem if the demand for computing power grows more slowly than expected or customers are unwilling to pay enough to cover the cost of providing it.
Reuters reported this week that market uncertainty and rising interest rates have prompted some companies involved in data-centre investment to delay or reconsider fundraising plans. Investors are becoming more selective, favouring businesses with established customers and long-term contracts over projects whose future earnings remain uncertain.
The risk is that investment decisions based on optimistic forecasts become difficult to justify if the expected revenue fails to materialise. A data centre may be technically impressive, but its owners still need customers willing to pay for its capacity over many years.
British businesses see opportunities, but measuring returns remains difficult
The investment boom is not confined to Silicon Valley. British companies are also increasing their spending on artificial intelligence, hoping to improve efficiency, reduce administrative costs and develop new products and services.
Research published by SAP in partnership with Oxford Economics on 14 October estimated that the average UK business was spending £20.6 million on AI in 2026, up from £16.2 million in the previous year's research. The survey also found that businesses expected average returns on their AI investment to rise from 22% this year to 39% within two years. These are survey estimates and expectations, not guaranteed returns or figures that can be applied to every company.
A separate survey of 100 UK chief executives by EY-Parthenon found that half identified AI tools as one of the leading contributors to productivity gains over the previous year. Yet only 14% said they had comprehensive, real-time visibility of their AI costs and returns, while nearly a third reported significant gaps in their ability to measure the benefits.
That is an important distinction. A business may find that staff can complete work more quickly, but the financial benefit depends on what happens next. Does the company produce more with the same workforce? Can employees spend more time serving customers? Are errors reduced? Does the business gain new customers or simply perform its existing work a little faster?
Without answers to these questions, it is difficult to know whether AI investment is delivering enough value to justify its cost.
Smaller companies may benefit without building enormous data centres
For Scotland's smaller businesses, the most useful opportunities may lie in applying AI rather than developing the technology itself.
A small manufacturer could use AI-assisted systems to help plan production, monitor equipment or prepare quotations. A tourism business might use it to answer routine customer questions, translate information or manage bookings. Accountants, solicitors and other professional firms could use AI to help search documents, prepare drafts and organise information, provided appropriate checks protect accuracy and confidentiality.
These applications do not necessarily require major capital investment. Businesses can begin with limited projects, assess the results and expand only when there is evidence that the technology improves performance.
There are risks, however. Subscription fees can accumulate, integration with existing systems can be expensive, and staff need training. AI-generated information can also be wrong, misleading or inappropriate for a particular customer. Businesses must consider data protection, cybersecurity and the consequences of relying on automated decisions.
For smaller firms, the sensible approach is likely to be to identify a genuine business problem first and then decide whether AI offers a cost-effective solution.
What does this mean for Scotland?
Scotland has an opportunity to benefit from both sides of the AI economy. Data centres and associated infrastructure can bring construction work, engineering contracts and investment, while businesses across the country can use AI to improve their own productivity.
But large infrastructure projects also require substantial amounts of electricity, suitable grid connections, land, water or cooling arrangements, and skilled workers. If those resources are unavailable or expensive, projects may be delayed or become less attractive financially.
There is also a question of how much economic value remains in the local area. A large data centre can represent a considerable investment without necessarily employing a large permanent workforce once construction is complete. The wider benefits will depend on local supply chains, skills, business activity and the services developed around the infrastructure.
For the Highlands, the opportunity may therefore be broader than simply attracting a major data centre. Local firms could provide engineering, maintenance, cybersecurity, software and other specialist services, while smaller businesses could use AI to overcome some of the disadvantages of operating in remote areas.
However, those benefits will not appear automatically. They depend on access to reliable digital infrastructure, suitable training and practical support for businesses that want to adopt the technology.
Could the AI boom turn into a financial problem?
The Bank of England has already warned that the rapid expansion of AI investment creates potential financial-stability risks. Its July 2026 Financial Stability Report highlighted the growing use of borrowing and other external finance to fund data centres and related infrastructure. It also warned that a sharp reassessment of the earnings expected from AI companies could cause significant falls in their market valuations.
That does not mean an AI crash is inevitable. New technologies often require substantial investment before their benefits become widespread, and today's infrastructure could support valuable services for many years. A fall in share prices would not necessarily mean that AI itself had failed.
The danger is that financial markets may have become too optimistic about how quickly the technology will generate returns. If expected profits fall short, investors could lose money, borrowing could become more expensive and some projects might be cancelled. The effects would depend on how widely the losses were spread through banks, investment funds and other financial institutions.
For the wider economy, there is a difference between a technology being useful and every company investing in it being profitable. AI may ultimately transform business while some of the businesses financing its development fail to earn an adequate return.
The next stage must be about results
Artificial intelligence is unlikely to disappear simply because investors have become more cautious. Its potential to improve productivity, assist workers and create new services remains substantial.
But the next phase of the boom will require more than announcements of investment and ever-larger data centres. Companies will need to demonstrate that their spending produces measurable benefits, that customers are willing to pay for AI services and that the costs of electricity, computing power and finance can be covered by sustainable revenues.
For British businesses, the opportunity is to use AI where it solves real problems rather than adopting it simply because competitors are doing so. For investors, the challenge is to distinguish companies with viable business models from those relying on optimistic forecasts.
The technology may indeed deliver major economic gains. The question now is whether the financial returns will be large enough, and arrive quickly enough, to justify the extraordinary sums being committed to it.