2025 in reflection: The Unravelling of the Digital Settlement

When I look back at 2025, the word that keeps coming back to me is not speed, or innovation, or AI. It’s fragility.

Not the fragility of technology itself — the systems mostly work — but the fragility of the political, institutional, and normative arrangements that once gave digital transformation a sense of direction. The guardrails that made the internet feel like a shared global project are still there on paper, but in practice they’re thinner, weaker, and far more conditional than they used to be.

For roughly three decades, the global internet rested on a loose but surprisingly resilient settlement: openness over control, interoperability over fragmentation, cooperation over dominance, and at least a nominal commitment to human rights. That settlement was never perfect, but it held. In 2025, it’s clear that it’s no longer holding in the same way.

What’s emerging instead isn’t a new order. It’s a patchwork. A fragmented digital landscape shaped by geopolitical rivalry, transactional diplomacy, corporate concentration, and a growing willingness by states to trade long-term openness for short-term advantage.

The story of 2025 isn’t that the internet is “breaking.” It’s that the world that sustained it is.

The Quiet Retreat from Internet Freedom

For much of the internet’s history, the United States acted — sometimes awkwardly, often inconsistently — as the system’s anchor. It defended the multistakeholder model in global forums, resisted efforts to formalise state control at the UN, and framed internet freedom as both a democratic value and a strategic interest.

By 2025, that posture has changed in ways that are hard to ignore.

Internet freedom has largely disappeared as an organising principle of US foreign policy. It hasn’t been rejected outright — it’s just no longer driving decisions. Instead, the focus has shifted to industrial competitiveness, export controls, national security, and technology rivalry with China.

You can see this shift clearly in multilateral negotiations. The US still shows up, but often selectively: pushing back against language that might constrain American firms and its own ideology, while showing much less appetite for sustained coalition-building around openness, global public goods, or institutional reform. Multistakeholder language remains, but the political energy behind it is thinner.

Funding priorities tell the same story. Global internet freedom initiatives that once sat near the center of US digital diplomacy now feel peripheral, while bilateral and minilateral tech deals absorb most of the attention. This isn’t abdication by accident — it’s reprioritisation.

The consequence is structural. When there’s no consistent advocate for openness, multilateral institutions drift. Sovereignty-first logic fills the vacuum. Consensus still happens, but more often through ambiguity than shared purpose. Power doesn’t disappear when leadership retreats — it simply moves elsewhere.
 
The End of Distance

This retreat is compounded by another shift that felt impossible to miss in 2025: the deep integration of Big Tech into the political orbit of Washington.

This is no longer about lobbying at the margins. US big tech and AI firms help frame national security debates while US chipmakers shape export control policy. The distance between corporate power and state power has collapsed.

That collapse is felt far beyond the US.

From Europe’s perspective, regulatory disputes with American platforms increasingly feel geopolitical rather than legal. Enforcement of competition rules, data protection, or platform accountability is often met with warnings about innovation, security, or strategic alignment. For many outside the US, it’s no longer clear where American public interest ends and corporate interest begins.

The result is a growing sense that global digital governance is asymmetric: rules apply downward, leverage applies upward. That perception matters. It weakens trust, fuels regulatory assertiveness, and accelerates fragmentation — even among allies.

Europe’s Digital Sovereignty Paradox

In 2025, Europe re-engaged in the pursuit of digital sovereignty—but it is still unable to figure out how to bridge the gap between policy and power. The re-emergence of this urgency is largely a reaction to the “Trump factor” as the prospect of a more transactional and unpredictable U.S. administration has reignited fears of technological blackmail. This has sparked intense infighting among member states regarding their dependence on foreign tech; while some push for a radical decoupling to build a homegrown "EU Stack," others argue that Europe lacks the industrial scale to survive without American hyperscalers.

Despite being a regulatory superpower through the AI Act and the DMA, Europe’s sovereignty agenda remains more rhetorical than operational. Ambitions for resilience are constantly undermined by a lack of investment in infrastructure and compute, leaving projects like the Chips Act struggling to gain traction against established global ecosystems. Sovereignty is currently functioning as a defensive reflex—a way to regulate against external influence—rather than a proactive strategy to lead in innovation. In a world that rewards speed and scale, Europe remains trapped in a cycle of regulating a future it has yet to build.

China’s Strategic Clarity

China, by contrast, ends 2025 with a level of coherence that’s hard to ignore. Its digital strategy spans infrastructure, platforms, standards, AI, data governance, and international engagement — and it’s embedded in long-term industrial planning and foreign policy.

This isn’t abstract. You see it in Digital Silk Road projects providing cloud services, smart city technologies, and surveillance infrastructure across Africa, the Middle East, and Asia. You see it in China’s growing influence in standards bodies, particularly around 5G and emerging technologies. You see it in coordinated positions in multilateral negotiations that consistently emphasise data sovereignty and state authority.

What’s different this time is that China isn’t just participating — it’s offering a full alternative. Countries are no longer choosing between connectivity and isolation. They’re choosing between digital ecosystems, each with different governance assumptions and political trade-offs.

AI as the Defining Issue of 2025

If one technology crystallised all these tensions in 2025, it was artificial intelligence.

