KEY POINTS:
- AI’s disruptive impact is spreading beyond the tech sector, creating compelling opportunities across the economy
- 88% of enterprises are using AI in at least one function, but few have rolled it out across the business
- Current market conditions share some features with the dotcom bubble but differ in critical respects
- Bottom-up fundamental research is key to identifying the winners and avoiding those being disintermediated by rapid innovation
Digital transformation is the most durable investment theme of our era. By 2030, the global AI market alone is forecast to reach $1.8 trillion, growing at a 36.6% compound annual growth rate.1 AI’s total economic impact, meanwhile, is projected to yield $22.3 trillion, or 3.7% of global GDP.2 Underpinning it all, capital expenditure on data centre infrastructure is set to reach $6.7 trillion by the end of the decade.3
Yet, in our view, the biggest opportunities may not lie where most investors are looking. As the technology sector grows and disrupts more and more parts of the traditional economy, its boundaries are dissolving. Simply identifying technology sector winners is no longer sufficient; the critical question is which companies across the broader economy are making the most of these rapid innovations and realising major efficiency and productivity gains as a result.
The cost of digital transformation
There is little doubt that AI has the potential to transform the global economy, and that adoption is still in its early stages. A 2025 McKinsey survey found that while 88% of companies had adopted AI in at least one function,4 most are still experimenting to identify the highest-impact use cases and few have rolled out the technology across the business.
At the same time, the companies involved in AI’s development – not just software firms but also semiconductor manufacturers and builders of data-centre infrastructure – have been able to deliver highly attractive earnings streams and capital growth to their shareholders. The capex of leading hyperscalers in the US is projected to lead to investment worth more than US$650 billion in 2026 – more than triple the total outlay in 2024.5
This spending is in response to strong demand. The contracted backlog for the top three hyperscalers (Microsoft Azure, Amazon AWS and Google Cloud Platform) exceeded $1 trillion of future revenue as of the end of 2025, almost doubling on a year-over-year basis.6 All three are delivering strong growth and accelerating revenue for their cloud platforms.
Over the course of the past year, however, questions have arisen around the sustainability of these returns, and around technology company valuations in general. Whether the potential long-term returns from AI-driven productivity gains can justify spending at this scale is one of the central questions for investors in 2026. Are current valuations a reasonable reflection of future cash flows – or are we in another dotcom bubble?
Risks in the current environment
There are a number of similarities between present market conditions and those in the lead up to the crash in internet and telecoms stocks in early 2000. As in the late 1990s, we are seeing intense competition between firms as they race to constantly update their AI models’ capabilities in order to be first to market. This could result in the overbuilding of infrastructure.
The huge up-front investment required by today’s hyperscalers and, in particular, the use of debt-financing vehicles such as private credit and off-balance sheet structures has the potential to add significant risk. In the dotcom era, cheap credit and widely available venture capital led numerous technology companies to expand despite being unprofitable.
Another red flag in the current environment is the emergence of circular financial arrangements between a number of major participants in the AI industry. Examples include hardware manufacturers extending financing to customers to fund purchases of their own products. Such practices that could increase systemic risk.
Confidence in the investment outlook
However, we believe there are several mitigating factors that mean we are yet to enter genuine bubble territory. For one, the cloud service providers that are leading the way in terms of AI capex are well-established companies boasting strong balance sheets and the ability to generate positive free cash flow. This is in stark contrast to the late 1990s, when much infrastructure investment was financed through debt, and company cash flows and balance sheets were much weaker.
It is also worth noting that much of the infrastructure needed to deliver AI-based services to end users – fibre networks, smartphones and other devices – already exists. During the dotcom boom, the infrastructure required for internet-based services was very much in its infancy, making it more difficult for providers to generate revenues.
At the same time, the AI boom is still at an early stage. ChatGPT was launched less than four years ago, and AI agents – which have the potential to drive a dramatic rise in process automation – are even more embryonic. We also see considerable scope for foundation models to integrate with physical devices in areas like robotics.
From our point of view, a final reason to retain confidence in the investment outlook for AI is that current technology stock valuations remain far less stretched than they were in the run-up to 2000. In fact, the technology sector’s modest performance in 2025, paired with sustained earnings growth, has actually led to cheaper valuations on a one-year forward basis. While valuations remain above the 2005-2020 average, this premium is supported by profitability metrics that are at an all-time high.
For these reasons, we maintain our view that AI is the most impactful digital transformation theme since the internet. That said, it is vital that investors remain aware of all potential risk factors, including a possible slowdown in the spending cycle. It is perhaps even inevitable that we will see a period of consolidation within the technology sector. Over the next few years, winners will emerge and specific applications or platforms will achieve market dominance.
Focus on sustainability, resilience and value
We therefore believe in looking past volatility to unearth the most compelling opportunities.
This broadening is already showing up in market performance. In 2025, only two of the ‘Magnificent Seven’ outperformed the broader market, even as technology stocks as a whole continued to deliver robust earnings growth. This points to the next phase of returns from disruptive technology being dispersed beyond the companies building AI infrastructure themselves.
We believe the real opportunities lie in businesses across healthcare, industrials, financial services, energy and consumer sectors that are applying novel technologies to transform their own operations and competitive positions.
We are particularly interested in adjacent disruptive technologies such as cloud computing, the Internet of Things – where the number of connected devices is projected to reach 39 billion by 20307 – and automation. Emerging areas also warrant close attention: quantum computing is expected to grow from $928.8 million to $6.5 billion by the end of the decade,8 while the gene editing market is forecast to reach $28.6 billion by 2032.9
To manage risk and increase portfolio resilience across the market cycle, we concentrate on companies that are sustainable in both financial and environmental terms. They should have a robust competitive moat, and – equally importantly – be attractively valued. Crucially, we have the freedom to invest across world equity markets, unconstrained by benchmark or market-cap requirements.
The changes we are seeing in both the global economy and financial markets can be a double-edged sword for investors. Innovation will increase productivity and create new markets. But the path to growth can be difficult to predict, and often brings with it uncertainty and market turbulence. As AI rapidly reshapes industries, bottom-up fundamental research is key to identifying the winners while avoiding those being disintermediated. Our goal is not only to interpret the changes taking place, but also to ensure our clients are able to navigate the evolving landscape with greater insight and resilience against volatility.
Source: BNP Paribas Asset Management, March 2026
[1] https://www.grandviewresearch.com/industry-analysis/artificial-intelligence-ai-market
[2] https://my.idc.com/getdoc.jsp?containerId=prUS53290725
[4] https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
[5] BNP PARIBAS DISRUPTIVE TECHNOLOGY, A thematic fund investing in the innovators and beneficiaries of worldwide digital transformation, March 2026, p.10
[6] BNP PARIBAS DISRUPTIVE TECHNOLOGY, A thematic fund investing in the innovators and beneficiaries of worldwide digital transformation, March 2026, p.11
[7] https://iot-analytics.com/number-connected-iot-devices/
[8] https://www.marketsandmarkets.com/Market-Reports/quantum-computing-market-144888301.html
[9] https://www.psmarketresearch.com/market-analysis/genome-editing-market
At the time of writing 20/3/2026, the Middle East conflict has not warranted any major changes to our base case macroeconomic outlook or investment recommendations. To follow our analysis of the events driving asset markets, go to Viewpoint at https://viewpoint.bnpparibas-am.com
Visit our fund centre
Learn more about our strategy
Combining perspectives to navigate volatility
Discover how structural shifts are fundamentally reshaping our world
