Highlights: Markets are debating between two opposing AI narratives: AI will disrupt incumbents and jobs or AI won’t deliver enough productivity to justify the investment. Our take sits in the middle — AI is proven to be useful and disruptive, but companies and workers aren’t defenceless. Recent volatility has reset expectations: IT valuations are lower and positioning is lighter, which can create entry points — but selectivity matters. If investors find it hard to pick winners, “picks-and-shovels” exposure in data centres and electricity providers can offer a steadier way to participate.

• Offence is the best defence
  “Agentic AI” could reshape enterprise software, but it’s unlikely to replace SaaS overnight. Large platforms still benefit from multi-year contracts, deep partnerships, proprietary data and IP. Many incumbents can defend and extend their position by embedding AI into products and workflows, innovating off their data advantage, and expanding their addressable markets. In practice, AI is already supporting productivity and margins through automation across functions and sectors
• Macro watch: labour market impact
  The macro debate has shifted from “will AI monetise?” to “what will AI do to labour markets?”. It remains too early to be definitive, but our base case is modest productivity gains with limited job creation in the near term
• Diversify, stay selective, and don’t ignore the infrastructure
  Selectivity matters as AI creates winners and losers. We favour combining exposures across public and private markets, and using hedge funds to broaden opportunity sets. If winner selection feels uncertain, “picks-and-shovels” exposure via data centres and electricity providers offers a more stable route as demand for compute, storage and power should persist across AI cycles