AI barbarians at software’s gate

Artificial intelligence has the potential to reshape productivity, labour markets, and our industrial infrastructure. Its impacts will be comparable in importance to the arrival of electricity, computing or the internet. In the meantime, AI has profound implications for the structure of the software industry, which has become a target in the debate around the technology’s potential to disrupt business models.

The recent market slump in software stocks has been broad and indiscriminate. Investors have reacted as if all traditional software will soon be replaced by a new generation of AI companies such as Anthropic or OpenAI. We take a more measured perspective, and believe the recent correction creates investment opportunities. Not all software firms will be displaced, and several areas of the sector deserve a closer look.

As a true general-purpose technology, AI will recast the software industry over time – shifting monetisation, reshaping pricing, creating new product classes, driving heavier compute demand, and unlocking broad productivity improvements. A recent study by the Organisation for Economic Co-operation and Development found that generative AI – a technology capable of producing original output – is already pervasive, continuously improving and innovative because it enables new applications and products.

A multi-year transition

The transition to AI will widen the gap between stock market winners and laggards. However, we expect to see a multi-year-style rotation within software, rather than a sector wide collapse of existing business models. Traditional application software certainly faces pressures, but not uniformly. AI will automate basic office tasks such as note taking, summarising meetings and low‑value workflows. We can therefore expect that operationally shallow software categories, or those easiest to replicate through an AI agent, are most vulnerable to displacement or pricing pressure.

Market concerns… look overdone

Market concerns that AI agents could replace incumbent software firms, depress demand for subscriptions as fewer employees are more productive, and lower barriers to entry look overdone. In our view, unlike current market sentiment, the implications are not uniformly negative.

Instead, it is more useful to think of AI as simply the next layer of software. It pushes productivity and automation higher but still depends on large systems that:

1. store ‘true’ data from customer details, financial data, and human resource records;
2. organise and manage information including version control;
3. deliver cloud and compute capacity through large-scale providers.

Incremental demand for more, not less software

AI therefore creates an incremental demand for more software, not less. This is because it shifts value within the software industry rather than replacing the industry altogether.

Why is this? Existing software platforms provide an anchor for a business around its customer activity, revenue, employees, and financial data. AI then generates more interactions, transactions, and data, all of which ultimately flow into these underlying systems. Companies with deep enterprise integration in customer management, personnel platforms, or financial and compliance systems, should therefore become more necessary rather than obsolete. They are the base layer that AI sits on, not a layer that AI can replace.

Many quality names now trade well below their historical ranges

Moreover, AI requires large volumes of high-quality data, clean and secure software code, along with strong governance and version control. We thus expect companies that provide data management, workflow control, and developer tooling to benefit from the proliferation of AI agents.

The recent equity market sell‑off has pushed software valuations back to levels that already discount a severe and prolonged disruption scenario. Many quality names now trade well below their historical ranges and, in several cases, at recession‑type valuation metrics. While fundamentals are in flux, such compressed valuations can act as a natural stabiliser for the sector, limiting further downside and creating an attractive entry point for long‑term investors as the AI transition unfolds.

In short, it appears that the market is currently mispricing the distribution of AI’s impact and the timeline for its effects. Selling pressure has been broad and indiscriminate, driven by fear rather than fundamentals. This misses the fact that AI’s impact is not evenly distributed: some companies are clear winners, some are structurally challenged, and many are simply misunderstood. We think there is opportunity for selective accumulation in names where the market has pricedin disruption that is unlikely to materialise.

Behavioural shift

At its core, generative AI remains software; written by developers and released in versions just like any other program. It is then deployed and maintained using familiar engineering practices. It runs on compute, follows underlying rules, and produces outputs from inputs. In that sense, generative AI represents the next major step in the long evolution of software, making it more like previous shifts towards high‑level programming languages, microservices, or cloud computing.

Software is moving from being a rigid tool to something closer to a collaborator

Where AI differs is in its behaviour. Traditional software is deterministic; the same input always produces the same output because the logic is explicitly programmed. Generative AI, in contrast, is probabilistic. It learns patterns from data rather than relying on fixed rules, which means that users may receive slightly different answers to the same prompt. The spirit of the output is consistent, but no answer is identical.

related content.

CIO Office Viewpoint

What US tariff policy shifts and Middle East risks mean for markets

February 24, 2026

investment solutions.

learn more

This shift means that software is moving from being a rigid tool to something closer to a collaborator, capable of generating ideas, drafting content, summarising information, or writing code. Rather than replacing software entirely, generative AI replaces specific pieces of the old rules‑based architecture, the underlying structural framework that traditionally governed how applications behaved, by working at a higher level of abstraction. The result is a more flexible, adaptive layer that can take on more cognitive tasks that were traditionally performed by humans or by single-purpose applications.

Technological innovation rarely rolls out over a timeline that investors expect. Typically, markets reflect too much excitement in the near term and fail to fully appreciate the significance of the long‑term effects.

Markets reflect too much excitement in the near term

With this in mind, we see attractive opportunities in quality growth stocks at a reasonable price within the tech sector following the recent consolidation, while keeping our overall neutral positioning in developed market equities. Overall, we keep a neutral stance on the technology sector – preferring healthcare, materials and utilities, which should benefit from innovation and productivity gains, as well as power demand related to the AI roll-out. We implement our broad overweight equity stance via emerging markets, where we see more attractive valuations and earnings growth.