Equity markets are rebounding from the latest AI-bubble fear sell-off. While this will almost certainly not be the last time investors pull back from the theme, we believe the worries are overstated.
Arguments supporting the idea of an artificial intelligence (AI) bubble can be evaluated by means of the price-earnings (P/E) ratio. Bubble concerns centre around both the ‘P’ and the ‘E’ of the ratio.
Let’s start with the denominator: ‘E’. The worries investors have are two-fold (and somewhat contradictory): Either earnings expectations are not high enough – that is, they will not rise sufficiently to generate an adequate return on the capital expenditure investment currently taking place; or else earnings expectations are too high and downward revisions will lead to a market sell-off.
Investment by tech companies has indeed increased at a rapid pace over the last year. Earnings growth rates have begun to pick up, but not yet as quickly, with a consequent decline in the expected return on investment (see Exhibit 1).

Whether tech earnings eventually do rise by enough is the fundamental question for investors.
However, given the vast potential for the use and monetisation of AI across the economy — from the development of new drugs to the substitution of capital for labour — we are still optimistic they will. It is important to note that capital expenditures are not forecast to continue growing at such a rapid rate, which should lead to a recovery in the return on investment (ROI) by 2027.
The rate of earnings growth forecast over the next few years is slightly above the historical average, but not dramatically so.
This suggests the second argument for an AI bubble – that earnings expectations are too high – is unfounded. Earnings growth for the tech-heavy NASDAQ index is forecast at 18% in 2026 and 17% in 2027, compared to a long-run average of 14%. This does mean, however, that the payback from today’s capital expenditures will not be immediate.
Valuations: ’P’
Even if the expected earnings do materialise, the price multiple on those earnings may be too high, leading to sub-par returns in the years ahead. The current forward P/E ratio for the NASDAQ is 27.5x, not far below the 30.9x level reached at the end of 2021. The sell-off in the market the following year was dramatic.
On a relative basis, the current valuation is indeed high, but less so than for several other markets. The z-score for the current P/E is 0.5, compared to 1.5 for the Russell Value index (see Exhibit 2).
Exhibit 2
Forward price-earnings ratio z-score

Data as at 1 December 2025. *Calculated from 2010; all other indices calculated from inception. BT = technology, broadline retail, interactive media & services. Sources: IBES, Bloomberg, BNP Paribas Asset Management.
We would still anticipate that the NASDAQ’s valuation eventually reverts to its mean value of 21.3x; the question is simply how. The 2022 scenario, when an increase in the policy rate by the US Federal Reserve triggered the index’s decline, seems unlikely next year as the fed funds rate is expected to fall. Similarly, economists do not expect a recession or sharp slowdown in growth.
A more benign evolution would see a more modest appreciation in the index price than the 21% gain so far this year (through 1 December), even as earnings continue to rise. A 10% gain in the index alongside the forecast 16% rise in earnings per share (EPS) would lead to a lower P/E ratio.
Hence valuations, while high, do not preclude further gains in the index, albeit at a slower pace than earnings growth.
Emerging markets
NASDAQ is not the only index through which investors can gain exposure to the AI theme. Emerging markets, particularly Taiwan, China and South Korea, offer promising opportunities.
While valuations relative to history are similar to those for NASDAQ, EM tech equities offer two key advantages: the valuations relative to the NASDAQ are comparatively low, with a z-score of -0.4 (see Exhibit 2).
Moreover, at 30% in both 2026 and 2027, EM tech earnings growth expectations are much higher than those for NASDAQ stocks. Even if these figures are too optimistic, earnings growth is nonetheless still likely to beat that of the US stocks.
Asset allocation views
- We remain constructive on risky assets which are supported by resilient global growth and accommodative monetary policies. US equities have benefited from a strong earnings season. As economic data releases resume after the shutdown, unforeseen developments may adversely affect the US equity market, which is now trading at elevated valuations. As the risk-reward dynamic is becoming more asymmetric, we have taken profits and tactically closed our position in US technology stocks.
- We maintain an overweight in Japanese equities, which should benefit from the structural changes taking place in the economy. We continue to see value in emerging markets across both equities and debt. We expect EM central banks to continue/resume their easing cycles in 2026. The heterogeneity of the emerging economies provides valuable diversification to our risk-on positioning.
- Our conviction, while reduced, is still positive on gold, which should benefit from the structural imbalance between supply and demand, especially from emerging market central banks.