We’ve talked a lot recently about big tech under threat from AI. It’s a tricky thing to visualise given that big tech companies are at the forefront of this kind of innovation. But this is what it looks like in reality.
Morningstar in a research note published this week has downgraded the long-term competitive advantages, or “economic moats”, of several major US software companies, arguing that the rapid advance of artificial intelligence makes the sector’s future less certain than it looked a year ago.
The research, led by senior equity research analyst Dan Romanoff, says the software industry still looks healthy in the medium term. Retention rates remain strong, margins are rising and enterprise customers are not rushing to replace their existing software stacks.
But the pace of technological change means analysts are less confident these companies will generate excess returns over the next decade with the same certainty as before.
Uncertainty rises as AI reshapes software
Morningstar has shortened the period during which it expects software companies to sustain strong competitive advantages, or their “moats”.
Romanoff said the firm has cut its moat duration assumption to 10 years from 20 years, reflecting uncertainty around how AI could reshape the economics of the industry.
“The future of software is less clear today than it was a year ago,” Romanoff said.
“We no longer think software companies will earn excess returns over the next decade with near certainty.”
The firm still views excess returns as probable over the next decade, but it has less confidence beyond that horizon.
Large platforms still hold an advantage
Morningstar argues the core enterprise platforms that act as “systems of record” should remain strategically important.
Companies such as Salesforce and ServiceNow already sit at the centre of customer relationship management and enterprise service workflows. That position gives them an advantage in distributing new AI tools across large installed customer bases.
Romanoff expects these platforms to evolve as AI becomes more widely adopted, with software vendors gradually shifting toward consumption-based pricing tied to AI usage.
Salesforce, for example, could capture new revenue if customers increasingly deploy AI agents inside its customer relationship and service platforms.
For now, the underlying data still supports strong competitive positions across much of the sector. Retention rates remain stable and measures such as annual recurring revenue and remaining performance obligations continue to grow faster than revenue.
Margins have also continued to rise despite concerns that AI computing costs could weigh on profitability.
AI still a small slice of software revenue
Despite the hype surrounding generative AI, it still contributes a small share of revenue for most software vendors.
Public disclosures from listed software companies suggest AI-related products account for about 1% to 5% of revenue or annual recurring revenue.
Romanoff said many management teams remain cautious because of the risk of hallucinations or autonomous systems making mistakes.
At the same time, AI-focused companies such as OpenAI and Anthropic are seeing rapid revenue growth, particularly from consumer subscriptions.
Is the bear scenario overhyped?
Morningstar believes some of the more extreme fears about AI disrupting the software sector are exaggerated.
There is not enough computing capacity today to replace large numbers of human workers with AI, and Romanoff does not expect that to change within the next five years.
The size of the global software market, about $1.4 trillion annually, also makes rapid disruption unlikely.
From a practical standpoint, most large enterprises cannot realistically replace their entire software stack and business processes within a few years.
Romanoff also pushed back on concerns about so-called “vibe coding”, where AI tools allow developers to quickly build software applications.
While developers may be able to produce applications quickly, enterprises still rely on proven platforms that are reliable, secure, supported and constantly improved.
Open-source software offers a useful precedent. It was once expected to undermine proprietary software vendors because it was free and rapidly evolving.
Instead, many proprietary platforms continued to grow by offering integrated systems, documentation, customer support and large user communities.
The most realistic long-term impact of AI may be on software seat growth.
As workers become more productive with AI assistance, companies may eventually need fewer licences per employee. That would slow growth for vendors that rely heavily on per-user pricing.
Morningstar expects any reduction in seat-based revenue to be offset, at least partly, by new consumption-based AI services.
