Top
We spoke last quarter about the sharp reversal in market direction and leadership following the market trough post the “Liberation Day” tariff announcements in early-April. The third quarter saw a continuation of that upward trajectory, with global equity markets returning +7%, taking year-to-date MSCI World Index performance to an impressive +17%, despite numerous remaining uncertainties in terms of policy and geopolitics.
The MSCI World Index is now on over 20x forward earnings, with the S&P 500 Index at 23x. These extended multiples are on forward earnings that are meant to grow by double-digits for the next two years on the back of margins improving even further from record highs. Indeed, when we consider what is priced into today’s historically high market valuations, the market is betting on a continuation of the vigorous AI boom and a macro backdrop strong enough to deliver the double-digit earnings growth, with confidence that easing policy and AI-linked productivity will keep margins elevated. In short, expectations are high. Yet the record gold price reminds us that uncertainties linger.
This 6-month growth-tilted cyclical rally leaves us with unprecedented underperformance of quality against the broader index, as demonstrated by the performance of the S&P 500 Quality vs S&P 500 Index. We have only ever seen quality underperform to this degree in the run up to the TMT bubble burst. In the past, periods in which quality has significantly underperformed have frequently been followed by a prolonged period of meaningful outperformance of quality vs the broader market.
We see a tug of war within markets, between the bull argument that AI will be visibly transformational to corporate profitability in the near term and/or the U.S. economy sharply accelerates, and the bear argument where these high expectations are not met. We believe the bear scenario may come from the scaled enterprise adoption of GenAI taking longer than expected, raising anxieties about the return on the hyperscalers’ massive investments or the macro environment not being strong enough to justify the double-digit earnings growth expectations. Our long-tenured team is also acutely aware of how painful it can be when elevated expectations reset downwards.
Taking data from the last 150 years, the market appears to be in its fourth “New Tech” era, with the associated extreme valuation, and the S&P 500 CAPE1 over two standard deviations above trend. Comparisons to the three previous episodes of extreme valuation, in the 1900s, the 1920s and most recently the dot-com bubble, highlight the risk of significant overall market drawdowns when market sentiment shifts (anything from a 15% to 50% drawdown). The most exposed areas suffer more heavily on the way down, while underappreciated segments get their turn in the sun; consumer staples in the dot-com crash, and potentially the supposed “AI victims” this time, be they in software or in data-rich financials and industrials.
While there are similarities to the over-exuberance seen during the internet “New Tech” era, we do see notable differences today: the companies at the centre of the boom are earning real money and their earnings momentum remains strong, while their current price-to-earnings ratios, though high, are not remotely extreme compared with 1999. Another critical difference is that today’s massive hyperscaler capital expenditure is largely being self-funded from operational cash flow, allowing for continued and even expanded investment with limited dependence on external funding.
However, uncertainty remains. There is a paradox currently at the core of the GenAI boom. It has garnered an unprecedented mindshare among C-suites for a new technology and the potential is clear to anyone who has used it, but the scale adoption and value realisation among corporates has been very limited. This could drive a classic Gartner Hype Cycle, with a shift from the period of “Inflated Expectations” to the “Trough of Disillusionment” as implementation proves hard and drawn out, even if it is eventually successful and transformative. In addition, the macroeconomic position is unclear given the high levels of policy uncertainty, not least around the eventual effect of tariffs, and worldwide geopolitical risks. It is worth remembering that while growth is positive, the macroeconomic outlook remains modest, with U.S. growth expected around 1.5%-2% for both 2025 and 2026 and EAFE markets closer to 1%.
During the quarter, markets became increasingly preoccupied with the question of whether AI will disrupt, in particular, data-centred businesses. The initial reaction has been quite broad-based, with investors indiscriminately punishing nearly all companies perceived to have exposure to data regardless of differences in business models, competitive positioning or adaptability. We believe this blanket approach by the market is wrong as it ignores important differences between the industries and companies involved. We carefully examine both the potential vulnerability to AI disruption and the revenue and cost opportunities on a case-by-case basis.
There are some general principles behind our company-specific analysis. In our view, those data-rich businesses that avoid disruption are likely to control proprietary datasets that cannot be imitated by GenAI bots scraping the internet, or are likely to be deeply embedded into clients’ workflows or even core to whole ecosystems. On the positive side, they should have the financial and technical capacity to integrate AI into their offerings in a way that enhances client value and also utilise the technology to remove significant costs, be it in client relations or coding. In the case of RELX, held in our portfolios, we are already seeing GenAI technology combined with its proprietary data sets accelerating revenue growth in its legal division. SAP, another holding, is buffered from disruption by being deeply embedded into mission-critical operations and by its well-established domain and industry expertise. Its Joule copilots and agents are potential sources of extra revenue, while GenAI innovation could speed clients’ lucrative transition to S/4 Hana, its next-generation enterprise resource planning system. It is precisely these sort of high quality, data-rich businesses we seek to own in our portfolio.
As the debate matures and the market develops a clearer view of which companies are truly vulnerable to disruption and which can harness AI as a competitive advantage, we expect to see much greater dispersion in returns across the sector. In the meantime, we see the broad-brush approach applied by the market as an opportunity to selectively upgrade some of our holdings in which some uncertainty exists on the impact of AI – where compounding babies have been chucked out with the disrupted bathwater.
In a market where investor “certainty” meets a very uncertain reality and valuations are stretched, we remain focused on companies we believe offer credible earnings per share growth, driven by strong revenue growth, which we consider a more reliable source of long-term compounding than supposed margin improvement.
Source for data cited, unless otherwise stated: MSIM, FactSet, as of September 30, 2025.
Featured Insights
