The sovereignty question is getting a real answer

For several years, government technology buyers — particularly in Europe and in regulated US agencies — have raised the same objection to hyperscalers: moving to the cloud means losing control of sensitive data. Google's response, until recently, was largely contractual.

That has changed. Google Cloud now offers what it calls a Sovereign Cloud portfolio with three distinct tiers. Google Cloud Data Boundary controls where data is stored and processed. Google Cloud Dedicated provides isolated infrastructure. Google Cloud Air-Gapped goes further, cutting off external network access entirely. Government CIOs no longer have to choose between cloud capabilities and compliance posture — they choose a tier that matches their actual risk classification.

The five commitments published alongside this portfolio — local infrastructure investment, digital resilience, customer-controlled encryption keys, open-source alignment, and regulatory compliance support — are an attempt to answer the sovereignty paradox that European and allied-nation governments have been raising. Google's explicit position, stated in a February 7 blog post: sovereignty and growth are not a trade-off. For enterprise IT leaders advising public sector clients, this matters because it removes one of the most durable procurement objections to Google Cloud. The conversation can move from "can we use Google?" to "which tier fits our classification level?"

AI on the flightline: paper forms are the first thing to go

Before the big partnership announcements, something quieter happened at Nellis Air Force Base in Nevada. Between January 12 and 16, 2026, the Air Force's Rapid Sustainment Office (RSO) — in collaboration with Google Public Sector — ran a five-day demonstration of a new application called the LITE Digital Maintenance Binder at the 57th Strike Aircraft Maintenance Unit.

The problem it solves is not glamorous, but it is real. Aircraft maintenance crews have been managing documentation on paper binders — manual forms, physical transport of documents, redundant data entry. The LITE system replaces that process with direct digital capture at the point of maintenance, with validation rules that prevent incorrect entries and automatic transmission to the Air Force's Integrated Maintenance Data System (IMDS).

During the event, an unplanned power outage hit the flightline. Maintainers using the Digital Maintenance Binder on mobile devices kept working. Those on legacy desktop terminals could not. Staff Sergeant Jacob Janusz from the 57th AMU put it directly: "Digital Binder just works. I wish we were using it right now."

This matters to enterprise technology leaders for a reason that goes beyond aircraft maintenance. When a government agency can demonstrate a 65 percent reduction in process time from a digitization project — with a clean before-and-after measurement, validated in a live operational environment — it provides the kind of return on investment evidence that justifies the next ten modernization proposals. The hard part of government technology adoption is rarely the technology. It is producing credible proof that change is worth the disruption. The RSO just produced that proof.

Gemini goes to work inside defense programs

On February 3, defense and government services company V2X (NYSE: VVX) announced a partnership with Google Public Sector to deploy Gemini-based artificial intelligence inside secure, on-premises, and isolated environments for US defense and intelligence agencies. The four named use cases are real-time multi-modal data analysis, AI-driven training and simulation, logistics and sustainment optimization, and supply chain risk detection.

V2X is a global company of approximately 16,000 professionals embedded in defense sustainment, logistics, and mission support programs. The practical implication of this partnership is that Google's generative artificial intelligence capabilities now have a path into mission-critical defense programs through a contractor with existing agency relationships and contract vehicles — without waiting for agencies to run their own procurement cycles for cloud AI services.

This is the classic systems integrator play in government technology, and it is strategically sound. Google gains mission-relevant deployments and accreditation experience. V2X gains a differentiated capability in a market where AI is increasingly a competitive qualifier, not just a feature.

Local governments are moving without waiting for federal permission

Google partnered with the US Conference of Mayors to publish an AI Playbook for local governments. The accompanying case studies from Chattanooga, Tennessee, and Mableton, Georgia, are not aspirational.

Chattanooga Mayor Tim Kelly described a specific goal: ingesting resident feedback from phone, email, 311, text, and web channels, then using artificial intelligence to identify patterns that prevent a small number of loud voices from driving city decisions. Mableton Mayor Michael Owens, whose city is one of the newest in the United States, described building AI governance first — including a generative artificial intelligence policy inside the city's HR handbook — before selecting specific projects.

"AI is already being used in your city whether you've acknowledged it or not. The worst position a government can take is to pretend it's not happening." — Mayor Michael Owens, Mableton, Georgia

Owens also said something that should land with any CIO managing a resource-constrained organization: AI is a tool that allows a small team to punch above its weight. That framing is precisely how under-resourced government IT departments are beginning to justify AI investment — not as transformation for its own sake, but as a force multiplier for teams that cannot grow headcount.

Iowa separately announced a partnership with the federal Administration for Children and Families to modernize child welfare technology. Child welfare case management represents one of the most technically and politically resistant categories of government legacy software in the country. If that category starts moving, it signals that the tolerance for delayed modernization — even in sensitive social services — is running out.

The educator initiative is an infrastructure investment in disguise

Google for Education, in partnership with ISTE+ASCD, announced free Gemini training for all six million K-12 teachers and higher education faculty in the United States. The training covers tools including Gemini and NotebookLM, serves over 74 million students indirectly, and will be delivered in short modular sessions that award micro-credentials on completion.

The curriculum is built around ISTE+ASCD's AI Ready Graduate framework. Richard Culatta, CEO of ISTE+ASCD, described the core problem plainly: too many teachers are being asked to navigate artificial intelligence without the training to use it effectively.

On the surface this looks like education outreach. Look at it differently: every educator who builds fluency with Google's AI tools becomes a channel through which the next generation of students, government employees, and voters forms their baseline expectations of artificial intelligence. Within five to ten years, the workforce entering government agencies will carry different default assumptions about which AI tools feel familiar and trustworthy. Those assumptions shape procurement preferences before a single request for proposals is written.

The competitive read

Google is running a three-layer strategy across these announcements. Against Microsoft, the play is on openness and model freshness — Gemini 3.1 Pro directly challenges the "Microsoft is the safe, familiar default" narrative that has historically driven federal technology decisions toward Azure and Microsoft 365. Against Amazon Web Services, the angle is the native AI layer: AWS has deep infrastructure penetration in federal agencies, but it does not have a consumer-facing frontier model with Gemini's name recognition. Google is betting that consumer visibility creates an adoption path that begins before formal procurement, when agency employees start using it on their own.

The educator initiative is the long game against both. It is building an installed base that no advertising campaign can replicate.

What IT leaders should be watching

The gap between Google's public sector positioning and actual agency procurement timelines is still wide. Most federal information technology acquisition cycles run eighteen to thirty-six months. The real test of this strategy is whether the V2X model — embedding accredited AI inside contractors with existing program relationships — can compress that timeline into something that produces results before the next model generation arrives.

The local government opportunity is more immediate. City governments are less encumbered by federal acquisition rules, they have urgent problems and limited staff, and they are increasingly willing to move quickly. The US Conference of Mayors AI Playbook is an attempt to give those governments a structured path forward rather than leaving them to figure it out city by city.

For CIOs and CTOs advising public sector clients, the practical near-term question is this: does your client have a clear answer to the sovereignty question? If not, the conversation about cloud AI cannot really begin. Google has now made that conversation easier to have.