We are now at or close to all-time highs again in world equity markets less than three months after the drama of the 2 April tariff announcements. The rally from the 8 April lows has been historically fast and furious, and focused on the US tech names which had been lagging into the 2 April announcement. The explanation for all of this is fairly simple. US GDPi growth expectations for 2025 have been cut from about 2.2% to 1.4% and EPSii growth for the S&P 500 from about 13% to about 8% since the announcement. Downward earnings revisions have however entirely excluded the large cap tech names which are for now immune to macroeconomic trends and driven almost entirely by the idiosyncratic AI trade. We think part of their appeal in 2023-2024 was the relative scarcity of EPS growth in a market that was broadly weak outside technology. While this had started to broaden in 2025, the tariff jitters have put an end to that for the time being.
We continue to look very closely into the economics of AI. We are now more than two years into the ‘AI era’ and while much is still not understood, the contours of the landscape are getting sharper. The technology is very different to previous waves of digital disruption in that it has high variable costs, very high capital intensity, and therefore does not scale in the same attractive way as past superstar digital businesses like Microsoft, Alphabet, and Meta did. For context, when Facebook was at Open AI’s current scale of monthly average users, it was roughly 70% smaller by revenue but enjoyed a 25% free cash flowiii (FCF) margin, whereas OpenAI expects to have a -54% FCF margin this year despite massive subsidies from its partners and a reliance on non-cash expensing (xAI, meanwhile, expects a -2,640% FCF margin). The big names at the heart of the AI category burn cash at a historic rate, which naturally cranks up risk for investors. This business model looks more like the Uber or WeWork generation of VCiv-funded ‘unicorns’. Investors in AI companies are betting that the economics will look like Uber in the long run, but the model of continual price cuts makes it very hard to work out if there is any underlying pricing power or not. Providers of compute power to the AI names are also seeing their gross margins erode and capital intensity soar as the recent Microsoft and Oracle results revealed; the big tech earnings growth we cite above is significantly ahead of FCF growth. The only big winner so far has clearly been Nvidia whose graphic processing units (GPUs) are the default backbone on which the industry has organised itself. It is however significant that Nvidia’s marginal customer has shifted away from the cashed-up hyperscalers, to OpenAI’s new compute partners who are highly levered and much more exposed to capital markets cycles, like Oracle, CoreWeave, and Soft Bank.
So, in brief AI has thus far been a ‘speedrun’ of classic tech features like high growth and low initial profitability in a way that reflects market participants absorbing the lessons of the past. The big difference is that it has not scaled in the same way, with so far low incremental returns on capital despite consumer user bases of hundreds of millions of people. The other surprise relative to the expectations of 2023 has been the rapid penetration of consumer AI applications but the relatively disappointing enterprise uptake. As we all learn more about the underlying technology, this is not that surprising. The technology tends to be approximate and unreliable in a way that consumers can accept but which has been so far more prohibitive for businesses which carry strict liability and regulatory oversight. While there are hopes that new developments within AI like retrieval augmented generation (RAG), agentic models and reasoning models can address these issues, the very high cost of these innovations brings us back to the exorbitant capital demands of the industry. Recent deals including Salesforce’s purchase of Informatica and Meta’s purchase of half of Scale AI, both of which we think are intended to improve the quality of pre-training data and post-training fine-tuning of models, highlight the steady inflation of the AI bill while big revenue and profit payoffs seem frustratingly always just around the corner.
This takes us back to the everyday reality of our portfolio and what we are doing. As ever, in order to keep the portfolio compounding, we want to find companies which can sustain their competitive advantage for as long as possible; we set more store by a 20% return on capital for twenty years than on 50% for five years. We also want to find companies where the competitive advantage is ideally improving at the margin, reflected in better pricing power and gross margins drifting gently upwards to enable more reinvestment in the competitive advantage.
We take the view that, as one well-respected information services company said to us, ‘AI is a UX (user experience) not a product in itself’. The natural language interface enabled by large language models can feel like an actual conversation, which makes it easier for non-technical users to surface content buried in large data sets. This is naturally useful to companies which control big proprietary bodies of information which are not available to the wider public and which can be used to help businesses become more efficient. We are well invested in this sector through classic data companies like RELX, Verisk, Wolters Kluwer, S&P Global, LSEG, and Experian. There are many ways AI can improve the productivity of these companies’ clients. For instance, Verisk is already selling a tool which allows faster review and processing of insurance claims documents. RELX similarly is selling tools which speed up the bread-and-butter tasks of lawyers in big white-shoe law firms, allowing quicker ‘Shepardisation’ of legal cases to see if they are valid precedents or not. Both have the common goal of improving the daily productivity of already highly skilled professionals. We are also interested in companies which have large logs of transactional data from dealings which happen on their platforms, including Mastercard, CME, ICE, and Marsh McClennan, which again can use these tools to aggregate insights from the data and make them available to clients.
