Amundi AI Outlook
3 steps1 pieces1 Sources
The Amundi Investment Institute, the research arm of French asset manager Amundi S.A., has released its 2027 AI investment outlook, positing a core judgment: the market will cease paying for capacity and begin demanding proof of returns. The article divides the AI value chain into three layers—bottleneck holders (power equipment, electrical grids, networking equipment, chip fabrication), spending dependents (new cloud providers, data center developers, colocation REITs), and the overlooked facilitator layer (IT services and software firms that translate AI capabilities into enterprise workflows)—arguing that the facilitator layer faces lower expectation thresholds and has a nearer monetization inflection point.
This shift is driven by accumulating warning signs in the current cycle. As free cash flow thins and leverage rises among US hyperscalers, every additional dollar of investment will face greater scrutiny. Amundi notes that while earnings breadth is improving, the IT sector is expected to account for 70% of net income growth across the MSCI ACWI index in 2027, up from around 40% in 2026. However, EPS leadership may diverge from market leadership as valuations become more demanding. The risk is not an abrupt end to the cycle, but a widening gap between capital deployed and returns generated.
For investors, being selective in 2027 follows a simple rule: limit exposure to industries requiring further commitments. Spending-dependent names may face direct revenue impairment if spending plateaus, whereas real infrastructure plays are more insulated due to physical scarcity and technological disruption risks. Meanwhile, the 'facilitator' middle layer is structurally positioned to monetize earlier, with transition to consumption-based pricing and agentic AI adoption pointing to a closer inflection point than the infrastructure layer.
A key external factor is China's role outside most portfolios but inside the returns debate. Chinese AI names have small weight in global portfolios, but their influence is far from marginal. Open-weight models compress frontier pricing power, and efficiency gains achieved under compute scarcity question the capex intensity assumed by the US build-out. This challenges the returns thesis itself, not just its timing.
What to watch next includes evidence of the monetization curve bending upward, renewed commitments confirming the cycle has room, or any wavering on spending guidance which could trigger a sharp re-rating of the AI infrastructure complex.