20th February 2026
The past few weeks have illustrated the benefits of the fund-of-funds approach. We cannot pretend to be masters of all of the subjects required to build a directly-invested multi-asset portfolio. Instead, we must be jacks-of-all-trades, allocating capital to specialist investors across both public and private markets. They may be regional specialists (India / UK etc), sub-asset class specialists (mortgage-backed securities / loans / logistics property / convertible bonds etc) or sector specialists (insurance or – the subject of this week’s blog: technology). The furore around the impact of AI on software businesses is a classic example of the benefits of outsourcing to people at the frontier of developments in AI and who have a far better knowledge of which companies will be impacted most. Here we attempt to summarise what we have learnt from our recent conversations with, and readings of the writings of these investors.
I must give particular credit to the Polar Capital Technology team, Barry Norris of Argonaut Capital, James Knoedler of Evenlode Investment, William de Gale of BlueBox and the teams at HG Capital and Oakley Capital.
First, addressing the huge capital expenditures by the so-called “hyper-scalers” (Microsoft, Google, Meta etc). These numbers have blown consensus estimates out of the water and naturally beg questions of whether there will be return on this investment as well as the secondary impacts on other industries like software.
Barry Norris reckons that the best estimates of AI Industry capex is c. $400-500bn a year. The beneficiaries of this spend are pretty clear (data centres, hardware and semiconductor equipment) – as are the industry bottlenecks (e.g. semiconductor memory). Speaking to Polar about this trend, the risks, if anything, are currently to the upside as capex spend keeps surprising to the upside.
The most direct way that the hyper-scalers can generate a return on this capex is via sales of AI software itself. The problem is, as Norris points out, that the hyperscalers are not deriving much revenue from this area at all – he estimates that the likes of Microsoft and Google are generating low single digit billions of dollars annually. Private companies like Open AI and Anthropic are doing better (and account for the majority of AI software revenues), but at the industry level, Norris estimate sales of AI software is running at c. $40-50bn. That’s a tenth of industry capex – a ludicrous ratio. Norris rightly questions the efficiency of this capital allocation.
On the other hand, there are different types of capex: defensive and offensive. Polar Capital believe the likes of Microsoft may be spending as a defensive measure (necessary to sustain their Office franchise for example). Hyperscalers have very conservatively-funded balance sheets and can add considerable amounts of debt (to fund further capex) without becoming too leveraged. If their capex spend is seen as defensive (i.e. protecting existing high margin businesses) then it need not have a high direct return. Meanwhile the spend of the private companies OpenAI and Anthropic is clearly offensive and disruptive. Polar have highlighted how these companies’ revenues are growing faster than expected and faster than the companies themselves have guided.
As a result of all this expenditure, AI models are improving at a very fast pace. So called “agentic” models can write code and complete tasks – thus allowing companies and individuals to circumnavigate the need to pay software companies for certain applications.
This leads us to by far the most significant investment implication of this spending is via the second order effects: the disruption of whole industries (software being the most obvious example). And as William de Gale puts it, it is impossible to prove a negative. Investors have priced in uncertainty with ruthless expediency. “Show me AI won’t disrupt your business model.”
We think software providers where the following conditions are met can continue to thrive:
- Highly skilled end-users
- Requirement for extreme levels of precision
- Very high cost of failure
- Ownership of an asset hard to replicate (e.g. a proprietary data set).
Application software stocks that don’t obviously possess any of the above have been hit hard. For companies with proprietary data sets, the jury is still out. Will AI providers have to pay for access? Or will the pace of improvement of AI models mean that some proprietary data sets will soon become redundant. In addition, many of these incumbent companies derive their revenue growth stories from cross-selling and tangential offerings and these may be disrupted by AI. That said, an argument can be made that the importance of brand, and the associated trust bound up in it, will mean continued success for these companies even if new breed AI models and companies can provide a service at a cheaper price.