The common thread binding the companies above is that they are not building the infrastructure for AI because they do not have to. The proprietary data they control are not reproduceable with AI so they can afford to be price makers on making AI tiers of content available to clients; in other words, they can price for the value it creates today. In some cases, the price is already low enough that it can be folded into the existing product at no added expense, whereas for others it needs to be a new tier of pricing. While this may be boringly conservative, we are not eager to take existential bets with client capital.
We do carry some exposure to some companies which are building large AI capacity notably Alphabet, Amazon, and Microsoft. The first two have very large captive consumer pools, excellent or good-enough proprietary AI models (respectively), and their own proprietary semiconductor businesses which run the entirety of their owned and operated AI effort (in the case of Alphabet) or a large and growing chunk of it (Amazon). These ingredients reduce their exposure to the open money pit risk that concerns us in AI. While there are ongoing concerns about consumers moving their time from Google Search to ChatGPT, our work using large-scale panel data suggests that 1) these products are more complementary than substitutive, 2) AI on the whole will benefit Google Search by improving the user experience and drive better outcomes from advertisers, and 3) Alphabet as a company is broadly benefiting from AI across its attractive non-Search businesses like YouTube, Cloud Platform, Workspace, Waymo, and the Play/Android franchise. Our Microsoft position has declined somewhat reflecting our concerns on 1) its high dependence on OpenAI both for foundation models and for consumer exposure, 2) its lagging position in custom silicon, and 3) the risk of declining margins in its Azure cloud franchise. Historically Azure has had the highest margins in enterprise cloud as its core client base of older-school companies tend to buy many software solutions from Microsoft, whereas Amazon and Alphabet’s more sophisticated digital-native clients like Netflix and Spotify tend to be more a-la-carte in their purchasing. As AI has grown, Azure’s mix of lower-margin business providing undifferentiated compute to consumer applications has grown, and we don’t expect this to have the same attractive cross-sell opportunity of higher-margin enterprise solutions. We continue to like the company but acknowledge its path ahead could involve lower gross margins.
In short, we are determinedly empirical in our approach to AI. While there is a lot of lofty talk of superintelligence and the Singularity looming around the corner, we take the view that when and if we get the ‘wild abundance of intelligence and energy’ recently promised in his blog by the CEO of OpenAIv, everyone will be a winner. We focus on the scenario where this doesn’t happen and hence, we continue to worry about balance sheets, cashflow, and valuations. For now, we all remain in a world where money and intelligence are relatively scarce and valuable. To be clear, we are not sceptics on the value of AI in its many guises. However, if it is to be a revolutionary technology, we assume that it will benefit a broad range of enterprises and consumers. So far its real impact has been muted outside a small pool of picks-and-shovels names. We view this as unsustainable – either it will produce substantial and widely shared benefits for the wider economy, or it will be deprioritised and moved back to more niche use cases e.g. coding copilots, language translation, natural language search tools, etc.
We will continue to update you as our thinking evolves.
Chris, James, Cristina, Gurinder, and the Evenlode Team
27 June 2025
Evenlode has developed a Glossary to assist investors to better understand commonly used terms.
Please note, these views represent the opinions of the Evenlode Team as of 27 June 2025 and do not constitute investment advice. Where opinions are expressed, they are based on current market conditions, they may differ from those of other investment professionals and are subject to change without notice. This document is not intended as a recommendation to invest in any particular asset class, security, or strategy. The information provided is for illustrative purposes only and should not be relied upon as a recommendation to buy or sell securities. For full information on fund risks and costs and charges, please refer to the Key Investor Information Documents, Annual & Interim Reports and the Prospectus, which are available on the Evenlode Investment Management website
iGross Domestic Product - a key measure of the total value of all goods and services produced within a country
iiEarnings Per Share - a financial metric that shows how much profit a company makes for each share of its stock.
iiiFree Cash Flow (FCF) - A measure of how much cash a company can generate over and above normal operating expenses and capital expenditure. The more FCF a company has, the more it can allocate to dividend payments and growth opportunities.
ivVenture Capital.