Global Data Center Capacity Set to Nearly Double by 2030

Global data center capacity is projected to nearly double by 2030, with almost 100 gigawatts of new capacity coming online, according to JLL’s 2026 Global Data Center Market Outlook. Investment in facilities and the computing and networking equipment they house could approach $3 trillion.

The forecast reflects growing demand for cloud services and artificial intelligence. But the shift is about more than the volume of computing power businesses need. As AI moves from analyzing information to taking action, companies also face more demanding questions about the reliability of their data and the systems using it.

From training models to running applications

JLL expects AI to account for around half of data center workloads by 2030, compared with approximately a quarter in 2025. It also forecasts that inference, the process of running trained models to generate responses or perform tasks, could overtake training as the main source of AI demand in 2027.

For infrastructure providers, this changes where capacity is needed. Applications serving users in real time require low latency, encouraging more geographically distributed facilities. Power availability is another constraint: JLL reports that average waits for grid connections in major data center markets exceed four years.

These forecasts depend partly on the emergence and adoption of AI applications that have yet to reach scale. They signal substantial anticipated demand, rather than guaranteed growth.

When AI acts on business data

For companies deploying AI, infrastructure is only part of the challenge. Systems that act on behalf of customers or employees make data quality directly relevant to the actions they take.

Speaking at the sixth hayaData conference in Tel Aviv, Guy Machlev, VP of R&D at Israeli fraud prevention company Riskified, described the transition from systems that primarily analyze information to agents that make decisions and execute actions. He said this introduces new challenges involving trust, identity and risk management, increasing the importance of accurate, real-time decisions.

Hayadata conference. Photo by Yair Sigron

Skills beyond the dashboard

AI is also changing who can work with data. Amit BenDor, Head of AI & Data Science at Unity, told the conference that employees across more roles can now build dashboards and perform tasks that once required dedicated specialists.

Greater access to those tools, he argued, makes deep domain knowledge more important: understanding the underlying problem, asking useful questions and drawing meaningful conclusions from results. Proficiency with AI tools across analytics, data engineering and data science is the other essential skill.

For business leaders, these developments raise practical questions: which decisions can be delegated to AI, what information those decisions depend on, and who has the expertise to assess the results. Access to computing capacity supports adoption; the ability to use it effectively also depends on data and judgment.