But one area where we think there is the strongest argument for resistance to AI disruption is among those software providers whose products possess “deterministic” characteristics (in the words of Matthew Brockman of HG Capital, these applications enable “mission-critical processes that require accurate, predictable outcomes (not probabilistic responses”). Heavily regulated sectors or sectors where even a small error can lead to fatal consequences (medicine, engineering) are likely to continue to use applications provided by companies that have a deep specialism in one area. Indeed, incumbents are likely to be able to defend, if not build out, their moats by utilizing the power of AI to augment what they provide already.
Another bull case for incumbent software companies involves using AI to increase target addressable markets (TAMs): either via the ability to service “the tail” (lower value customers) or replacing human labour. HG Capital is a listed private equity investment company that specialises in owning software businesses with a combined enterprise value of $185bn. They have an inhouse AI product incubator to help their investee companies increase their TAMs. The implications for graduate jobs (already a horribly tough market) are not lost on us. But even here there is a bull case – a kind of Jevon’s Paradox in which the efficiency gains from AI leads to more not less employment. Willian de Gale has cited the radiology profession as a live example.
Finally, the way companies will pay for software will change. From subscription (“per seat”) pricing to outcome-based invoicing (tasks completed or labour time saved). Companies that are unable to adapt to these new economics will struggle.
Notice, I haven’t once used the word “valuation” – the north star of our investment process. When discussing these matters with fund managers, we are always looking for a margin of safety. We were already lowly-weighted to software companies, where multiples had expanded too far. But, as you will know from Dan’s recent blog, we have recently been shifting our UK equity exposure from value towards quality – including funds that hold some of the companies recently hit over concerns that their business models are being disrupted. We think there will be opportunities for discerning active managers who are on top of the pace of growth of AI models. But for now, we are intent on avoiding taking a strong view and will not be increasing exposure to those companies with question marks over their earnings 2 or 3 years’ from now, or indeed their terminal values. The last 1,000 words show you that arguments on either side of the debate are strong. Thankfully there are plenty of other areas of markets where margins of safety are far easier to assess, and where upside is not at risk from the disruptive forces of AI.
Ben Conway – Head of Fund Management
Subscribe to receive our latest commentary and insights straight to your inbox.
For professional advisers only. This article is issued by Hawksmoor Fund Managers which is a trading name of Hawksmoor Investment Management (“Hawksmoor”). Hawksmoor is authorised and regulated by the Financial Conduct Authority. Hawksmoor’s registered office is 2nd Floor Stratus House, Emperor Way, Exeter Business Park, Exeter, Devon EX1 3QS. Company Number: 6307442. This document does not constitute an offer or invitation to any person, nor should its content be interpreted as investment or tax advice for which you should consult your financial adviser and/or accountant. The information and opinions it contains have been compiled or arrived at from sources believed to be reliable at the time and are given in good faith, but no representation is made as to their accuracy, completeness or correctness. Any opinion expressed in this document, whether in general or both on the performance of individual securities and in a wider economic context, represents the views of Hawksmoor at the time of preparation and may be subject to change. Past performance is not a guide to future performance. The value of an investment and any income from it can fall as well as rise as a result of market and currency fluctuations. You may not get back the amount you originally invested. FPC26647
Source: MSCI. Neither MSCI nor any other party involved in or related to compiling, computing or creating the MSCI data makes any express or implied warranties or representations with respect to such data (or the results to be obtained by the use thereof), and all such parties hereby expressly disclaim all warranties of originality, accuracy, completeness, merchantability or fitness for a particular purpose with respect to any of such data. Without limiting any of the foregoing, in no event shall MSCI, any of its affiliates or any third party involved in or related to compiling, computing or creating the data have any liability for any direct, indirect, special, punitive, consequential or any other damages (including lost profits) even if notified of the possibility of such damages. No further distribution or dissemination of the MSCI data is permitted without MSCI’s express written consent.
