AI’s Next Chapter Depends on More Than Building Data Centers | CRTV Episode 142

AI infrastructure is expanding, but the harder questions are about where capacity should live, how organizations will get value from it, and whether they can govern it effectively.

AI investment is entering a more demanding phase. The question is no longer simply how many data centers the industry can build. It is whether those investments can be executed, matched to real workloads, and translated into business value.

That was a central thread in CRTV Episode 142, which brought together a discussion of data center development, a cloud optimization case study, and an enterprise technology briefing.

Data center demand remains—but execution is the test

The discussion of AI data centers focused less on whether financing is available and more on the challenges of building and operating the facilities. Equipment supply, networking, cooling, electricity, staffing, and site construction all create potential bottlenecks.

The panelists also debated whether political resistance and a slowdown in construction could curb future capacity. Larry Dignan argued that some announced projects were unlikely to be deployed as planned, and that a pause could help the market avoid overbuilding. Others emphasized that enterprise AI adoption is still early: many organizations have yet to determine how to use AI effectively, while future workloads—including inference and robotics—could drive substantial demand for compute.

The market may change shape as it grows. The conversation suggested that large public cloud providers could gain pricing power if construction slows, while some neoclouds may survive by building closer relationships with enterprise data. Rather than supporting hundreds of providers, the market could consolidate to a smaller group in each category.

That makes the central question not just how much capacity will be built, but where it will be located and which workloads it will serve. Data access, infrastructure economics, and the ability to execute will all matter.

Cloud optimization can fund transformation

In an interview, Parag Jain described how Picaro re-optimized its AWS environment over a 12-month period. The work included reducing end-of-life support costs, resizing and consolidating infrastructure, and aligning uptime requirements with where the company’s fleet operated and where customers needed service.

Jain said the initiative generated about $500,000 in savings, but framed the project as more than cost reduction. Picaro redirected the savings toward AI, security, and enterprise platforms—funding transformation without making a separate request for AI dollars.

The approach followed a practical cycle: discover, analyze, act, and measure. It also included challenging vendor recommendations and comparing alternatives, rather than accepting the first proposed solution. For Jain, this helped move IT from a reactive cost center toward a more strategic role in the business.

One example was the rollout of Microsoft Copilot across the company’s Microsoft 365 environment. Jain reported 88% utilization across the suite. Looking ahead, Picaro’s priorities include centralized governance for AI agents and reducing shadow AI—so the organization can observe and manage what its agents are doing.

Productivity gains need a destination

The board update highlighted a gap between AI investment and realized value. It reported that technology investment rose 20.6% year over year between the third quarters of the previous and current year, with much of the spending going toward infrastructure and some toward services.

At the same time, only 13% of CEOs reportedly said their organizations had achieved significant or transformational value from AI implementations. The update also emphasized a practical budgeting challenge: organizations need clear rules for stopping initiatives that are not working and reallocating money as conditions change.

Productivity presents a similar challenge. AI may release employee capacity by automating processes and transactions, but the update said only 21% of that released capacity is being redirected toward innovative work. The gains matter most when organizations decide where to put that capacity—whether toward growth, better service, or greater resilience.

Governance must keep pace with deployment

The board update also pointed to a gap between plans for agentic AI and organizations’ readiness to manage it. It reported that 71% of organizations expect to deploy agentic AI in the next 12 months, while only 21% have governance capabilities to manage those deployments.

Cyber resilience raises another concern: the update said 84% of CIOs expect cyber budgets to increase, but only 40% have a fully operational cyber continuity plan. It argued that governance and continuity need to be built alongside deployment—not added after systems are already in use.

The priority: build selectively, measure value, and stay adaptable

Across the episode, the message was not that AI infrastructure investment will stop. Rather, the next phase will test whether companies can make better choices about what to build, what to fund, and what to change when results fall short.

For technology leaders, that means tying infrastructure decisions to real workloads, treating optimization as a source of investment, directing productivity gains toward new capabilities, and ensuring governance and resilience keep pace with AI adoption.