Software’s sudden derating

As with any investment, your capital is at risk.

The first weeks of 2026 delivered a sharp, at times unsettling, rotation in equity markets.

Shares in software and other digital information businesses fell hard, even as large parts of the broader market held up better.

The extent of the derating has been remarkable. A recent analyst note highlights that software and IT services are now trading at only a 9 per cent premium to the market, the lowest gap since the 2008 financial crisis.

The AI shock moves through the stack

The catalyst for this rotation was the market’s renewed focus on advances in AI, and whether recent progress will fundamentally change the economics of software development.

While coders have used AI in software production for some time, the newest models – notably Anthropic’s Claude Code – take this a step further. Programmers can now release code entirely written and tested by AI, with little or no human intervention. What’s more, they’re doing it faster than ever.

The negative bear case writes itself: why would companies pay for software if they can ask a small team using AI to build a custom app for them? At the very least, it seems likely to increase competition and lower barriers to entry.

But still, an overreaction

Frankly, even if AI can write more code, it doesn’t mean that all software companies will become worthless.

These businesses don’t just sell code; they sell a complete service. This includes support, compliance, integration with other tools, and the overall user experience. Those are precisely the complicating factors that make it unlikely that enterprises will replace their software subscriptions with custom AI-built tools anytime soon.

If enterprise customers instead build their own internal tools using AI, they would need to stay on top of the constant maintenance, security updates, record-keeping and documentation – and be willing to carry all regulatory and operational risk rather than sharing any potential liability with a third-party software provider.

And it’s revealing that even major AI providers frame their role as enabling, rather than replacing existing enterprise systems. Take OpenAI’s announcement of Frontier, a new business platform that helps companies create and manage AI agents. OpenAI emphasised that Frontier “works with the systems teams already have, without forcing them to re-platform.”

This vision implicitly assumes that core business software will continue to exist – particularly software like customer management databases, finance programs and HR platforms. The businesses that sell these products represent around 50 per cent of the total enterprise software market, yet their share prices have fallen in line with the wider industry.

What we own in software today

While AI’s threat to software caught the market off-guard, we’ve been thinking about it for some time.

For the past 18 months we’ve treated traditional business software as riskier. Last year, we sold our holdings in companies such as Sprout Social, Xero and Blackline because we think there is a risk that AI could copy their features or increase competition.

The software companies we still own are ones we believe AI can’t easily replace, such as essential databases and critical workflow systems that businesses depend on.

One example is JFrog, which makes applications that help companies manage software development processes. JFrog’s management describes it as “the system of record for how modern software is built, secured, and deployed in the era of AI.”  Once installed, it benefits from customer lock-in as it’s expensive and time-consuming to switch to an alternative. In its third quarter 2025 results, the company reported total revenue up 26 per cent year‑on‑year (to $136.9m) and cloud revenue up 50 per cent (to $63.4m).

Another example is Amplitude, a company that develops digital analytics software which tracks how customers interact with a business’s website. It reported third quarter 2025 revenue up 18 per cent year‑on‑year (to $88.6m) and annual recurring revenue up 16 per cent (to $347m).

Similarly, Appian, which provides complex workflow automation to companies in regulated environments, reported third quarter 2025 cloud subscription revenue up 21 per cent year‑on‑year (to $113.6m) and total revenue up 21 per cent (to $187.0m).

These are not the numbers of declining businesses. JFrog, Amplitude and Appian may face strategic challenges down the road, but their fundamentals show no signs of the market’s concerns about disruption. Instead, they show that these companies are still delivering real value to customers.

On top of these examples, we are actively searching for quality software businesses where the recent sell-off has created an attractive entry price.

What if we’re wrong?

It is also important to recognise the asymmetry in the broader portfolio.

If AI is more disruptive to software than we currently expect, it is very likely to be positive for the capital-intensive infrastructure needed to run AI models: datacentres, power supplies, networking equipment, and specialised hardware.

More widely, investors are increasingly grappling with the scale of spending by the big tech companies, with US annual spend on AI software, computing power and datacentres now exceeding $1tn. Naturally, investors are demanding evidence that this investment will pay off.

As investors who are precluded from holding the tech mega-caps due to their size, this debate doesn’t directly affect us. That said, we believe that computing power will remain a rare and valuable commodity for the foreseeable future, given the proliferation of AI models, which are themselves using increasing amounts of data at an unforeseen rate.

Yet we are clear that these massive spending commitments should result in material opportunities for many of our current holdings.

Take IREN, the emerging cloud provider, which has signed a five-year agreement worth about $10bn to supply Microsoft with large-scale AI infrastructure built in Texas using Nvidia chips. Microsoft paid 20 per cent upfront, and the deal is expected to generate around $2bn in revenue each year.

Similarly, Astera Labs is poised to benefit. Its technology links graphics processing unit (GPU) chips together to increase the rate that data is transferred. This is crucial because modern AI models are less limited by raw computing power than they are by data bottlenecks, which Astera’s technology helps to eliminate.

Its share price rallied in early February, not because of its own results, but as result of Meta, Alphabet, Microsoft and Amazon’s combined commitments to spend $650–700bn on AI-related capital expenditure in 2026.

Key takeaways

Recent weeks have shown once again that markets react to news stories faster than they assess fundamentals.

The change in how software gets built is genuine. We must recognise this and adjust our mental models accordingly. It does not, however, automatically impair every software business model.

Our role is to remain discerning: reducing exposure to companies that are most threatened by AI and instead focusing on companies that are less vulnerable, such as systems of record and regulated workflows where replacement is slow and risky – while also taking advantage of short-term volatility to search for mispriced long-term opportunities.

Finally, we’re not putting all our eggs in one basket. If AI proves to be more disruptive sooner, the portfolio’s long-term AI infrastructure holdings should benefit from the same forces unsettling parts of software today.

Important information

This communication was produced and approved in February 2026 and has not been updated subsequently. It represents views held at the time of writing and may not reflect current thinking.

This article does not constitute, and is not subject to the protections afforded to, independent research. Baillie Gifford and its staff may have dealt in the investments concerned. The views expressed are not statements of fact and should not be considered as advice or a recommendation to buy, sell or hold a particular investment.

Baillie Gifford & Co and Baillie Gifford & Co Limited are authorised and regulated by the Financial Conduct Authority (FCA).

Some of the views expressed are not necessarily those of Baillie Gifford. Investment markets and conditions can change rapidly, therefore the views expressed should not be taken as statements of fact nor should reliance be placed on them when making investment decisions.

Baillie Gifford & Co and Baillie Gifford & Co Limited are authorised and regulated by the Financial Conduct Authority (FCA). The investment trusts managed by Baillie Gifford & Co Limited are listed on the London Stock Exchange and are not authorised or regulated by the FCA.

A Key Information Document is available at bailliegifford.com.