Market volatility: what it is, what it isn’t, and what we can do about it
Highlights: Earnings have been strong, but markets are instead focused on cash flows – especially the mismatch between current AI investment and uncertain future earnings. In a high-for-longer environment, this is understandable, but it doesn’t signal broad-based credit stress. And AI’s rapid adoption means that compute demand continues to exceed supply, supporting margins and EPS. It is hard to know when markets bottom, but the 6-12 month outlook remains positive.
What it isn’t about:
- The sell-off is not driven by EPS: Q3 earnings have been strong, especially in tech, but also other in sectors. Any negative comments from the megacaps were not about the cloud or semi demand but about their older advertising and retail businesses
- The market volatility does not signal a cyclical downturn. PMIs are strong and economic surprises are positive. While the economy is K-shaped, there is a lot of support from structural trends, which is where we continue to focus
- There is no broad-based credit stress: CCC bond spreads have widened significantly, but BB and IG much less. We focus on quality in both public and private markets
What it is about:
- Worries about the mismatch between the huge current AI investments and the future revenues, leading to a negative current cash flow impact. This is natural in any tech innovation wave and less damaging for hyperscalers than the neo-cloud
- A realisation that AI constitutes a big part of people’s portfolios, leading to rotation into other sectors: as a result, the MSCI AC world is ‘only’ down 4 per cent from its peak
- Some profit taking and de-risking ahead of the mid-terms. We will soon know more
We remain optimistic for 2027 as AI adoption will continue to accelerate and users are willing to pay more than the current meagre average of USD16/month. This leads to a rapidly growing profit pie, but how will it be distributed? Adopters / applications are well placed, followed by the cloud and datacentres. A bottle neck in semis should continue to benefit their margins, but they are prone to headline risks if the data centre build-out is delayed.
The key is to look across the AI ecosystem in order to diversify, and to focus on quality (strong balance sheets and strong innovators) to achieve appropriate differentiation.