At the turn of the year, the outlook for investors is unusually difficult to call. Signs of a possible investment bubble in AI-related stocks have prompted some falls in technology stocks, but often followed by a rally. Valuations are stretched, and investment in data centres by the hyper-scalers is ambitious, but while individual institutions building AI capability may falter, many investors are looking at a broader picture in which AI may improve or even transform business performance across the economy.
BlackRock, the world’s largest asset manager, declared itself continuing to be ‘pro-risk’ in its Outlook for 2026. It remains bullish regarding US tech stocks, but selectively so, stating that it is a good time for active investing, ‘for those with insights on who will capture the revenues’. The ‘micro’ has become macro – the scale of the investments in building AI capability by a small number of large tech firms is such that it has a macroeconomic impact.
The projected investment in AI is projected to be as high as $8tn by 2030, while the increase to revenues for the AI hyper-scalers is estimated to be around $1.6tn per year. This, however, is not the full story. One reason that high AI-related valuations do not obviously constitute an investment bubble is that the technology has applications across the economy. BlackRock compares the AI revolution to earlier industrial revolutions such as the steam engine and mains electricity, increasing productivity and spurring economic development.
Taking on debt to front-load the investment is necessary to realise the returns over a longer period, but this does come with risk to the financial system, at a time of high public sector debt. ‘Bond yield spikes could pose a risk to this financing,’ the report states. A structurally higher cost of capital raises the cost of AI-related investment.
A small number of stocks with potentially high but uncertain prospects mean that divesting or staying is a big active call: There are no easy passive diversification options.
The report notes that the long-term trend growth rate in GDP for the US has been around 2% per year for the past century. If the AI revolution significantly increases this, it would be historic. Such an outcome is a ‘tall order’, but is feasible, BlackRock concludes. The likely uneven impact on AI, with some businesses realising the gains more than others, calls for active investing and a rethink of traditional approaches to portfolio construction.
If the AI spending fails to lift the trend growth rate above 2%, then it could crowd out non-AI investment, and potentially cause inflation to rise.
A major constraint in the US is energy capacity. AI data centres could be consuming up to a fifth of electricity generated in the US by 2030. On energy, China may be winning the competition as it is proving to be able to build scale affordably in different energy sources, including hydropower, solar and nuclear as well as coal. DeepSeek, the highly competitive Chinese large language model, is more efficient on energy use than US rivals.
The Franklin Templeton Institute, in its Global Investment Outlook, describes the ‘Age of Intelligence’ as a major theme. The AI revolution is just beginning, the report advises, and the potential for economy-wide efficiency gains is ‘enormous’.
Franklin Templeton identifies three themes, which are interlinked: Broadening, steepening and weakening.
Broadening: US equities have outperformed others for many years, and while prospects remain strong, there are opportunities emerging more broadly both within the US, and internationally. The profits of small cap US businesses are set to rise, owing to reduced interest rates and lower debt servicing costs. Assets in emerging markets show potential for attractive returns.
Steepening: The combination of falling interest rates while inflation is above target will result in steeper yield curves. The Federal Reserve is likely to continue cutting interest rates in 2026, while two other factors are set to cause the yield curve to rise: Increased demand for capital to fund investment, especially in AI and energy capacity; and continued increases in borrowing by governments in the largest economies.
Weakening: The US dollar weakened by around 10% in 2025, and is set to weaken further. Inflation and interest rates are lower in Europe there will be a lower cost of currency hedging. Commodity prices will likely rise, with the exception of oil. The weakening dynamic further encourages the broadening effect by enhancing potential returns in non-US assets. In emerging markets, import costs and inflation will fall allowing central banks to ease policy, underpinning bond prices. Credit ratings in emerging markets have improved.
Franklin Templeton also pointed to the rise in government intervention: In trade, through tariffs and other barriers, and subsidising or otherwise protecting certain businesses or sectors. While there are legitimate reasons for some policies that curb globalisation, such as security, in general markets are better than governments at allocating capital, the report states.
It is notable that neither report makes a definitive call of a bubble in AI stocks, although they do identify risks in the scale of investment, as well as systemic financial risks around leverage and high government borrowings.
While there is a rational basis for being bullish, perhaps ‘carefully bullish’ might be an apt phrase.
The author is a Qatari banker, with many years of experience in the banking sector in senior positions.
