AI & Technology - The Emerging Investment Theme

By EC Assets Research Team, Technology Themes · Published 19 May 2026 · Updated 16 October 2026

AI & Technology: Artificial intelligence and broader technology themes have become a primary institutional investment focus since 2022. The category spans direct AI infrastructure (chips, data centres, models) and AI-affected sectors across the economy.

What the AI and Technology Theme Is

Artificial intelligence became a central investment theme after generative AI tools reached a mass audience in late 2022. The theme runs along a physical and economic value chain: semiconductors and the equipment to make them; data centres and the real estate they occupy; electricity generation and grids, part of infrastructure; networking; the developers of AI models; software that embeds AI; and companies in every sector whose costs or competitive positions may change as they adopt it.

Investors reach the theme through listed equities, where a few very large technology companies dominate, through venture capital and growth equity in private model developers and applications, and through infrastructure and property funds that build data centres and power capacity. AI also changes investment management itself, as covered under quantitative investing.

How It Works

The investment case rests on productivity. If AI substantially raises output per worker across the economy, the companies that supply and use it effectively can earn large profits for years. Early evidence is encouraging: Brynjolfsson, Li and Raymond (2025) found that customer-support agents using a generative AI assistant resolved 14 percent more issues per hour on average, with the largest gains among less experienced workers.

The investment risk lies in prices. Pástor and Veronesi (2009) show that during technological revolutions, uncertainty about a new technology's productivity can push the prices of innovating companies up sharply, and that prices can fall as the uncertainty resolves and the technology is adopted across the economy, even when it is a success. Railways, electricity and the internet all followed this pattern. The question for investors is therefore not only whether AI matters, but what growth is already priced in, and who will capture the profits: chip makers, model developers, platform companies or the businesses that use AI and compete away some of the gains for their customers.

Supply chains and policy add their own risks. Manufacturing of the most advanced chips is concentrated in a few companies and locations, notably Taiwan, and export controls on advanced chips and chipmaking equipment have become an instrument of policy. Regulation of AI itself is developing: the EU AI Act, in force since August 2024, phases in obligations for high-risk and general-purpose systems over several years.

Spending on data centres and chips must eventually earn a return. If revenue from AI services grows more slowly than capacity, the result can be overcapacity and write-downs, as with telecommunications networks after 2000, when much of the fibre laid during the boom sat unused for years.

Worked Example

Suppose an investor wants an 8 percent annual return over ten years from a stock, and expects its price-to-earnings ratio to fall back to 20, a level typical of mature companies. Ignoring dividends, the earnings growth required depends on the starting valuation:

Return target 8 percent a year, P/E ending at 20, no dividends; a different target or ending multiple changes the figures.

Growth of 15 to 20 percent a year for a decade is rare. Chan, Karceski and Lakonishok (2003) find that the number of companies sustaining above-average growth for many years is close to what chance would produce, and that analysts' long-term growth forecasts are too optimistic for the companies with the highest expectations.

History shows what happens when growth disappoints. The internet transformed commerce and communication, yet the Nasdaq-100 index fell 82.9 percent from 27 March 2000 to 7 October 2002 and regained its 2000 level only in November 2015, 15.6 years after the peak, measured without dividends.

When It Applies (and Limitations)

The table is arithmetic, not a forecast. It shows what a price implies; it does not say whether a company will achieve it.

Different from 2000. Many leading AI-related companies today are highly profitable, unlike many internet companies in 2000, which makes the analogy imperfect.

Uncertain timing. Productivity gains from general-purpose technologies have historically taken years to spread, which affects who profits and when.

Physical constraints. Electricity supply, chip manufacturing capacity and grid connections can limit growth, shifting returns towards owners of scarce inputs.

Why It Matters for Institutional Investors

Concentration. Market-cap-weighted equity indices hold large positions in a few technology companies, so a passive portfolio already carries substantial AI exposure; the asset allocation should account for it before adding more.

Implied expectations. Comparing the growth that prices imply with base rates of corporate growth keeps enthusiasm in check.

Look across asset classes. Data centres, power generation and grid investment offer exposure through contracted or regulated cash flows, with different risks from equity in AI companies.

Diversify across the value chain and scenarios. Positions sized to survive both rapid adoption and a slower, more competitive outcome are more robust than concentrated bets on either.

References

1. Pástor, Ľ., & Veronesi, P. (2009). Technological Revolutions and Stock Prices. American Economic Review, 99(4), 1451-1483.
2. Chan, L. K. C., Karceski, J., & Lakonishok, J. (2003). The Level and Persistence of Growth Rates. Journal of Finance, 58(2), 643-684.
3. Brynjolfsson, E., Li, D., & Raymond, L. (2025). Generative AI at Work. Quarterly Journal of Economics, 140(2), 889-942.
4. López de Prado, M. (2018). Advances in Financial Machine Learning. Wiley.