What if AI doesn’t work out?

AI has been a dominant investment theme in 2025, but legitimate questions exist around take-up and pay-off. Even if AI doesn’t deliver the sweeping technological revolution many investors hope for, other strong themes exist which can provide market leadership.

26 February 2026

2025’s equity market return for the US at least was in large part driven by earnings and expectations related to artificial intelligence (AI). While the S&P 500 Index gained 17.9%* over the year just gone, the Communication Services sector returned a whopping 33.6%* and the Information Technology sector an almost-as-eye-popping 24.0%*. The premise behind the moves will by now be familiar to everyone but it essentially states that we are on the cusp of a technological revolution that will automate nearly every aspect of our lives. The ‘picks and shovels’ suppliers of this revolution, whether it’s the large language models or the chips themselves are perhaps best personified by the ‘Magnificent Seven’ stocks – Alphabet, Apple, Amazon, Meta Platforms, Microsoft, Nvidia and Tesla. The scale of the investment they are making into AI is enormous, with conservative estimates for Alphabet’s spend alone standing at USD 175 billion. We believe, this is undoubtedly exciting on one level but some investors are concerned at the growing argument that AI may not be as revolutionary as previously hoped. This, combined with the fact that the S&P 500 is no longer ‘’cheap’’ on virtually any fundamental metric - 27x forward price/earnings* (P/E)¹ ratio as of 16 February 2026 - suggests that the risks of market progress stalling are no longer minimal. At the very least, the issue warrants further exploration.

Expensive as chips – S&P 500 valuation is undoubtedly elevated:

From 31 Dec 1997 to 16 Feb 2026

The case against the market enthusiasm for AI is multi-faceted, but probably rests on one major pillar – adoption, or lack thereof. As aggregated by The Economist, rare harmonised cross-country survey² of corporate officials in the US, UK, Germany and Australia by the Atlanta Federal Reserve (Fed), Bank of England, Bundesbank and Macquarie University respectively revealed that nearly 75% of businesses are now using AI in some way but at the same time 86% of respondents reported that there had been virtually no productivity improvement.

Mind the Adoption Gap

This in all likelihood speaks to the way that AI is being taken up (so far, at least), mostly by employees using it as an adjunct rather than firms themselves integrating it into their platforms ‘agentically’, ie in a way that satisfies customers’ requirements end to end rather than just improving single workflows. A rare example of full corporate AI adoption is South East Asian ride hailing firm Grab which recently undertook a deep rewiring of its core systems at enormous expense. Most western firms are balking at the investment required to truly integrate AI and therefore remain in ‘pilot mode’ for now, perhaps bolting on a smarter chatbot to existing systems to help customers but not truly running tasks autonomously. This lack of AI take-up can arguably be seen in the very positive employment data coming out of the US. Truly effective AI should be resulting in a major shift of the employment landscape but, since 2022, inflation-adjusted wages in the service sector, particularly accountants and salespeople, have increased rather than cratered as expected. Surely if customer service was being replaced by AI then employment in the service sector would be struggling? Other issues postponing AI investment decisions include the unreliable nature of some large language models (LLM). ‘Hallucinations’, in which AI insists factually incorrect output is accurate, can have dire consequences, and the ‘beta’ nature of the current technology may be putting corporations off. Clearly, anything less than widespread penetration of AI into all facets of human life, including of course business, is going to play poorly in a richly valued equity market which has been dominated by nothing else. The risk is that the billions of dollars going into AI data centres and powering ever faster models could possibly go the way of the mania for fibre optic cable in the late 1990s. While eventually used, the obsession with fibre infrastructure became totemic of the Dot.com bubble, and its popping in early 2000.

Smart glasses – seeing technology through a different lens

However, there are reasons the S&P 500 Communications and Tech sectors might prove more robust than investors think. First, they are nothing if not adept at constant innovation of their business models. To this point, AI is not the only story in town. There is, for example, a meaningful opportunity in the next couple of years to capture how consumers interact with the digital world which may involve elements of disruption and areas of potentially profitable business activities. For years now it has been assumed that the smartphone is the only way to access your digital life, but, spurred on by the incumbent Apple and Google duopoly, Meta and Amazon are now looking to muscle in with complementary technologies. Meta has been focused on smart glasses which can display messages and also augment reality via integrated cameras, while Amazon has been rolling out the Echo+ assistant to its smart speakers and (soon) its glasses and earphones too. The opportunity appears very substantial, in my opinion. By one estimate there are just 15 million users of smart glasses in the world while last year fully 250 million iPhones alone were sold. If these new technologies can successfully claim a fresh slice of consumers’ budgets then they could prove a neat offset in the event of disappointment with AI specifically.

AI is only one of many drivers of stock market returns

The other comfort that investors can draw on is that AI is not the only theme in the broader S&P 500 Index today. Sector leadership this year has broadened out to include, for example, consumer staples. They may be criticised as boring or stodgy but the sector has gained 15.8%* so far this year, powered by erstwhile stocks such as Walmart, Costco Wholesale, Procter & Gamble and Coca-Cola. In early February, Walmart’s market capitalisation hit a cool USD 1 trillion*. Materials, industrials and utilities have also fared well as investors have started to show an openness to steady earnings streams beyond technology. And of course there is literally a whole world of other regions beyond the US for investors to tap into, whether Germany with its rearmament and infrastructure story, Japan with its new government focused on growth or the emerging markets set to benefit from potential further US dollar weakness.

Boring is back – 2026 showing signs of more US market breadth:

From 31 Dec 2024 to 13 Feb 2026

We believe careful investors are right to question whether the momentum behind AI will be justified given the patchy record on adoption so far. With valuations stretched and Big Tech even going to the bond markets to partly finance this year’s capital expenditure (see Alphabet), the prospect of AI somehow not meeting expectations should at least be considered. The good news is that for investors who remain engaged in equities, even at the S&P 500 or global MSCI AC World index level, there are multiple sources of potential new market leadership which may automatically be captured. This could be from within the Technology sector itself, the broader sectors of the S&P 500 or in relatively undervalued international stock markets. Equities in this sense reflect a more diversified opportunity set than the recent apocalyptic commentary around AI suggests. Of course, this assumes that AI doesn’t work out - which itself is something nobody can be certain about.

Julian Howard is Chief Multi-Asset Investment Strategist at GAM Investments. This article represents the views of GAM’s Multi-Asset team.