Disclaimer: Views in this blog do not promote, and are not directly connected to any Legal & General Investment Management (LGIM) product or service. Views are from a range of LGIM investment professionals and do not necessarily reflect the views of LGIM. For investment professionals only.
An unlikely AI case study?
Walmart* may not be glamourous, but the American retail giant is showing the way ahead for creating potential value with artificial intelligence (AI).
Another AI blog? Snooze.
Investor fatigue is understandable, given how this theme has dominated financial markets over the last two years. While we’re still positive on the long-term outlook, we would struggle to say we didn’t in some way feel it too.
When investors do engage, the focus has moved from technical scepticism to commercial scepticism and a concern over whether corporates will realise a meaningful return on their AI capex spend.
This leads me to the perhaps surprising example of US supermarket chain Walmart. I have highlighted some comments from their Q2 2025 earnings call (excuse the length, but I think they are deeply insightful for this debate) that showcase the ways they are utilising AI.
“We're finding tangible ways to leverage generative AI to improve the customer, member, and associate experience. … One example is that we've used generative AI to improve our product catalogue. The quality of the data in our catalogue affects nearly everything we do, from helping customers find and buy what they're looking for, to how we store inventory in the network, to delivering orders.
We've used multiple large language models to accurately create or improve over 850 million pieces of data in the catalogue. Without the use of generative AI, this work would have required nearly 100x the current headcount to complete in the same amount of time.
Customers and members are already enjoying AI-powered search on our app and site. And now they'll have even more help with a new shopping assistant that provides advice and ideas, answering questions like, which TV is best for watching sports? Looking ahead, the assistant will be able to respond with more specific follow-up questions like, how's the lighting in the room where you'll place the TV?”
The key point here is that no single AI application, be it customer bots that drive purchases, more efficient shelf stacking/inventory management and delivery, or devices for store and warehouse employees, is a standout driver in itself. It’s more the accumulation of AI applications being embedded across the entirety of the business that aggregate into something truly meaningful.
We think investors sceptical of the commercial logic of making significant investments in AI therefore need to look beyond individual line items in financial statements and at the broader impact on business operating conditions. There is a risk here that investors can’t see the business opportunity because the spreadsheet is getting in the way.
Let’s turn to the technology sector more broadly, specifically the hyperscalers at the centre of this debate. In our opinion, Microsoft exemplifies the above ‘problem’ perfectly, given that AI is so deeply embedded in everything that they offer. Yes, they can explicitly state OpenAI and generative AI’s contributions to Azure cloud growth, but the reality is that AI goes far beyond this. It’s just incredibly hard to disaggregate and quantify the isolated value of something that is so integrated across their product platform, and which ultimately results in a superior offering.
Yes, the magnitude of the dollar spend on AI capex is truly extraordinary, but so are the size of these technology companies. We therefore think it’s useful to scale this spend by looking at it from a relative intensity perspective (capex / revenues):
Examining this chart, Microsoft* clearly shows consistent execution of elevated capex intensity driving enhanced return on invested capital (ROIC). Put differently, spending has typically equalled value creation, at least if we look backwards.
We expect that, for most companies, the benefits derived from AI will be seen gradually overtime in overall performance metrics rather than a grand reveal ‘iPhone moment’. This quiet raising of standards and improving of business performance (or deepening of economic moats) will become apparent only to those most willing to embrace this new frontier – but does raise the risk of accusations of there being no material use case.
Either way, we think there will likely be plenty of disruption to come, given the nascent state of generative AI. It will be fascinating to see where we collectively end up – skilled stock-picking could pay off.
*For illustrative purposes only. Reference to a particular security is on a historic basis and does not mean that the security is currently held or will be held within an LGIM portfolio. The above information does not constitute a recommendation to buy or sell any security.