Valuation Cycles and Risk Management – Q3 2026
New technologies reshape the future. Yet the expectations placed on them often develop faster than their economic reality. History shows that major innovations first spark enthusiasm and high valuations, long before their true economic value has fully unfolded. Over time, expectations and reality realign, the euphoria fades and valuations normalise – while the innovation itself endures.
The internet changed the world profoundly. Even so, many of its eventual winners lost more than 80 per cent of their market value during the dotcom correction before their businesses grew into those valuations years later – while numerous others disappeared from the market altogether. Innovation and valuation rarely move at the same pace.
For the present, this yields an important observation. Market corrections do not arise solely in phases of economic weakness or recession. They can equally occur when expectations rise faster than economic reality. In such phases it is not the innovation that is called into question, but its valuation.
High valuations do not inevitably lead to falling markets. They do, however, alter the balance between opportunity and risk: the remaining upside narrows, while the downside in the event of disappointment widens. This does not mean a correction is imminent. It means, rather, that the starting position has changed.
Precisely this tension between innovation, expectations and valuations can be observed in every major technological phase. The current AI investment cycle is no exception.
A Current Example
The long-term potential of artificial intelligence is beyond question. At the same time, the question arises whether the pace of investment and the expectations attached to it have, in certain areas, already been partly anticipated.
This is evident in a mechanism recently discussed as the AI financing loop. Put simply, some large chip and cloud providers invest in AI companies. Part of this capital then flows back, as those same companies purchase computing power and chips from precisely these providers. This gives rise to capital flows that partly circulate within the same value chain. A comparable pattern emerged around the turn of the millennium in the telecommunications sector, when equipment makers helped finance the growth of their own customers – until demand fell short of expectations and these interdependencies unwound.
Part of the reported growth thus rests on capital flows within the same ecosystem. As long as these investments continue, the mechanism persists. Should it slow, growth expectations may adjust accordingly.
Estimates from various market analyses suggest that the mutual financing and investment interdependencies accumulated to date amount to well over US$800 billion. Measured against the US equity market as a whole, this figure appears modest at first. What matters, however, is less the absolute amount than the concentration of this capital in a small number of companies that today shape a significant part of index performance. In the event of disappointment, what would primarily be repriced there is not a single sum, but the growth narrative that underpins these valuations.
The AI investment cycle is therefore less a forecast than a current example of how expectations, capital flows and valuations can interact. Whether a future repricing actually originates there or is triggered by other developments remains open. The underlying mechanism, by contrast, is timeless.
Known Risks Are Not Automatically Priced In
The real challenge therefore lies less in the risk itself than in how it is interpreted. The relationships described have long been known and are widely discussed. To conclude from this that they are already fully reflected in valuations would be premature. Being known and being priced in are two different things.
Capital markets have repeatedly shown that risks do not lose significance simply because they are talked about. What matters, rather, is how expectations, valuations and the behaviour of market participants interact. As long as these do not change, the starting position too remains largely unchanged.
In phases of great confidence in particular, a high concentration on a few themes or companies often emerges. In itself this is neither unusual nor problematic. Yet the more one-sided expectations become, the more sensitively valuations react to disappointment – not because the long-term prospects fundamentally change, but because a great deal of good news had already been taken for granted.
Here lies the real heart of the matter. A surprising event finds market participants differently positioned: some are exposed, others hedged, still others stand on the other side and buy. The pressure is dispersed. With a widely known risk, the reverse holds. Because everyone shares the same assessment yet remains invested, most are positioned in the same direction – on the same side, in the same few companies. Should that assessment change, it is not individual participants who wish to sell, but nearly all at once. And when everyone tries to leave through the same door at the same time, there is no one on the other side to buy. The market then falls not in an orderly fashion, but abruptly. This is precisely why, under one-sided positioning, even small shifts in expectations can trigger disproportionately large movements.
The current AI investment cycle illustrates this interplay to a particular degree. It combines high expectations, exceptional investment and a strong concentration of capital flows in a few companies – exactly the conditions under which one-sided positioning can arise.
Concluding Thoughts
New technologies will continue to drive innovation, productivity and economic progress. Of that there is little doubt. Equally, there is much to suggest that capital markets will, in future too, repeatedly experience phases in which expectations run ahead of economic reality and valuations adjust accordingly.
For investors, the challenge therefore lies less in predicting the next turning point. What matters more is to reassess expectations, valuations and risks on an ongoing basis and to draw decisions from this with the necessary discipline.
A normalisation of valuations would, from this perspective, be no extraordinary development, but part of a recurring market mechanism. Innovation and valuation rarely move at the same pace – and therein lies the challenge of long-term risk management.
Disclaimer. This publication is issued by Livalor Vermoegensverwaltung AG, Zurich and Vaduz, for general information only. It constitutes neither a solicitation, an offer, nor a recommendation, express or implied, in respect of any financial instrument, issuer, sector, market or asset allocation, nor an invitation to engage in any transaction. Market mechanisms described by way of example are illustrative only and are not intended to identify or assess any particular company or instrument. This publication is not investment, legal or tax advice and is not tailored to any recipient; before making investment decisions we recommend consulting a qualified professional. Reading it creates no client or advisory relationship. The information is based on sources we consider reliable, although we do not guarantee its completeness or accuracy. It reflects the position at the time of preparation, may change at any time and will not be updated. Past market phases are not a reliable indicator of future results; forward-looking assessments rest on assumptions and are subject to uncertainty, so that actual developments may differ materially. Livalor Vermoegensverwaltung AG and the portfolios it manages may hold positions in instruments of the kind referred to herein. This publication is not directed at any particular jurisdiction and constitutes, in any jurisdiction, neither an offer or solicitation nor the provision of investment advice or portfolio management services. Livalor Vermoegensverwaltung AG offers its services only where it is authorised to do so. Any person accessing this publication, or contacting us in response to it, does so on their own initiative and is responsible for observing the laws applicable to them; no such contact obliges us to provide any service. To the extent permitted by law, we accept no liability for any loss arising from reliance on this publication. No part may be reproduced without our prior written consent.