Insights
By:
Turning promises into profit will take time
- AI investment is accelerating faster than enterprise adoption and monetisation, as organisations take time to integrate AI into workflows amid infrastructure, governance and compliance constraints.
- As adoption, infrastructure and regulation mature, AI technology will create new growth opportunities in digital advertising, AI agents, identity management and zero‑trust security.
AI sits at the centre of a multi‑year Digital Transformation megatrend. Although capital is flowing into the sector at software speed, adoption is moving much more slowly. Though consumers can adopt new technologies very quickly, organisations can only harvest productivity and returns as fast as they can integrate AI into workflows. And that can take a lot longer.
The reason for this is that adoption of AI is not something that can happen at a uniform pace. IT applications can scale rapidly, but infrastructure almost inevitably evolves more slowly due to chokepoints around data centre capacity and power requirements, while governance and compliance frameworks change on multi-year timelines. Together these will naturally limit the deployment of AI even as budgets increase.
Monetisation adds another layer of complexity. When ChatGPT launched, it reached 100 million users within weeks and by late 2025 boasted around 800 million users[1], yet only ~5% pay for a subscription according to the FT[2]. This gap between usage and paying customers highlights the challenge of converting enthusiasm into durable revenue.
Despite these challenges, some areas have already found ways to monetise AI effectively. The clearest success so far is digital advertising, which is already a vast market. Google’s ‘Performance Max’ and Meta’s ‘Advantage+’ both leverage AI to optimise targeting and bidding, delivering significant efficiency gains and margin improvements. This is tangible proof of AI driving operating leverage – doing more with the same resources.
Other areas also offer growth potential. AI’s next phase is likely to focus on AI agents, software that acts on our behalf to book travel, make purchases, schedule meetings or analyse data. Each agent must be verified and monitored, which requires a secure digital identity. In the physical world each person has one identity, but in the digital world a single individual might manage dozens of agents, and agents will interact with agents. This creates a new infrastructure layer: identity management. Spending on identity and security systems is already growing briskly as businesses prepare for this shift.
Security will be a critical challenge illustrated by examples like the ‘EchoLeak’ vulnerability of 2025, which showed how a Microsoft AI assistant could be manipulated to leak sensitive data without any user action. The risk is structural: AI agents operate at machine speed with broad permissions and can be tricked in ways that humans might not anticipate. This is driving demand for ‘zero-trust’ security models, which assume no system or agent is trusted by default, and use strict identity checks, temporary credentials and continuous monitoring.
In short, beyond the obvious beneficiaries of the AI revolution, like hyperscale cloud platforms, digital advertising and semiconductor firms, AI is opening new growth avenues. Some will materialise quickly as in security and identity systems, but others will need time for the infrastructure, adoption and governance to catch up.
Alastair Dean is a Director responsible for portfolio management across a number of equity, multi-asset and thematic mandates. He also contributes to the firm’s equity fund research and manager selection efforts.
Hear more from Alastair
Disclaimer: Opinions expressed here are as of the date of publication and subject to change without notice. Stonehage Fleming Investment Management Ltd shall not be responsible for any trading decisions, damages, or other losses resulting from, or related to, the information, data, analyses or opinions contained herein or their use, which do not constitute investment advice, are provided as of the date written, are provided solely for informational purposes and therefore are not an offer to buy or sell a security.
Any information which could be construed as investment research has not been prepared in accordance with legal requirements designed to promote the independence of investment research.
Past performance is not a guide to future returns.
All investments risk the loss of capital.
Issued by Stonehage Fleming Investment Management Limited (SFIM). Authorised and regulated by the Financial Conduct Authority (194382) and registered with the Financial Sector Conduct Authority (South Africa) as a Financial Services Provider (FSP No. 46194).
[1] TechCrunch – October 2025
[2] FT – OpenAI makes 5-year business plan to meet $1tn spending pledges - October 2025
