Concerns that artificial intelligence will rapidly supplant software engineers and hollow out the IT services industry are, in the view of Bernstein analysts, overstated. The firm signals that, to date, actual adoption of AI coding tools remains constrained and the productivity gains reported so far are modest.
Bernstein highlights that writing code represents under 20% of a typical software engineer's responsibilities. The remainder of engineers' time is devoted to tasks such as documentation, deployment, systems integration and management of development processes. In that light, even if AI systems can generate code more quickly than humans, they would only automate a minority of the overall job.
Survey data referenced by the analysts indicates that nearly 80% of developers now use AI tools in some form. However, a substantial share of those users report that code produced by the tools still needs refinement and debugging before it can be deployed. That gap between initial generation and production-ready code dampens the immediate disruptive potential of these systems.
Layoffs in the technology sector have remained elevated since the post-COVID hiring surge, with more than one million workers cut worldwide since 2020. Bernstein attributes much of this workforce reduction to over-hiring during the hiring boom and weaker demand conditions. The firm notes, however, that sectors tied to AI have accounted for a higher proportion of recent cuts. Some companies may have trimmed headcount partly to signal AI adoption, yet Bernstein emphasizes that the tools themselves are not yet substituting for teams at scale.
The analysts explicitly reject comparisons between the current AI environment and the late-1990s dot-com bubble. They point out that leading AI companies today are established, profitable firms that are investing from robust cash positions, rather than speculative startups dependent on speculative capital.
Looking ahead, Bernstein expects AI to layer incrementally into enterprise technology stacks rather than to function as an immediate replacement for human knowledge work. "The current market reactions perfectly capture two fallacies - the slippery slope and the bandwagon, and once the dust of the AI storm settles, we perceive a correction to emerge," the analysts write. They add that, while longer-term disruption cannot be ruled out, current market sell-offs in AI-exposed sectors imply an economic shock that is not yet visible in corporate behavior or macro data.
On the investment side, Bernstein suggests that the heightened fear in markets can create buying opportunities. "For the investors, periods of heightened fear offer opportunities to buy beaten down sectors," their note concludes.
This assessment frames AI as a technology that will likely augment enterprise capabilities over time, with immediate effects concentrated in specific coding tasks rather than across the full spectrum of engineering functions. The trajectory and timing of broader disruption remain uncertain, and market pricing appears to have moved faster than underlying corporate actions.