Enterprise AI spending on track for $2.5 trillion this year, yet most companies remain stuck in silos

Global enterprise spending on artificial intelligence is projected to reach $2.5 trillion this year, a 44% increase over 2025, but the surge in budgets is not translating into business value. According to a report from MIT Technology Review Insights produced in partnership with Uniphore, most organizations are still failing to grow revenue through AI or fundamentally change how they operate. The bottleneck is not the models — whose capabilities are advancing faster than most enterprises can absorb — but fragmentation: intelligence accumulates in silos, so sales teams cannot see open support tickets, and marketing systems personalize content without knowing what finance already knows about the customer.

The structural problem

The report frames the jump from "AI as a tool" to "AI as an operating model" as an "agentic shift" and argues that the primary obstacle is structural, not technological. Companies generating sustained return share a common discipline: they treat process redesign as work that precedes model selection, and they build for how the technology will evolve rather than retrofitting roles and workflows after deployment. For them, the agentic shift begins with the operating model, not with a benchmark of one LLM or another.

Data readiness versus volume

The second finding concerns the gap between data volume and data readiness. Most organizations discover too late that possessing data and possessing AI-ready data are entirely different things. The report proposes a sovereign, composable foundation — one that retrieves and prepares data where it sits, without migration or centralization — as the way to turn raw data estates into intelligence that AI agents can act upon. As data-residency rules, multi-cloud environments, and organizational complexity make centralization increasingly impractical, sovereign control over where models run and where data lives is what preserves that flexibility.

What the architecture requires

Three architectural shifts are identified as necessary conditions: first, rebuilding data infrastructure for accessibility rather than volume; second, replacing fixed tech stacks with composable architectures that can evolve as models and tools change; third, resolving questions of AI sovereignty — where intelligence runs, who controls it, and how it operates across organizational and legal boundaries. Without all three, the "agentic shift" remains a slogan without execution capability.

Report context

The content was produced by MIT Technology Review's custom content arm, Insights, not by the editorial staff, and was written by humans with any AI tools limited to production processes under human supervision. Uniphore is the commercial partner behind the project.