Each January, innovation conferences follow a familiar ritual. Over 100,000 people gather in Las Vegas for CES, where companies showcase products you can touch, examine, and, occasionally, struggle to lift. A week later, the conversation moves to Munich, where DLD convenes European thought leaders to process what CES revealed. Then, a few days after that, the circuit culminates in Davos, where heads of state and Fortune 500 CEOs gather in the Swiss Alps to discuss governance, inequality, and the policy frameworks needed to manage transformation.

This three-stage progression – products, then reflection, then policy – isn’t a coincidence. It’s a structured sense-making process, our collective attempt to metabolise technological change, to examine it through increasingly abstract lenses. But the journey from concrete to conceptual reveals something uncomfortable about how institutional power relates to transformation.

The infrastructure reality

CES 2026 made the constraints visible:

Strip away the product launches, and CES 2026 demonstrated that AI progress has hit a wall – not of imagination or algorithms, but of physics and economics. The breathless promises of AGI within years run headlong into the mundane reality that we don’t have the electricity, cooling capacity, or manufacturing throughput to support anything close to universal AI deployment.

This is a concrete reality. You can measure it in megawatts, count it by fabrication plants, and map it geographically based on who controls rare-earth minerals and chip-making equipment. Jensen Huang lifting server racks, AMD showcasing 7,000-pound infrastructure monsters, OpenAI admitting compute constraints – these aren’t abstract concerns. They’re physical limits determining what’s actually possible.

The European interlude

Between Vegas and Davos, DLD Munich convenes to process what CES revealed. The gathering brings distinctly European perspectives on digital sovereignty, adds cultural dimensions around technological change, and discusses ethical frameworks for AI governance. Over 2,000 attendees, 250+ speakers, conversations that matter.

But watch what happens in this middle space. The conversation shifts from “can we build this?” to “should we build this?”; from infrastructure constraints to value questions. The physical reality of 7,000-pound server racks becomes abstracted into discussions of responsibility, innovation ecosystems, and Europe’s role in global technological advancement.

This isn’t wrong – these are questions that matter deeply. But something gets lost in translation. Concrete infrastructure challenges become conceptual “sustainability considerations”. The geographic concentration of semiconductor manufacturing capability – something you can map with precision – becomes “digital sovereignty concerns” that feel more aspirational than actionable.

The discussion is sophisticated and necessary. But it operates one level removed from the constraints that actually determine what’s possible.

The policy abstraction

By the time the conversation reaches Davos, the abstraction is complete. Technology becomes something to be “governed,” “regulated,” and “aligned with democratic values”. The specifics dissolve into systems-level conversations about inequality, geopolitical competition, and the future of work.

This matters because Davos is where actual policy gets shaped. It’s where finance ministers meet central bankers, where trade frameworks get negotiated, and where regulatory approaches find political backing. The influence is real.

But here’s the problem: the gap between what was launched in Vegas and what can be governed in Davos keeps widening. Democratic institutions process change more slowly than technology ships. By the time policy frameworks emerge, the infrastructure reality has already shifted.

CES 2026 demonstrated that China’s BYD overtaking Tesla as the world’s largest EV manufacturer wasn’t primarily about better technology – it was about manufacturing capacity, battery supply chains, and domestic market scale. The infrastructure to build and power EVs existed. Western policy conversations about EV adoption happened whilst China built the factories.

What the arc reveals

January innovation conferences reflect our hope that transformation flows sequentially: products first, reflection second, policy third. That if we gather the right people in the right rooms, we can guide change toward preferred outcomes.

But transformation doesn’t actually work this way. Infrastructure gets built (or doesn’t) based on capital availability, geographic accident, and political will operating on different timelines than product launches or policy debates.

As we documented throughout 2025, the pattern is consistent across domains:

We have the technical solutions – cheap clean energy, AI-augmented healthcare, abundance-creating automation. What we’re missing is the political capacity to implement them at scale. We can build the abundance we need – we’re choosing not to, through regulatory inertia, incumbent resistance, and bureaucratic dysfunction.

Incremental improvement, transformative restructuring, revolutionary change – all happening at once, governed by constraints the January circuit processes through those increasingly abstract lenses.

What the January arc really reveals is a confidence gradient. Certainty decreases as you move from product demonstrations to policy frameworks. In Vegas, you can lift the server rack. In Davos, you’re drafting principles for AI governance whilst the server racks are already being deployed at scale in data centres you’ll never visit.

The most significant insights emerge not in the keynotes but in the gaps. The distance between “we built this” and “we can govern this” grows annually. The political capacity gap – democratic institutions’ inability to keep pace with technological deployment – isn’t a temporary lag. It’s structural.

Living in the gap

The January circuit provides frameworks and vocabulary for collective sense-making. Each gathering contributes essential perspectives. But taken together, they reveal something we’re reluctant to acknowledge: the future isn’t knowable through sequential processing. It’s already arriving unevenly, shaped by infrastructure constraints and capital flows that conferences can observe but rarely influence.

We’re learning to inhabit what we’ve called permanent liminality – the threshold that doesn’t resolve. Not because we lack expertise or goodwill, but because transformation operates across multiple simultaneous timelines that don’t neatly synchronise with policy frameworks.

The real question isn’t whether the January circuit serves its purpose, because it does. No, the real question what kinds of insight this system enables versus what it obscures. We excel at synthesis, vocabulary-building, and relationship formation. We struggle with genuine uncertainty, transformation that hasn’t been named yet, and insights that emerge from infrastructure constraints rather than strategic planning.

Perhaps the conferences that open the year are the ones most trapped in the previous era’s assumptions – that futures are predictable, that expertise conveys authority, that gathering the powerful in rooms enables them to guide change. Meanwhile, the actual infrastructure that determines what’s possible is built in factories, power plants, and semiconductor fabs that don’t send delegations to Munich or Davos.

The gap between products and policy, between infrastructure reality and governance aspiration, isn’t closing. It’s the territory we inhabit now. And learning to work within that gap – rather than pretending we can bridge it through better convening – might be the only work that matters.

Photo by Oli Hale from Pixabay