This article is co-authored by Kelly Ommundsen and Stephan Mergenthaler from the World Economic Forum, Timo Graf von Koenigsmarck and Marc Reinhardt from Capgemini, and Manuel Kilian from Global Government Technology Centre Berlin.

One year after publishing our first flagship GovTech report, we take stock: where has GovTech come from, where did 2025 take us, and where is it all heading? A reflection on the milestones, the shifts in thinking, and what the rise of agentic AI means for what comes next.

This article marks the one-year anniversary of our flagship report, The Global Public Impact of GovTech: A $9.8 Trillion Opportunity, produced jointly by the World Economic Forum, Capgemini, and the Global Government Technology Centre Berlin.

Throughout writing the report, we kept asking ourselves: are we overstating the case? Is GovTech really as important as we claim?

So we gathered the evidence. What we found was a $9.8 trillion opportunity — the public value governments could unlock through GovTech by 2034. This is not market revenue. It is impact that flows to society and business at large if governments implement GovTech responsibly and at scale. $5.8 trillion through efficiency, $1.1 trillion through transparency, $2.9 trillion through sustainability.

That figure gave us confidence in the scale of what's possible. What became clearer over the past year is how unevenly that opportunity is being pursued, and how quickly the technological landscape has shifted beneath our assumptions.

However, we were equally clear in the report: this figure is meaningless without execution. The opportunity is reserved for governments that actively reorganize to capture it. Supporting governments on this journey has been the cornerstone of our joint initiatives throughout 2025.

2025: The Year GovTech Became Undeniable

As we look back at the past twelve months, it is clear that 2025 was the year GovTech further transitioned from a promising frontier to an operational necessity. We have identified six fundamental shifts that defined this year — moving from the high-stakes game of global geopolitics to the rigorous demand for real-world results. Underpinning each of these shifts is a single, evolving reality: the digital transformation of the state is no longer about testing what is possible, but about executing what is essential.

1. The Emergence of the Agentic State

Agentic AI — systems capable of perceiving situations, reasoning through problems, and taking autonomous action within defined boundaries — emerged as the most consequential technological development of 2025. The speed and scale of these advancements are without precedent.

For governments, the implications are profound. A significant portion of public sector activity consists of codified workflows: processing applications, verifying eligibility, issuing permits, and enforcing compliance. These are precisely the structured, rule-bound processes where agentic AI offers great transformative potential.

This marks an inflection point in public sector digital transformation. The possibilities were explored this year in The Agentic State, a vision paper produced by a global group of government leaders and experts with the support of the Global Government Technology Centre Berlin and the World Bank.

However, unlike DPI, where a shared blueprint already exists, we are only beginning to understand how to effectively integrate agentic AI into government operations. How should a state manage systems that act rather than merely advise? What does accountability look like when an AI agent initiates a process? We do not yet have established patterns to follow.

While our original report anticipated that new technologies would continue to expand the horizons of GovTech, what we did not anticipate was the velocity. The speed at which agentic AI has moved from a laboratory concept to a commercial reality — and its sheer potential for public service — suggests that our $9.8 trillion valuation of the GovTech opportunity may look conservative in the future.

2. Less Experimentation, More Execution

GovTech has further matured in 2025, and we saw a continued shift away from pilots for their own sake toward a harder question: does this actually deliver value?

Part of this shift is driven by global fiscal reality. With budgets tightening, the appetite for exploratory projects without a clear return on investment has diminished compared to five years ago. But confidence has also grown. There is now compelling evidence that well-executed GovTech solutions deliver substantial, measurable impact. This track record has raised the bar; governments today increasingly demand results rather than just experiments.

Our report argued that “GovTech implementation does not happen by chance, but through strong execution.” We are encouraged to see this principle becoming the operational standard.

The open question is how the rise of agentic AI will influence this momentum. Agentic systems demand experimentation, but of a different kind. Governments’ ability to pilot rigorously, learn quickly, and decide when to scale will be tested by technologies that behave fundamentally differently from what came before.

The stakes have changed. Governments can no longer afford pilots that do not have a clear path to production, nor can they afford to wait years to explore technologies that competitors and adversaries are deploying at scale. This tension, between the need for rigor and the pressure for speed, will define GovTech execution in 2026.

3. GovTech Moved Closer to the Centre of Government

In 2025, Germany, the world's third-largest economy, finally established a dedicated ministry for digital, consolidating responsibilities previously scattered across six ministries. The signal was clear: GovTech is important enough to require and deserve a ministry of its own.

At the start of 2026, Ukraine's Mykhailo Fedorov was nominated to move from digital minister to defence minister. Fedorov built Diia — the digital state in a mobile app. His transition underscores how digital leadership is increasingly seen as core statecraft, not a technical specialty.

At the same time, across many governments, ministerial titles are only meaningful if they come with actual authority. What matters is whether a digital minister can reshape procurement rules, unblock data sharing, or implement workflow reform across entrenched ministries. In some countries, progress is visible. In most capitals, digital leadership has a seat at the table, but rarely the final word.

Our report argued that GovTech is a whole-of-government discipline, not an IT function. What we saw in 2025 was precisely this shift: governments elevating digital from a technical function to a strategic priority. The question for 2026 is whether that elevation translates into power or remains symbolic.

