EBITDA multiples across the technology services sector have compressed roughly 30% from their 2023 peaks. At the same time, firms across the sector are reporting real AI-driven productivity gains, from automated security operations to AI-assisted code generation to AI-first customer service. Productivity is rising. Valuations are falling. The gap points to something structural in how services firms turn productivity into profit.
The compression itself is temporary. For firms that redesign how they price and deliver, the opportunity ahead is larger. But the transition will not happen on its own, and firms that delay will lose ground before the expansion arrives.
The reason sits in the commercial model. Technology services still runs predominantly on time-and-materials billing: clients pay for the hours engineers spend, so revenue scales with headcount and billing rates. When AI tools compress the time required to deliver work, the firm delivers faster, bills fewer hours, and generates less revenue from the same scope. The productivity gain accrues to the client as a lower cost. The firm absorbs it as lower revenue. Across application development, testing, infrastructure management, and business process services, the same dynamic is playing out at different speeds.
The traditional talent pyramid compounds the problem. Built on a broad base of junior engineers performing volume work, its economics depend on the gap between what clients pay per hour and what junior staff cost. As AI-driven automation absorbs the work that justified the pyramid’s base layer, the foundation erodes, and with it, the margin mechanics the industry has relied on since the era of global delivery. The shift toward outcome-based contracts, IP licensing, and subscription-based commercial models is what allows AI productivity to translate into margin for the firm, not just savings for the client. Investors already understand this. In a recent diligence engagement, a leading growth equity firm evaluating a data and analytics services asset looked past top-line growth entirely. What they assessed instead was automation yield, IP and subscription revenue share, platform-spine maturity, and the mix of outcome-based pricing. The question underneath all of it: does this firm have a credible path from selling hours to selling outcomes?Â
Enterprise technology budgets are rotating, not contracting. The spend that traditional service lines are losing is moving into new categories, and collectively, they are much larger. The growth is organized around four areas:
Individual pockets within these four categories range from USD 30 to 300 billion. Zinnov estimates the total addressable opportunity at USD 700 billion to 1.2 trillion. Realistically, services firms are likely to capture USD 260 to 440 billion of that by 2030, with the expansion growing more meaningfully after that. The scale is already attracting new entrants: OpenAI’s expansion into enterprise services, and the growing role of foundation model providers in direct client engagements, signal that services firms are no longer competing only with each other.
At the center of much of this new spend sits the orchestration layer, the architecture that will govern how AI agents interact across enterprise systems. That layer is already crystallizing around platform vendors like Microsoft, Salesforce, ServiceNow, and Google. For services firms, the message is straightforward: become the firms that design, implement, and govern these orchestration architectures for enterprise clients, before the window closes.
The firms moving early are making four interconnected bets.
These four bets do not land all at once. In our work with technology services leadership teams, transformation succeeds when it is sequenced so that each step funds and enables the next. Skip a step, and the one after it stalls.
Three player archetypes are now converging on the same expansion opportunities: traditional services firms repositioning around AI, platform vendors building services capabilities, and AI-native entrants built without legacy constraints. Based on Zinnov’s research, the window for traditional services firms to establish defensible positions is roughly 18 to 24 months. First movers are already converting early engagements into multi-year client relationships. The net outlook for the industry is positive: the expansion is larger than the compression. But the gains will not spread evenly. They will go to the firms that get the operating model, the commercial architecture, and the domain positioning right.
That kind of timeline is not new for this industry. Technology services has reinvented itself roughly once a decade: mainframes and hourly consulting in the 1970s, client-server and outsourcing in the 1980s, ERP and global delivery in the 1990s, managed services in the 2000s, cloud and digital studios in the 2010s. Each cycle rewarded a different capability.
This one will reward the firms that can sit at the intersection of deep domain knowledge, software-like economics, and the ability to design and govern the AI architectures that enterprises will run on for the next decade. Three archetypes are racing toward that intersection. Not all of them will get there.
THANK YOU
FOR YOUR RESPONSE
