The CIO’s existential moment: Sovereign AI, boardroom relevance and the end of “steady state”

Gartner has long argued that as much as 80% of IT budgets are locked into sustaining systems, people and processes. Most CIOs will acknowledge that historical KPIs, operational commitments and legacy obligations have limited their ability to adapt, innovate and deliver exponentially greater value year over year.

In a stable world, that might be acceptable.

But the world in front of us is not stable.

The notion of “steering a steady ship” is now diametrically opposed to the realities enterprises face. Agentic AI and generative AI will add an estimated $17 trillion to global GDP in less than three years. This is an existential moment for every enterprise, every CEO and every CIO.

In fact, 95% of enterprises globally say they want to build their own AI and data platforms within the next three years—or roughly 800 working days.

Jamie Dimon, CEO of JPMorgan Chase, didn’t mince words about what this shift looks like in practice:

“We took AI/Data out of Technology. It’s too important. We put it at the management table. The woman running AI & Data now reports to me and our President … [because] there will be no job, no function, nothing that won’t be affected by [AI].”

This shift from technology-led to leadership-led illustrates the speed and the stakes of the revolution underway. What JPMorgan Chase is doing today, many banks, insurance providers and financial services firms will follow tomorrow.

And it raises a critical risk. The CIO and CTO can easily become trapped in a “maintain and sustain” mode while the rest of the organization charges forward with AI-based workloads, new agentic applications and accelerated use-case deployment.

The winners are pulling ahead

Enterprises winning with agentic AI and GenAI today are seeing dramatically stronger returns. They are achieving five-times the ROI of their peers and running two-times the agentic and GenAI workloads in mainstream production.

“The enterprises succeeding with their GenAI and agentic AI have chosen a very deliberate architecture for success: sovereign by design, built on open source (Postgres®), secure, agile, compliant and with a clear blueprint for how they measure and manage,” said Quais Taraki, CTO at EnterpriseDB (EDB), a leading data and AI platform company. “The CIOs and CTOs in these organizations are moving from steady-state orchestrators to agentic architects for their AI and data platforms—and enablers for the success of the whole enterprise.”

These findings hold consistently across global economies, based on primary research involving enterprises with more than 500 employees. The 13% who are pulling ahead with these early gains are doing four things differently. And in every case, the CIO and the CTO play a critical role.

Factor one: Sovereign AI and data infrastructure requires CIO and CTO leadership.

More than 93% of the most successful enterprises have maintained a mission-critical commitment to sovereignty over their AI and data. This means access anywhere and anytime, secure by design, compliant by default and agile at scale.

Sovereign AI and data isn’t a technology preference. It’s a leadership position. This is how CIOs/CTOs stop being downstream service providers and start being enterprise co-architects again. The winning organizations are making deliberate, foundational moves to build sovereign AI and data infrastructure now. If sovereignty is real, it can’t be built at the department level.

Factor two: Escape the single-cloud trap—or get trapped in the past.

You cannot be sovereign if your data is beholden to a single cloud ecosystem.

Your AI models, agentic workflows and LLM-powered applications must operate close to your data. Keeping AI and data siloed may work during experimentation, but production-grade agentic AI requires proximity to operational data in a secure, compliant and agile architecture.

That reality pushes enterprises toward hybrid and multi-environment architectures where the enterprise—not the provider—sets the rules. Leaders that fail to evolve beyond a single-cloud dependency risk being trapped in yesterday’s operating model.

Factor three: Build width, not density, to create the AI flywheel effect.

We examined 15 agentic and GenAI workloads across 18 departmental functions. These patterns emerged:

The winning 13% take a pragmatic approach to ROI expectations. They rarely expect performance or productivity gains above ~15% in any one function. But they consistently beat that across 10 or more mainstream agentic and GenAI deployments.

Their CIOs and CTOs set realistic expectations and avoid the over-ebullience seen in enterprises that fail to convert pilots into measurable outcomes.

Once ROI targets are reached, these leaders expand deliberately into additional use cases—carefully designing systems to create the “agentic flywheel effect,” in which momentum builds and adoption accelerates.

Factor four: Measure ROI in multiple ways—then move on (there’s a ceiling).

The winning 13% measure success across a broad range of ROI outcomes, including:

OpEx reduction
Margin improvement
Revenue growth
Speed of innovation
Market share expansion
Confidence in outcomes
Organizational confidence and adoption

They outperform by nearly 50% their peers focused on only one or two use cases and only one or two metrics.

Winning CIOs and CTOs understand there is a ceiling to value in any single use case. They know how to define the shape of success, balance outcomes and guide the organization through a deliberate and scalable path forward.

From steady-state orchestrator to agentic architect

The CIO and CTO roles are being rewritten in real time—transforming from operational guardians into agentic architects of the enterprise’s AI and data platforms. Your decades of experience—security, reliability, governance, architecture, discipline—are suddenly the accelerant, not the anchor.

But only if you step back into the game.

The trap is clear: You can spend the next 800 working days maintaining yesterday’s commitments, while your competitors use that same time to rebuild how work happens.

If AI and data are now existential, then the CIO/CTO mandate is existential, too:

Build a sovereign AI + data foundation.
Design for hybrid reality (not single-vendor dependency).
Scale width to create compounding returns.
Measure value broadly, and move fast once ceilings appear.

Claim your AI and data sovereignty now. Choose your ROI. And move back into the boardroom, where the future is being decided. Take 15 seconds to test yourself to see if you are on the right path. Then get these free blueprints for success here.