Artificial intelligence is the story which has propelled markets this year. Where to from here? There’s no doubt AI innovators – and technology companies generally – are highly priced in relation to company earnings, by comparison with long-term averages. But is it an AI bubble?
Markets move in cycles, and investor enthusiasm for AI could continue for some time. Clearly AI technologies are here to stay, and will cause some fundamental changes in society. So there is also risk in sitting on the sidelines. How should investors react?
Valuation levels are high
Exhibit A is the valuations of pure AI firms – which the Economist said in June were “verging on the unhinged”. That article cited Perplexity’s valuation at $14 billion as an example of challenging valuations.
Things have only risen since then. Perplexity reportedly raised money at a $20 billion valuation in September, an estimated 100 times’ its annual sales.
ChatGPT’s owner OpenAI, meantime, reportedly raised money from Softbank earlier this month based on a valuation of $500 billion, a multiple of perhaps 50 times’ annual sales.
AI bubble or not, the technology is driving wider valuations
The AI excitement has also driven large movements in technology stocks more generally. In fact, AI businesses have performed strongly for investors since 2023.
And we’re nearly all invested in technology, because it has become such a large part of the market, in the US in particular. The Bank of England estimates that AI stocks now account for roughly 44% of the S&P 500 market capitalisation.
This brings some specific risks.
Firstly, the valuations of many of these players are ‘priced for perfection’: if they produce an earnings disappointment (eg Netflix’s 10% share price drop last week) or a challenge emerges (the arrival of Deepseek caused Nvidia’s value to fall nearly $600 billion in the first half of this year), then there’s a strong change of rapid revisions to their valuations.
Secondly, big tech has become increasingly interdependent, with some companies buying shares in their customers as well as selling them goods. For example, Nvidia received money from OpenAI by selling it chips – but it’s also injecting money into OpenAI as a shareholder. How long can this cycle continue?
Thirdly, on an operational level, we are seeing increasing interdependency. That was evident in the recent seize-up of Amazon Web Services systems, where numerous other software companies, from Zoom to Coinbase to Perplexity experienced problems in delivering their services.
The increasing linkage between these firms means the popping of an AI bubble would have larger knock-on effects for the whole market.
AI bubble warnings are growing
The voices expressing concern about AI bubble risks are growing – as are comparisons to the dot.com boom and bust. In recent weeks there have been three notable examples:
Jamie Dimon, CEO of JPMorgan and the world’s most powerful banker, told the BBC he was “far more worried about [a fall in US markets] than others”.
The Bank of England said “The risk of a sharp market correction has increased” in a report on 8 October on risks to financial stability.
The International Monetary Fund’s chief economist, Pierre-Olivier Gourinchas, has also drawn comparisons with the dot.com era in an interview with Reuters.
Safe might mean sorry
So should you exit now? The hard part is that valuations can stay stretched – and indeed grow – for a long period. AI bulls contend we are only in the foothills of finding out what the technology can do for growth and productivity, and that there is more to come.
Certainly tech valuations have continued to outpace the wider market despite the warning voices. Over the last month, for example, the main US index, the S&P 500 is up 2.2% in dollar terms. But the Nasdaq 100 index which focuses on technology stocks is up 3.5%.
So the boom could go on for some time – and there is money to be made along the way. That calls for a balanced approach.
Four actions to consider
Here are four actions to consider in order to balance your exposure to the risk of an AI bubble.
Diversify: it’s the only free lunch in investment. The idea is that a portfolio constructed of different kinds of investments will, on average, yield higher returns and pose a lower risk than any individual investment found within the portfolio. That’s the case more than ever when the overall stock market is dominated by highly-valued technology stocks.
Assess your total exposure: for example, if your job in a multinational tech company investing in Ireland, and your pension are also heavily invested in AI, you’re in an AI bubble in one sense already. You may want to build less-correlated exposures such as dividend-rich stocks, commodities, bonds and infrastructure.
Avoid AI hangers-on: there’s no doubt AI is a powerful set of technologies which are here to stay. That’s doesn’t mean every business claiming AI capability will succeed. Right now, many businesses offering AI services are doing so at a loss – and won’t succeed in their search to do so at a profit. To judge by the dot.com bubble, some will dominate, some will survive but won’t be spectacular for investors, and some will fail.
Consider taking profits if you have done well in recent market conditions. This could mean trimming your exposure rather than exiting entirely. In effect, you’ll be re-balancing your portfolio, and likely de-risking in the process.
