Expectation is the new volatility

For years, investors reassured themselves that while markets could be excitable, valuation ultimately rested on something observable. Earnings might wobble, margins might compress, but multiples would move in response to hard evidence. Then an AI announcement lands and an insurance distributor, previously trading comfortably at 18 times earnings, is suddenly marked at 11 times. Nothing dramatic has happened to current revenue. What has shifted is the market’s expectation of what revenue might look like in a world where underwriting, claims handling or client interaction is increasingly automated.

That shift is the real process change AI is introducing into investing. It is less about replacing analysts and more about widening the distribution of possible futures - and accelerating the speed at which markets price those futures in.

Inside investment firms, AI is largely being used to improve rigour: synthesising diligence, structuring portfolio data, back-testing historical underwriting against realised outcomes. That is healthy. But outside the firm, in the market, the reaction function has changed. The moment a credible narrative emerges that a revenue pool might be structurally altered, multiples adjust. Discount rates rise not because earnings have fallen, but because the range of outcomes has expanded.

This is expectation risk.

For short-term investors, expectation risk is destabilising. If sentiment shifts faster than fundamentals, prices can move violently before there is evidence either way. An 18x multiple becomes 11x on plausibility alone. Volatility increases because imagination moves more quickly than cash flow.

For long-term investors, however, expectation risk can create entry points. If the repricing reflects fear rather than structural erosion, a compressed multiple may represent opportunity. The challenge, of course, is epistemic: how certain can one be about the future shape of that revenue stream? Is AI nibbling at the edges, or is it rewriting the business model? That is the uncomfortable part.

AI does not simply increase uncertainty; it accelerates the market’s reaction to uncertainty. It compresses the time between hypothesis and price move. In that environment, investing becomes less about reacting to earnings prints and more about assessing which narratives will prove durable.

Expectation risk cuts both ways. It punishes complacency quickly, but it also offers disciplined capital a chance to step in when sentiment outruns substance.

The difficulty is not seeing that volatility creates opportunity. The difficulty is knowing, with any conviction, whether the future is becoming clearer - or simply louder.

Rebecca Lewis is Partner & Co-CEO at Arisaig Partners. The views expressed are personal.