Barclays Research today released the 70th edition of the Equity Gilt Study, a flagship annual publication that combines market-leading macro analysis with a unique multi-asset data set spanning over 100 years. In this year’s report, our Research analysts examine some of the broader, long-term effects of the spread of AI – from the energy needs of the hyperscalers to the redefinition of white-collar work.

• Minds and minerals. Just a handful of countries tend to dominate supplies of inputs critical in the development of AI, such as rare earths and skilled talent. Such concentration is not always risky, in itself. But geopolitical conflicts, trade wars and natural disasters can quickly turn dependencies into vulnerabilities.

• Your job is probably safe (for now). Eight years of evidence suggests that the adoption of AI is not resulting in fewer employment opportunities for human beings. Instead, it is redefining expectations of the skills and tasks required, especially in higher-paid, white-collar roles. But these are early days. More radical macro effects could become apparent over time.

• Deep impact. The sudden emergence of DeepSeek from China was a “Sputnik moment” for the US and other developed markets, galvanising a new wave of investment in AI infrastructure. We think this capex arms race has much further to run – not least because of the vast computing power needed for emerging, multi-step workloads.

• From macro to markets. The spread of AI could have far-reaching implications for investment returns if it bears similarities to previous periods of rapid technological progress, such as the post-WW2 revolution in durable goods, the rise of the internet in the 1990s and the data boom of the 2010s. In that case, we might expect larger gains in equities, upward pressure on yields across the curve and a moderately stronger dollar.