A race without one finish line
The phrase race to superintelligence suggests a simple contest with a visible leader. The reality is less convenient. Organisations can lead in model performance while trailing in distribution, infrastructure, capital efficiency or public trust.
Any useful view of the race therefore needs several dimensions. A single benchmark or funding announcement can be meaningful, but it cannot explain the complete strategic position.
Capability is necessary but unstable
Model capability is the most visible dimension. Benchmarks, evaluations and observed performance help show what systems can do. Yet results are sensitive to the test, the conditions and the distinction between a research result and a capability that users can access reliably.
The relevant question is not simply which model scored highest. It is whether the capability is repeatable, economically usable and connected to products that can improve from real interaction.
Compute, capital and talent shape the pace
Frontier development depends on infrastructure, energy, specialised hardware and the ability to coordinate expensive research programmes. Capital creates options, but spending alone does not create a durable lead. Execution, research judgement and access to scarce talent determine how effectively resources become capability.
These inputs also have different time horizons. A model can change perceptions overnight. New data centres, energy supply and research organisations take much longer to build.
Distribution turns capability into influence
A technically strong model can have limited impact without distribution. Existing products, developer ecosystems, enterprise relationships and consumer habits influence how quickly capability reaches real work.
Distribution also generates feedback and revenue. That can strengthen the next development cycle. It is therefore both an outcome of technical progress and an input into future competition.
Governance belongs on the board
Safety, regulation and geopolitical access are not external commentary on the race. They influence what can be trained, deployed, exported and trusted. A major safety failure or regulatory restriction can change an organisation's position even when its technical capability remains unchanged.
The Radar treats leadership as an editorial assessment built from multiple observable signals. It should show its sources, confidence and methodology rather than pretending to provide a permanent scientific ranking. The value lies in making movement intelligible, not declaring a final winner.