For the past two years, our public-sector predictions have tracked governments’ increasingly uncomfortable relationship with global technology power. Two years ago, we rightly predicted that governments would ramp up regulation attempting to curb the might of the tech giants, with middle powers acting collectively because going alone wouldn’t be enough. Last year, we predicted that this contest would shift from regulating technological power to reducing dependence on it as governments chasing digital self-sufficiency embraced increasing tech nationalism.

Neither force has subsided. Both have intensified. In the last year, AI has supercharged sovereignty concerns just as the international order has become more multipolar. At the same time, Silicon Valley’s once-formidable social license has been eroded by the increasingly political, sometimes imperialistic behavior of technology leaders and persistent consumer skepticism about AI. It’s a potent mix in which governments want the economic productivity improvements that AI promises while people expect them to retain control over technology that increasingly affects their day-to-day lives. The problem? Governments of all persuasions are claiming digital sovereignty when, in fact, none truly have any.

Our Global Sovereignty Forecast, 2025 To 2030 shows that digital sovereignty requires the autonomy and agency to both develop and deploy technology using critical resources, infrastructure, data, and talent independently of the decisions and influence of foreign governments. No nation today can achieve this alone. Instead, alliances, partnerships, open technologies, diverse capital sources, and selective investments are needed to manage unavoidable and unwanted dependencies.

In the next year, AI will expose that sovereignty gap even further. Governments will discover how little control they actually have over the technology, suppliers, decisions, and skills underpinning public services. Their response? More intervention, not less, further fragmenting already diverging technology markets, adding increased obligations on suppliers, and frustrating consumers. But governments will see these efforts as necessary to protect public services, digital rights, and national resilience and to respond to the demands of their people.

As a result, we predict that 2027 will be a year in which:

  • Political intervention will derail $40 billion in government technology procurement. Forget procurement based simply on functionality, price, security, and value for money. Politics is now part of the evaluation panel. Political leaders will increasingly overrule CIOs and procurement officials when a technology supplier’s ownership, AI model provenance, data access, supply chain, or political exposure creates unacceptable sovereign risk, even when career officials assess that the actual manifestation likelihood and impact are low. We are already seeing this with how France’s sovereignty agenda helped local cloud provider Scaleway win a health-data contract over Azure, while some Australian lawmakers have called for Palantir to be banned from government contracts over its political alignment. In the AI era, governments will increasingly ask: Who owns this technology? Who controls it? And what happens if today’s trusted dependency becomes tomorrow’s strategic liability? Government technology leaders should identify political sensitivities before issuing RFPs and assess shortlisted providers against sovereignty, resilience, competition, and public trust.
  • At least one G7 country will air-gap a sensitive system after an autonomous AI attack. For more than a decade, government technology strategies have pushed toward cloud adoption, connectivity, data sharing, and interoperability. AI will force leaders to confront the other side of that equation. We predict that a state-sponsored autonomous AI attack will cause at least one G7 nation to move an intelligence, defense, identity, or critical-service system into a fully air-gapped, disconnected, or tightly controlled sovereign environment. This doesn’t herald a mass retreat from cloud. Instead, selective isolation becomes a legitimate resilience strategy. Security leaders should establish which additional sensitive systems could operate for 90 days without public cloud services, public models, shared agent services, or continuous external connectivity and apply Zero Trust and deception controls to those that must remain connected. Sovereign resilience will increasingly be measured not only by how quickly government can connect but by how safely it can disconnect.
  • Two-fifths of OECD countries will implement AI labor-market transition strategies. The AI workforce debate has become something of a circus, with leaders encouraged to choose between extraordinary abundance and mass technological unemployment. Governments can afford neither the hype nor the paralysis. AI is bigger than just another skills gap to fix with training courses. It will change occupations and how workers move between them. Countries will follow early moves by Singapore, South Korea, and California and combine workforce planning, reskilling, employer incentives, and social protections into funded AI transition strategies. Australia, Canada, and the Nordics have an advantage because of their established workforce infrastructure, but success must be measured through worker placement, mobility, retention, and wage growth, not course completion. And this is a sovereignty issue, too. A nation without the people required to build, secure, govern, and operate AI systems is little more technologically sovereign than one without the infrastructure on which they run.

Digital sovereignty cannot mean technological self-sufficiency. No one country has the infrastructure, suppliers, capital, or talent required to reproduce the entire technology stack domestically. Sovereignty must instead mean exercising strategic agency: knowing what must remain under national control, where trusted interdependence makes sense, which alliances can fill capability gaps, how domestic or global regulations can be enforced, and when governments need to accept that they have to walk away.

2025 was about regulating the power of technology companies. 2026 was about reducing dependence on them. In 2027, governments will discover whether they really have control. And AI is about to show us which ones really do.

Read our full Predictions 2027: Public Sector And Government report to get more detail about each of these predictions and read additional ones. Set up a Forrester inquiry or guidance session to discuss these predictions or plan out your 2027 strategy.

If you aren’t yet a client, you can download our complimentary Predictions guides, which cover more of our top predictions for 2027. Get additional complimentary resources, including webinars, on the Predictions 2027 hub.

Share