Federal CIOs are navigating a critical moment. The push for artificial intelligence (AI) is intensifying just as cost-cutting mandates tighten budgets and infrastructure demands mount.

MeriTalk recently sat down with Suri Durvasula, vice president, federal, Dell Technologies, to discuss the findings of the Tech Tonic: FY26 Federal CIO Forecast. Our conversation covered lots of ground: shifts in federal IT priorities, modernization while reducing costs, the foundation for effective AI implementation, and why industry should come together to accelerate government progress.

MeriTalk: The Federal CIO Forecast shows AI jumping to the No. 1 priority and infrastructure to No. 2, and cybersecurity remaining foundational. What’s driving the reshuffle, and what’s the biggest risk that could derail agencies’ progress in the new fiscal year?

Durvasula: Government has a mandate to modernize and drive efficiency at the same time. Federal CIOs are under pressure to improve service quality, streamline operations, and even shrink the government footprint. In the past, modernization didn’t always lead to efficiency, but today, technologies like agentic AI are helping close that gap.  That’s why infrastructure has become central – it’s the foundation that enables AI.

It’s interesting that cybersecurity has dropped to the third pillar, but I don’t see that as a downgrade. Cybersecurity surrounds everything we do with AI and infrastructure. It’s baked into the whole ecosystem.

The reshuffle is also being driven by AI mandates from the White House. The real challenge for government today is making the upfront investments needed to realize long-term benefits. Modernization savings aren’t instant, and if agencies lose sight of the mission outcomes, whether that’s better citizen services or improved national security, they risk getting derailed. Technology must align with mission outcomes. That’s where past modernization efforts have often fallen short.

MeriTalk: Despite its critical role in enabling AI, infrastructure modernization faces one of the largest leadership support gaps. What conversations and metrics should CIOs bring to secretaries and their deputies to close that gap and secure buy-in?

Durvasula: It’s a tough sell because IT investments often look like sunk costs. Secretaries care about mission outcomes, not server racks. CIOs need to tell a compelling story that links infrastructure investments to those outcomes.

Think about it like education. When we used to advocate for one-to-one student device programs, we didn’t talk about the hardware. We talked about improving graduation rates. That’s the shift we need in government. CIOs have to connect the dots between technology and the results that matter to their department’s leadership. AI makes this easier, but it’s still a skill CIOs need to develop: aligning IT to mission impact.

MeriTalk: With every CIO pursuing savings – and 80 percent under explicit savings mandates, according to the Federal CIO Forecast – which levers tend to deliver the most value without compromising modernization? What pitfalls should agencies avoid to ensure that short-term savings don’t lead to long-term costs?

Durvasula: Agencies need to evaluate their portfolios – their infrastructure and applications, and opportunities for automation. You might be running 4,000 applications when you really only need three critical ones. Rationalizing those assets can unlock savings that can then be reinvested into AI or automation.

There’s no magic pot of money. Agencies have to find savings within their current budgets. That means setting up a modern, efficient IT “house” first. Without that foundation, you’re just layering on complexity.

App rationalization is a great example. Agencies may be maintaining dozens of bespoke applications when a commercial off-the-shelf product would do the job better and at less cost – plus, it’s likely more secure. Short-term investments here can lead to long-term efficiency and lower support costs. But if agencies skip that foundational step, they risk higher costs and weaker cyber postures down the line.

MeriTalk: According to the Federal CIO Forecast, nine in 10 CIOs support an enterprise approach to Federal IT. What practical steps, such as standard platforms, shared services, or enterprise contracts, can help agencies simplify operations while improving mission delivery?

Durvasula: The key is to break down the silos. Today, agencies have fragmented data centers and duplicate systems. Until you consolidate your data, rationalize your applications, and simplify your IT environment, you can’t unlock the full value of enterprise IT.

Think about application virtualization. It’s not just about reducing your physical footprint. It’s about creating a common platform that can scale. The compute power available today is incredible. A single server can now do the work of seven from just a few years ago. That means agencies can dramatically reduce their infrastructure, centralize data, and prepare that environment to run AI workloads efficiently.

The simpler and more unified your environment, the easier it is to deploy emerging technologies. That’s the path to true modernization.

MeriTalk: Cybersecurity dropped on the priority list on the new Federal CIO Forecast, but CIOs say they are embedding zero trust across the board. How can agencies integrate security into AI, data, and infrastructure upgrades so it’s built in from the start, not bolted on after the fact?

Durvasula: Fortunately, a lot of today’s technologies come with security built in, from hardware to cloud services to software. The focus needs to be on aligning those capabilities with the agency’s cybersecurity posture, especially their zero trust implementation.

When agencies buy commercial off-the-shelf or open-source products, they need to hold vendors accountable for integrating cyber principles. It’s not about layering security on top after the fact. It’s about making sure the tools you bring in are secure by design. Cyber should be a decision criterion for every procurement, up front.

MeriTalk: AI and infrastructure are deeply connected. What does an AI-ready infrastructure and data foundation look like for federal workloads, and how should CIOs sequence their investments to set their agencies up for long-term success?

Durvasula: Data is the fuel that makes AI go. If your data isn’t clean, accessible, and governed, you won’t get the outcomes you want from AI. Agencies need a data platform that spans cloud, on-premises, and edge environments. Without that, they’ll only get partial insights.

Governance is huge, especially in secure environments. Who can access which data matters a lot. And quality matters, too. Think about your personal computer: It’s full of junk and a few meaningful files. You want your AI to focus on the meaningful data, not the noise.

CIOs should start by building a data platform that covers all their environments and then layer on AI capabilities. Don’t get distracted by proofs of concept. Think enterprise-scale from the beginning.

MeriTalk: Where can industry make the biggest impact when helping CIOs with their 2026 priorities? And specifically, how are agencies leveraging offerings like the Dell AI Factory with NVIDIA to build secure, scalable AI environments that align with enterprise simplification and cost-efficiency goals?

Durvasula: Industry can make the biggest impact by lightening the lift for government. Too often, we compete against each other instead of coming together. That’s why we built the Dell AI Factory with NVIDIA, a validated platform that includes everything from hardware to services, where agencies can run multiple AI workloads within an environment.

We also partnered with Deloitte and Equinix to launch a solution called Silicon to Service, which is powered by the Dell AI Factory with NVIDIA. It’s housed in a secure Equinix environment and lets agencies test drive a compliant AI platform with built-in cybersecurity. They can explore it, validate it, and then replicate it next to our environment using a full bill of materials.

The goal is to give the government a blueprint for AI implementation that is secure, scalable, and supported by a team of partners.

When we work together, we can achieve a lot more and help the government move a lot faster.

Read More About
Recent
More Topics