Like the early internet, AI exposes and deepens inequalities — but not just in connectivity. The divides are now about data, skills, energy, and access to compute.

And unlike the internet, AI is capital-intensive and centralised. Frontier model development is concentrated in a handful of firms, overwhelmingly in the US and China. Access to GPUs, cloud infrastructure, and large-scale datasets determines who can innovate and who can’t.

Export controls and pricing structures shape participation in very real ways. Universities, startups, and public institutions in many developing countries aren’t excluded by censorship — they’re excluded by cost.

The digital divide has returned, measured in compute hours and megawatts.

Chips, NVIDIA, and the New Dependency

No company captures the political economy of AI better than NVIDIA. By 2025, control over advanced AI chips has become a gatekeeping function — not just for innovation, but for economic development and national capability.

Governments now negotiate directly with chipmakers and hyperscalers to secure access to compute. Export controls shape alignment as much as security. For many countries in the Global South, the problem isn’t talent or ambition — it’s access.

This isn’t just a market issue. It’s a governance problem. We’re relying on private actors to supply what has effectively become critical infrastructure.

Infrastructure, Standards, and Quiet Power

Beyond chips, power in 2025 flew also through less visible channels: subsea cables, cloud regions, satellite constellations, internet exchange points, and technical standards.

Standards-setting, in particular, has become one of the most decisive battlegrounds and will continue to be for the forthcoming future. AI safety benchmarks, digital identity systems, and smart city protocols embed governance assumptions long before any law is passed. China, India, and a growing group of middle powers invest heavily here. Western engagement often feels fragmented and reactive.

Infrastructure is no longer neutral. It’s strategic.

The Global South Isn’t Waiting Anymore

One of the most underestimated shifts of 2025 is the agency of the Global South.

Rather than waiting for leadership from Washington or Brussels, many countries are building their own digital pathways. India’s digital public infrastructure — Aadhaar, UPI, data-sharing frameworks — is being adapted elsewhere. The African Union is pushing for continental coherence on data policy. South–South cooperation on skills and AI capacity is growing.

The posture has changed. Engagement with China, Western firms, and multilateral institutions is pragmatic and selective, driven by development needs rather than ideology.

Transactional Politics, Everywhere

All of this sits within a broader transformation in international relations. In 2025, digital cooperation is openly transactional.

Market access is exchanged for regulatory concessions. Infrastructure financing comes with political strings. AI partnerships are framed as security arrangements. Trust erodes, long-term cooperation weakens, and fragmentation becomes the default rather than the exception.

Human Rights: Still There, But Thinner

Perhaps the most worrying shift of all is what’s happening to human rights.

For decades, human rights provided the normative backbone of internet governance: privacy, freedom of expression, due process. In 2025, those principles are still referenced — but their practical force is fading.
National security, child protection, and misinformation are routinely used to justify surveillance, content restrictions, and AI-driven decision-making, often without meaningful safeguards. This isn’t limited to authoritarian systems. Democracies increasingly deploy similar tools, narrowing the gap in practice even if intentions differ.

AI accelerates the problem. Algorithmic systems shape access to welfare, credit, jobs, migration, and justice — but accountability becomes diffuse, harm probabilistic, and redress elusive. Human rights institutions themselves are under strain, underfunded, and politicised.

The risk is that rights become decorative language rather than real constraints on power.

Multilateralism Turns Inward

One of the clearest through-lines of 2025, for me, was how visibly multilateral digital governance became more state-centric. WSIS+20, the UN cybercrime convention negotiations, and the Global Digital Compact all pointed in the same direction. None of these processes collapsed. In fact, on paper, they can be read as successes: consensus was preserved, institutions were reaffirmed, and cooperation was repeatedly invoked. But taken together, they also revealed something deeper — a quiet but consequential rebalancing of power away from the multistakeholder model and back toward governments as the primary, and increasingly exclusive, decision-makers.

WSIS+20 captured this tension perfectly. The Internet Governance Forum was given permanent status, which mattered enormously for continuity and institutional memory. Data governance was acknowledged as central. And yet, the price of consensus was ambiguity. The outcomes stabilised what already exists rather than opening new space for shared problem-solving. In parallel, the UN cybercrime convention exposed the risks of this shift far more starkly. Despite widespread concern from civil society, technical experts, and parts of the private sector, the process remained overwhelmingly state-driven, with weak safeguards and broad definitions that many fear will be abused in practice. Meanwhile, the Global Digital Compact echoed familiar language about inclusion and cooperation, but ultimately reinforced a vision of digital governance anchored firmly in intergovernmental negotiation, with non-state actors consulted rather than meaningfully empowered.

Governments increasingly see digital issues — from data and AI to cybercrime and content — as extensions of sovereignty and security, not shared global challenges requiring distributed governance. This will continue in 2026.

Fragility as the Defining Condition

If there’s one takeaway from 2025, it’s this: the fragility of the international order and the fragility of the internet are inseparable.
​
As trust between states erodes, so does the willingness to maintain a shared digital commons. As institutions weaken, power concentrates with those who control infrastructure, capital, and leverage.
The future of the internet won’t be decided by technology alone. It will depend on whether we can rediscover a minimum commitment to cooperation over control, rights over expediency, and legitimacy over dominance.