The starting point is the growing demand for computing power and data storage. As AI systems become more advanced, they need far more processing power and memory to handle complex tasks. This is driving strong demand for computer chips and the equipment used to manufacture them.
Importantly, this surge in demand is happening at a time when the semiconductor industry is already operating at close to full capacity. That combination of strong demand and limited supply supports profits and gives companies the confidence to invest in new factories and equipment, helping to fuel growth across the wider AI supply chain.
Yes, very much so. One clear example is energy. The rapid expansion of data centres is increasing electricity demand in certain regions, creating opportunities for utilities that supply this power.
AI is also likely to improve productivity across many parts of the economy. In more tightly regulated industries, such as banking, established companies may be able to benefit from these efficiency gains while facing less threat from new competitors. That’s why we see AI not as a narrow technology trend, but as a long‑term theme with the potential to influence a wide range of sectors.
The concern is understandable, particularly given comparisons with the dot‑com boom of the early 2000s. The difference today is who is doing the spending. Most of the investment in AI is coming from highly profitable, cash‑generative companies, particularly large US technology groups, which are reinvesting their own profits rather than relying heavily on borrowing.
History shows that major technologies often require large upfront investment before their full potential is realised. With AI, however, the foundations are already in place. Smartphones, cloud computing and high‑speed mobile networks mean new AI tools can be used immediately, rather than waiting years for demand to catch up. As a result, AI data centres are already fully utilised and generating revenues, which gives us confidence that current levels of investment are largely rational and commercially driven.
Some newer AI companies are not yet profitable and rely on external funding. In a few cases, that funding has come from larger technology companies that are also customers or partners. While this can raise concerns, there are important differences from past technology cycles.
Rather than lending money to customers to buy equipment, we are seeing established companies take minority ownership stakes in emerging AI businesses. This approach has clear precedents in previous technology waves and allows larger firms to support innovation without taking on excessive financial risk. While this is something we continue to monitor closely, we believe the current structure of AI investment remains more disciplined than in past bubbles.
