Enterprises Will Spend $2.5 Trillion on AI This Year. Most Can't Turn It Into Revenue.

Global AI investment is on track to reach $2.5 trillion in 2026, up 44% from the previous year, according to an MIT Technology Review Insights report published this week. Model capabilities are advancing faster than most organizations can absorb, and the cost of AI performance keeps falling. And yet: most enterprises are still not growing revenue through AI, the report says.

The structural problem

The report's diagnosis is blunt. The scaling problem isn't technological — it's structural. Many enterprises have invested heavily and ended up with fragmentation: AI accumulates in silos, with sales agents unaware of open support tickets and marketing systems personalizing content without visibility into what finance already knows about a customer. Each function may perform well in isolation, but the company as a whole learns little. As the report puts it, intelligence that can't flow across the organization isn't intelligence — it's overhead.

The process-first playbook

What separates the companies generating sustained returns, according to the report, is a shared discipline: they treat process redesign as the work that comes before model selection. These "process-first" companies rebuild their operating models for how the technology will evolve, rather than retrofitting roles and workflows after deployment. The shift from AI as a tool to AI as an operating model — what the report calls the "agentic shift" — starts with architecture, not algorithms: rebuilding data infrastructure for accessibility rather than volume, and replacing fixed tech stacks with composable ones that can evolve as models change.

Data readiness, not data abundance

A second finding worth attention: having data and having AI-ready data are very different things. Most enterprises discover that distinction too late. The report argues for what it calls a sovereign, composable foundation — querying and preparing data where it resides, without migration or centralization. That matters because data residency laws, multicloud environments, and structural complexity are making centralization increasingly impractical. AI sovereignty — where models run, who controls them, how they operate across borders — is becoming a board-level question.

Why it matters

The $2.5 trillion figure makes the headline, but the real message is about management, not money. Enterprises that treat AI as a procurement problem — buy the model, deploy the tool — are the ones stuck with impressive demos and flat revenue. The ones pulling ahead redesigned the company first. For investors, the signal is that the AI trade's next winners may not be model builders at all, but the enterprises disciplined enough to reorganize around them.

Written by the BidAsk Newsroom. Based on reporting from the sources below.