4. Digital Public Infrastructure as the Foundational Blueprint

The concept of Digital Public Infrastructure — common, re-usable building blocks for identity, payments, and data exchange — has further moved to the established blueprint for the digital transformation of the state.

The UCL Institute for Innovation and Public Purpose recently published the first comprehensive global assessment of this shift, finding that 64 countries now have DPI-aligned digital ID systems, 97 have digital payment systems, and 103 have established data exchange platforms (see here). Furthermore, the 50-in-5 campaign, launched in 2023 to help 50 countries deploy DPI by 2028, has seen its membership grow from 11 to 30 nations. At 60% of its target, the momentum is significant.

Equally important, DPI has become a shared vocabulary: a way for governments to learn from one another, rather than build in isolation.

But quantity does not guarantee quality. In some jurisdictions, rapid deployment has outpaced interoperability, privacy safeguards, and inclusion.

The question for 2026 is not just how many more countries deploy DPI, but whether the systems being built will withstand the weight we're about to put on them. When agentic AI starts making decisions based on identity verification, payment confirmation, and data exchange, flawed infrastructure will not merely slow delivery — it will amplify failure at scale.

DPI is rightfully understood as the foundation on which everything else builds: tax collection, social protection, healthcare access. Without central identity, payment rails, and underlying data platforms, the efficiency and transparency gains our report described remain out of reach. What this underscores is a simple but consequential distinction — building DPI is no longer the challenge; building DPI that is interoperable, inclusive, and resilient is.

5. Technology Became Geopolitics

Technology has become political. This is not new, but 2025 made it clearer than ever. And for governments, the implications are direct: the most immediate way a state interacts with technology is not by regulating it or subsidising it — it is by using it. By running public administration on it. That is what GovTech is.

This places GovTech at the centre of the global conversation around digital sovereignty. Digital sovereignty, not to be mistaken with the idea of full technological autarchy, is increasingly seen not just as a policy goal, but as a requirement for the infrastructure a government runs on. Throughout 2025, it evolved from a conceptual debate into a matter of operational resilience.

We have seen this reflected in a growing move toward localised and diversified infrastructure. For instance, the Dutch parliament has called for a pause on certain cloud migrations to prioritise domestic or regional alternatives, while Austria has shifted collaboration tools to European providers to ensure closer alignment with local data standards. We are beginning to see similar questions emerge around Large Language Models: who provides them, under what conditions, and what continuity looks like as these systems become more embedded in government operations.

But the more revealing signal came from what governments stopped doing. We saw procurement processes stall not because vendors couldn't meet technical requirements, but because decision-makers couldn't answer fundamental questions about supply chain resilience. What happens if access to this model is withdrawn? What if the training data reflects values we don't share or sourced from contested sources? What if the provider is subject to legal obligations that conflict with ours?

These aren't hypothetical concerns. The question is no longer whether to use AI in government, it's whether governments can use AI without creating dependencies they can't afford. And for many governments, the honest answer is: we don't know yet.

Our original report framed the value of GovTech through efficiency, transparency, and sustainability. 2025 has added a critical fourth dimension: sovereignty. The technology governments use to deliver essential services is now inseparable from the question of who controls access to that technology — and how to smartly manage interdependencies that come with it.

6. A Larger GovTech Market, a Stubborn Bottleneck?

Our report estimated the GovTech market at $606 billion in 2024, growing to $1.4 trillion by 2034. That estimate was tied to IT spending in the public sector, which is a conventional and defensible approach.

But in light of what we have seen with agentic AI, this framing may need revisiting. If AI systems can take over tasks previously performed by service providers or internal staff, then GovTech is no longer competing only for IT budgets. It is eating into the broader share of services bought by the public sector — operations, case processing, compliance, customer service. The market we described may be larger than we thought, and it may look quite different.

The paradox, however, is stark. As capability accelerates, governments’ ability to buy and deploy technology has barely changed. Procurement cycles designed for buying furniture are being asked to evaluate AI systems. Risk frameworks built around implementation failure haven't adapted for algorithmic bias or model drift. Procurement cycles are long, risk appetite is low, and vendor access is uneven.

This isn't a problem technology can solve. It's a problem of state capacity, institutional adaptability, and whether governments can move at the speed their citizens will increasingly demand. Whether the demand catches up to the technology, or the technology outpaces the market's ability to absorb it, is something we will be watching closely.

2026: What Comes Next

Our shared ambition remains unchanged: turning technology into public impact. This principle guided our work throughout 2025, and it will continue to do so in the year ahead.

For us, this means empowering governments to make the best possible decisions on technology. Our role is not to advocate for technology for its own sake, but to provide the evidence, frameworks, and convening power that allow governments to act with clarity and confidence.

In 2026, the greatest appetite for knowledge and insight centers on how to deploy agentic AI responsibly and at scale. This is where we will focus a good part of our collective energy.

In the coming weeks, we will bring this discussion to the World Economic Forum Annual Meeting in Davos and subsequently to the World Government Summit in Dubai. By the end of the first quarter, we will publish our second flagship GovTech report, titled The Agentic State Decoded: A Function-Based Assessment of Government Readiness — a study that maps the agentic AI opportunity across the full breadth of government activity, identifying where the potential is highest and where the barriers remain.

The $9.8 trillion opportunity we identified a year ago is more real than ever. So is the urgency. We look forward to continuing this work, alongside governments, technology providers, and the broader GovTech community, to help turn that potential into reality.