Across the United States, Europe, and the Gulf, households are beginning to feel the effects of the artificial intelligence boom through higher electricity bills. Regulators warn that the energy consumption of data centers, especially those powering generative AI workloads , is now influencing both policy and pricing. Reporting from the Financial Times shows how data-center growth is reshaping national energy strategies, with some countries reconsidering grid investments and industrial policy to support rising AI demand.
(Financial Times)
Governments frame new data centers as catalysts for economic expansion: more jobs, more investment, stronger digital competitiveness. But underneath the policy language sits a quieter and often invisible constraint the thousands of workers who wait, sometimes for weeks, to begin the jobs they were hired to do because onboarding and compliance workflows move slower than project timelines.
This problem is almost never discussed, even though it determines whether AI infrastructure gets built on schedule.
A growing body of analysis describes the tension. Research by McKinsey calls this the “data-center balance,” noting that the global race to build AI capacity comes with tradeoffs: energy consumption, water strain, land availability, and long-term exposure to single-sector risk.
(McKinsey)
What the models rarely measure is labor activation — the time between when a worker is hired and when they can legally and operationally begin earning wages.
Across data center buildouts, from Northern Virginia to Ireland, the Netherlands, Quebec, Singapore, and the Gulf states employers describe the same bottleneck:
Workers arrive on site.
Workers complete safety training.
Workers pass background checks.
Workers clear technical skills tests.
Then they wait.
Sometimes the delay is one missing document, an incomplete police clearance from another country, or a verification step still handled manually inside an outdated HR system. Other times, the delay is unexplained. But the outcome is always the same: no work, no pay.
For migrant workers rotating across Europe, the Gulf, and Asia, this waiting period can become financially destabilizing. They continue paying for lodging, food, transport, and in many cases loans taken out to secure employment. Employers meanwhile face schedule overruns. Governments fall behind on productivity targets tied to national AI and industrial strategies.
The public rarely sees this friction, even as they experience the downstream effects through rising energy costs.
The strain is especially evident in high-growth AI construction hubs where permitting, grid planning, and workforce onboarding are all running at full capacity. Ireland and the Netherlands are reassessing data-center limits.
(Irish Department of Environment: Grid Review 2025)
(Netherlands Data Centre Policy Update)
Singapore has introduced additional controls to manage power and land allocation for new digital infrastructure.
(Singapore MTI / Data Centre Policy 2024)
And Gulf states are simultaneously expanding AI zones and addressing labor-market vulnerabilities tied to cross-border recruitment.
(Gulf Labour Markets & Migration Programme)
Yet in all of these policy conversations, worker activation goes unmeasured.
This is where compliance-focused workforce platforms have begun intervening. Companies like FirstWork, led by co-founders Vardhan Kapoor and Shubham Choudhary, build systems that shrink the waiting period between “hired” and “on the job.” Their tools automate identity checks, document validation, contract verification, and cross-border compliance workflows that traditionally require manual review.
Kapoor has described these delays as “an invisible tax on workers.” Choudhary notes that “most onboarding failures aren’t about skill; they’re about fragmented systems that cannot keep up with the speed of modern infrastructure.”
These solutions reveal what a more efficient labor market could look like:
Real-time verification.
Portable, machine-readable worker files.
Cross-border documentation analysis.
Audit trails that allow employers and regulators to see exactly what is missing — and why.
Faster activation helps employers meet project timelines, reduces wage insecurity for workers and gives governments clearer visibility into the real economic cost of the AI buildout.
The global expansion of data centers will continue accelerating, and the strain on utilities will remain part of the public debate. But policymakers cannot make informed decisions without addressing the human side of the equation.
AI infrastructure is not built by algorithms. It is built by people navigating systems that were never designed for projects moving at this speed.
Before governments commit billions more to power and land allocation for AI facilities, they must recognize and measure the earliest bottleneck in the chain: the one that prevents workers from starting work at all.
