AI’s enterprise moment could belong to IT services

The recent merger between ITC Infotech and Happiest Minds may foreshadow a consolidation among mid-market IT service providers as they navigate the challenges posed by AI-driven disruption. The market value of leading global firms has plummeted by nearly 25% since early 2026, blurring the landscape. Moreover, the trading multiples of major IT companies now resemble those of low-end staffing firms.

While investing in AI-centric capabilities is crucial, the industry needs a solution to its own problem and scale alone may not provide it. The excitement surrounding large language models (LLMs) and generative AI (GenAI) is palpable, but the real story lies in how AI gets embedded within global firms.

Partnerships between front-runner model providers such as OpenAI and Anthropic and IT giants like TCS, Accenture, EPAM, Cognizant, and DXC indicate a growing realization of this need. Enterprises are grappling with an estimated $1.5 trillion in technical debt, while legacy system maintenance consumes around $500 billion annually. This burden, however, also presents an opportunity.

AI's initial significant contribution may not be glamorous but rather optimizing and automating the legacy layer. This could generate efficiencies that fund the next wave of AI adoption. Nevertheless, the IT sector must adeptly manage the resulting revenue contraction and redistribution to meaningfully participate in new value pools.

AI is an ecosystem, with value generated across various levels - from silicon to data infrastructure, from energy generation to foundation models, and ultimately to enterprise applications. While early discussions centered on foundation models, sparking concerns that application software might become commoditized, this is not the end of the line for enterprise software. As AI migrates into platforms, some applications will become thinner, while others will become richer by embedding AI deeply into workflows.

The AI-native enterprise is likely to evolve around four interconnected systems: engagement, agents, work, and context. Each of these layers will require tailored solutions, and IT partners have historically played a crucial role. As AI software becomes probabilistic and capable of self-improvement, the question arises: can AI-native systems coexist with legacy IT infrastructure designed for human operators?

The technology behind autonomous vehicles has been available for years, but adoption has been hindered by the challenge of running manual and autonomous vehicles simultaneously. Similarly, AI-native systems may struggle to reach their full potential while operating within workflows tailored for human decision-making. This could necessitate a complete reengineering of the systems stack, presenting new revenue opportunities for tech-services companies.

The evolution of Software as a Service (SaaS) provides a useful precedent, with most successful SaaS companies emerging in three categories: universal needs (email, calendars, collaboration, search), new consumer categories (ride-sharing, food delivery, home-sharing), and B2B specialization (deep vertical or functional expertise).

AI appears to be following a similar trajectory, with potential opportunities in domain-specific models, industry-specialized applications, and intelligent systems deeply integrated into business processes. As with SaaS, this layer is unlikely to yield a single winner but rather numerous companies serving specialized markets. The trajectory of IT companies will increasingly hinge on enabling these domain-specific AI solutions, from AI agent-driven loan processing to autonomous cyber defense.