- Software and data analytics stocks have been put through the mincer
- Worries over how Anthropic’s Claude Cowork tool could affect demand for their proprietary data and services persist
- But their share prices had already peaked, in some cases a year ago, weighed down by high valuations
- Lofty price tags and growth expectations left these names exposed to any sudden, unexpected loss of confidence
- Issue now may be when valuations reflect too much caution and too little optimism as investors start to differentiate between perceived AI ‘winners’ and ‘losers’
“Investors are still trying to work out exactly what AI and the launch of productivity tools such as Anthropic’s Cowork and OpenAI’s Frontier platforms may mean for the data, intellectual property and business models of analytics, software and publishing companies, but the debate over the risk of disruption neglects a further important issue, namely valuation,” says AJ Bell investment director Russ Mould.
“This tool flagged the dangerous combination of lofty price tags and equally optimistic growth expectations at the top and it could yet signal when excessive pessimism on both counts may mean there is a chance to reassess.
“The economist and writer John Kenneth Galbraith pinpointed the importance of this issue in his text, A Short History of Financial Euphoria, as he analysed prior periods of market convulsion.
“Galbraith argued: ‘The least important questions are the ones most emphasise. What triggered the crash? This is not very important, for it is in the nature of a speculative boom that almost anything can destabilise it. Any serious shock to confidence can cause sales by those speculators who have always hoped to get out before the final collapse but only after all probable gains from rising markets have been reaped.’
“This is not to suggest that high-quality stocks such as RELX, London Stock Exchange or Experian were in a ‘bubble,’ but it does, in the current circumstances, flag how expensive these stocks were, using one-year forward price/earnings (PE) ratios as a metric.
“Investors had latched on to these names for, amongst other things, their proprietary data, sticky customers and pricing power.
“Even here the danger was that by paying top-dollar for quality names, investors mistook solidity of a business model for safety, and rendered these stocks unsafe by paying multiples of earnings and cash flow which meant there was actually little margin of safety if anything ever went wrong, and not as much upside as expected to compensate, even if everything went right.
“Although the past week’s share price tumbles have attracted all of the headlines, many of the stocks affected by a bout of AI jitters peaked some time ago. This suggests that investors had already begun to struggle to find reasons to pay higher valuations, even before the AI narrative – whether it proves accurate or not – began to chip away at their confidence.
Source: LSEG Refinitiv data
“This week’s loss of confidence is even more clearly reflected in the derating they have suffered, as the multiples of earnings investors are prepared to pay have fallen sharply. Premium ratings have narrowed and, in some cases, even gone to a discount, so investors will now start to look at how those premiums (or discounts) compare to historic averages, and whether the current ratings mean more bad news is now priced in than good.
Source: Marketscreener, consensus analysts' forecasts, LSEG Refinitiv data
“The onus is now on the companies to prove that their business models, profits and cash flows will not be disrupted by AI but enhanced by the adoption of it. It will take time to offer this reassurance, but the lower the multiples paid, the more time there is on investors’ side, at least unless business models are indeed about to be broken.
“The hard part then, if these companies do reaffirm their reliable growth credentials, will be the challenge outlined by long-time US equity market analyst Walter Deemer, who once observed, ‘When the time comes to buy, you won’t want to.’”
