Institutional-Grade Research
on the AI Agent Economy

Agent banking rails · Monitoring

Strong operator velocity, but distribution quality still lags infrastructure ambition.

Bankr has a credible primitives layer for agent-native treasury flows, but the current growth mix still depends too heavily on launch-cycle attention instead of sticky protocol volume.

Agent launch infrastructure · Speculative

Fast attention capture, weak defensibility.

ClawPump is effective at compressing launch friction, but we do not yet see durable moats in distribution, retention, or developer lock-in.

Inference coordination · Threshold watch

Close to premium coverage, still missing proof on cost-quality persistence.

Relay Arc is near the 80 threshold and benefits from serious architecture decisions, but the margin story depends on inference routing advantages that have not fully persisted through volatility.

Agent consumer network · High conviction

Distribution strength and ecosystem pull justify premium coverage.

Virtuals remains in Siro's high-conviction tier. The ecosystem has enough evidence of sustained demand and narrative control to keep the full report subscriber-only.

Agent infrastructure · Core coverage

Execution quality and strategic leverage keep this in the premium tier.

Autonome Cloud stays locked to subscribers because the score remains firmly above 80 and the research contains Siro's highest-conviction operating view.