Governance went from rare to nearly universal in one year
NASCIO's 2026 State CIO Survey, now in its 17th year and drawing responses from 51 state and territory chief information officers, documents one of the fastest governance shifts we have seen in any public or private sector technology survey: 98 percent of states now have formal enterprise AI policies and procedures in place, up from 76 percent just a year earlier. Procurement has moved in step, with states writing generative AI terms and conditions into vendor contracts jumping from 41 percent to 61 percent over the same period.
That speed is notable precisely because state government technology rarely moves this fast. Budget cycles, procurement law, and interagency coordination typically slow adoption curves that move quickly in the private sector, yet AI governance closed a 22 point gap in twelve months. The explanation most state CIOs would likely give is pressure rather than enthusiasm: agencies were deploying AI tools with or without policy, and governance had to catch up before exposure became a public incident.
Agentic AI is the technology CIOs actually expect to matter
Sixty four percent of state CIOs named agentic AI as the technology most likely to reshape their work over the next two to three years, compared with just 14 percent who said the same of generative AI. That is a striking reallocation of attention for a sector that spent the last two years building generative AI pilots, and it suggests state CIOs view the generative AI phase as largely settled rather than still emerging.
Current agentic AI deployments remain concentrated in lower risk, internally facing categories: internal workflow automation, IT operations, application development, and customer service. None of those use cases put agents in direct control of citizen facing decisions yet, which tracks with a sector where the cost of a visible AI failure is measured in public trust rather than just revenue, and where governance teams understandably want evidence from internal use before extending agentic AI outward.
Funding has not caught up with ambition
Only 55 percent of surveyed states report having dedicated funding for citizen digital services, meaning close to half are running digital transformation efforts without a reliable funding line to support them. CIOs cited budget uncertainty, fragmented funding sources, reliance on one time grants, and demand that consistently outpaces available resources, a familiar list to any CIO in any sector but one that lands harder in government, where raising prices or cutting services is rarely an option.
Hawaii CIO Christine Sakuda, who administers the state's Office of Enterprise Technology Services, put the tension plainly: her state has "a ton of funding, but I think not enough" to fully support digital services, and she is prioritizing a strategic plan for a unified digital experience across government services. That is the honest version of a budget conversation most CIOs recognize, where the absolute dollar figure sounds adequate until it is divided across the full scope of what citizens now expect.
Infrastructure is shrinking and expanding at the same time
Forty three percent of states are downsizing or right sizing their own data centers, while 73 percent plan to expand public cloud use and 55 percent are building hybrid models that blend state owned infrastructure, public cloud, and third party providers. Those numbers describe a sector moving away from the binary choice between owning infrastructure and renting it, toward a deliberately mixed model shaped by cost, workload sensitivity, and the specific compliance requirements that come with handling citizen data.
NASCIO executive director Doug Robinson noted that the public response to data center decisions has sometimes reached a "ridiculous level of response" from citizens, a reminder that infrastructure choices in government carry a political dimension private sector CIOs rarely face. A data center closure or cloud migration that would pass without comment in a private enterprise can become a local news story in state government, which adds a communications burden on top of the technical one.
The role itself is being redefined, not just the toolset
Texas CIO Tony Sauerhoff, who leads the state's Department of Information Resources, told NASCIO that "modernization" no longer captures what the job has become, describing it instead as closer to digital transformation than modernization, with responsibilities expanding well beyond traditional IT operations into policy, workforce strategy, and citizen experience design. Sixty eight percent of surveyed CIOs agreed in substance, saying they expect more turbulence and disruption to the role over the next five years as the scope of the job keeps widening faster than the title or the org chart around it.
One unnamed CIO's line captures the current mood better than any statistic: "The threat is high and the defenders are thin." That sits alongside protecting critical infrastructure, which remains a high priority for most respondents even as cybersecurity competes for the same limited budget and staff that AI governance, cloud migration, and citizen service modernization are all drawing from simultaneously, leaving state CIOs stretching the same headcount across a growing list of roles that used to belong to separate offices entirely.
What private sector CIOs can take from a public sector survey
The gap between state government and private enterprise technology maturity has narrowed in ways this survey makes concrete. A sector that moves slowly by design closed a 22 point AI governance gap in one year, which should reset any private sector CIO's assumption that public sector technology lags theirs by default. The accessibility finding, just 35 percent of states help local governments with digital accessibility, is the kind of unglamorous gap that persists precisely because it never generates the urgency that AI governance did.
The transferable lesson sits in the funding mismatch more than the technology choices. States built governance and cloud strategy faster than they built the budget lines to sustain either, and plenty of private enterprises are repeating that pattern with their own AI rollouts: writing the policy before securing the recurring budget needed to operate it well past the pilot stage. A CIO in any sector reading this survey should treat governance and funding as a single line item to plan, not two separate ones to solve in sequence.
