Over the month the fund retraced by 3.75% in GBP. The strongest markets were Saudi Arabia and Thailand; the weakest was China. The best performing sector was Industrials and Consumer Discretionary the weakest. Year to date the fund has returned 19.5% which is behind MSCI Emerging Markets.
Last month we mentioned that we saw a deteriorating risk/reward in our technology exposure given the price moves, very high volatility and diminishing yields. Whilst the index heavyweights remained firm, there’s been increasing signs that the ‘all-in’ AI trade may have begun to apex.
With elevated optimism and crowded positioning (c$40bil has flowed into memory and single stock ETFs this year), the key issue has always been when the narrative shifts from AI demand, to return on investment. The former is relatively easy to monitor through factors such as token consumption and the annualised revenue numbers coming from the leading labs, the latter is much harder to quantify as it requires both revenue and margin projection far into the future. This has come into sharper focus recently now that funding the capex super-cycle is increasingly coming from the credit markets rather than hyperscaler’s cashflows.
There’s also other over-arching issues such as power availability (more chips being produced than new power generation) and data sovereignty (not passing-on your firm’s IP to a frontier lab) which have led to a rapidly changing view of the moat held by leading AI labs. This is being compounded by the remarkable advances made by Chinese open-source models that achieve very similar results at a fraction of the cost.
Our equity exposure in technology is almost entirely centered on the picks and shovels of the AI arms race, and they are in a new demand paradigm that extends far beyond HBM memory. Ironically some of the tightest supply deficits we are hearing from management are more in the legacy areas of the supply stack rather than in the advanced technologies, but they are all ultimately dependant on a continuation of the waterfall of capex flowing from these US companies.
Many in the team are naturally sceptical that the current run rate of this capex is sustainable having had first-hand experience of the misallocation of capital in the dot-com era and the creative funding techniques of the GFC era, so we’ve been watching the credit markets closely. This has gained further relevance since the delay of OpenAI’s IPO into next year and the capital raisings by Alphabet and SpaceX, both of which are now below issue price.
With AI related hardware stocks accounting for around 40% of the EM index, and much more in terms of returns, exposure to the sector overwhelms any other active decisions in the portfolio. However, one of the core philosophies of our portfolio construction is not to have a single sector or theme dominate our portfolio, especially one with such high volatility, so we’ve had relatively low exposure and have recently been consolidating further as we shift towards alternative investment themes. We are not of the opinion that this is the end of the AI hardware trade, but there can certainly be a period of consolidation. This first half has set a new record in relative country returns where MSCI South Korea has more than doubled whilst China and India are both down around 10%. This spread of major market moves is more than double the nearest equivalent in H1 2016 (Brazil +46% vs China -5%). We are adding selectively to India and China where valuations and yields have now returned to very attractive levels while economic performance has been relatively resilient in the context of the Middle East conflict.
We have also increased exposure to some of the smaller markets such as Thailand, where new-found political confidence, clear economic policies, government stimulus measures and a revival of tourism have underpinned domestically-oriented sectors. We are also starting to add to Indonesia which has been the weakest market in EM this year, down around 40% and driven by the reciprocal of almost all the positives mentioned above. This has resulted in question marks over its MSCI index status prompting many to exit the market entirely, leaving valuations at levels not seen for over a decade.
