The Great AI reckoning: when innovation becomes disruption
The stock market is experiencing a dramatic bifurcation. While traditional sectors surge – energy companies are up over 20 per cent and materials firms 15 per cent – a different story is unfolding in the technology and financial services sectors. Established players are watching their valuations crumble as investors grapple with the answers to a simple question: which companies will survive the artificial intelligence (AI) revolution?
The S&P 500’s near-flat performance this year masks a seismic shift beneath the surface. Energy and industrial stocks are rallying while software companies, financial data providers, and insurance brokers haemorrhage value. Some stocks have plunged 20-30 per cent year-to-date (YTD), and the most extreme have fallen 50 per cent.
By way of example, S&P Global (SPGI) dropped 25 per cent in a month after 2026 earnings guidance of $19.40-$19.65 missed the $19.96 consensus. The company now trades at 21 times consensus, its lowest multiple in five years.
Figure 1. S&P500 vs Software (1 Jan 2025 = 100)
Source: LSEG Workspace
The selloff accelerated when AI companies began unveiling tools, and plugins in particular, that undermine the Software-as-a-Service (SaaS) versions owned by established industry players. Each announcement triggered fresh waves of panic:
When advanced AI assistants demonstrated capabilities in legal research, non-disclosure agreement (NDA) triaging and contract writing, financial analysis, and sales automation, software companies saw their stocks crater. Application software firms dropped nearly 23 per cent, while systems software players fell 15 per cent. Dominant financial data providers such as Thomson Reuters and FactSet Research have seen their shares decline by 30-60 per cent as investors question whether AI can simply replicate their core services.
The contagion spread to asset managers and investment firms with significant exposure to software companies. With roughly US$25 billion in software loans trading below 80 cents on the dollar, firms like KKR, Blackstone, and Blue Owl watched their stocks decline 8-16 per cent this year alone – and by as much as 50 per cent from the most recent highs.
Next, it was tax-preparation AI tools that analyse uploaded documents and generate personalised tax strategies, raising alarm about AI’s potential to automate accounting and financial advice. H&R Block’s shares have halved. Advisory firms have slipped 7-10 per cent from their peaks.
The most recent blow came when AI chatbots began offering personalised insurance quotes within conversational interfaces. Major brokers like Aon and Arthur J. Gallagher have declined 10-18 per cent year-to-date (YTD), with some down nearly 40 per cent from their highs.
Adapt or die
This will all feel eerily familiar to anyone who remembers the internet’s arrival. Back then, retailers, media companies, and service providers faced similar existential questions. Some adapted and thrived (Walmart). Others emerged from obscurity to dominate (Amazon). Many, such as newspapers, suffered as classified revenue streams were eroded and audiences fragmented.
It has always been the case that new General-Purpose Technologies (GPTs) disrupt. As Table 1 shows, throughout history, general-purpose technologies have transformed economies by making entire industries obsolete.
Table 1. General Purpose Technologies (GPTs) and disrupted industries
|
Technology (Period) |
Industries that went under |
|
Steam Engine (1760s-1850s) |
• Hand loom weavers • Canal boat operators • Sailing ship builders • Traditional blacksmiths • Stagecoach services |
|
Electricity (1880s-1920s) |
• Gas lighting companies • Ice harvesting industry • Water wheel manufacturers • Telegraph operators (partially) • Manual elevator operators |
|
Internal Combustion Engine (1890s-1930s) |
• Horse breeders and stables • Blacksmiths and farriers • Buggy and carriage manufacturers • Livery stables • Railway short-haul freight |
|
Transistor/Semiconductor (1950s-1980s) |
• Vacuum tube manufacturers • Slide rule manufacturers • Mechanical calculator companies • Typewriter manufacturers • Analog switchboard operators |
|
Personal Computer (1980s-2000s) |
• Typewriter industry • Encyclopedia publishers (print) • Film-based typesetting • Video rental stores • Music stores • Bookstores (many independent) • Print newspaper classifieds |
|
Internet (1990s-present) |
• Dial-up ISPs • Video rental chains (e.g., Blockbuster) • Print yellow pages • Traditional print media • Landline telephone services • Traditional retail (many sectors) • Traditional taxi services • Traditional hotel booking |
Within the emergence of AI and associated disruption a pattern has emerged. That pattern suggests domain expertise may be the differentiator. AI will disrupt companies whose primary value lies in technical execution. But enterprises offering deep knowledge of specific industries – their workflows, regulations, and nuances – maintain defensible positions, for now at least.
For now, chief technology officers (CTOs) are unlikely to entrust mission-critical operations to unproven AI startups, no matter how impressive their demonstrations.
A valuation opportunity?
Where things get interesting is where the forward price-to-earnings (P/E) ratios have compressed dramatically. Software and data companies aren’t yet being priced as if the underlying business models are terminally broken, so the opportunity doesn’t slap you in the face, but…
According to Ed Yardeni, application software has fallen from 35x to 24x earnings, P/E ratios for systems software has dropped from 36x to 23x, financial data providers have seen their P/Es decline from 30x to 23x, Insurance brokers from 25x to 16x and asset managers have slipped from 19x to 16x.
For now, these companies are still projected to grow earnings by 8-20 per cent. Provided P/Es don’t fall further, and provided those earnings growth estimates prove correct, investors can make an 8-20 per cent Internal Rate of Return (IRR). The market is pricing in substantial uncertainty about whether AI competition will force downward revisions to those projections as contracts are renewed.
For investors who are confident the incumbents will successfully adapt, these valuations present opportunities. The risk, of course, is that the earnings estimates themselves prove too optimistic or that P/Es decline further.
The extent of actual disruption, however, will depend almost entirely on we deploy AI.
In financial markets, for example, AI threatens to commoditise information services while potentially opening new opportunities for companies that successfully integrate it with proprietary expertise.
Companies that view AI as a tool rather than a threat and are willing to experiment now with how to integrate it will be best positioned for what comes next. Those who assume their historical advantages grant immunity are today’s ostriches with their head in the sand. They will see their assumptions – and valuations – vaporise in real time.
The question isn’t whether AI will transform industries. It’s already happening. The question is: who adapts quickly enough to remain relevant when the transformation is complete?