What it means
Agents ramp up, drift out of spec without anyone noticing, do damage, get flagged, queue for review and get retired. Sprawl cost is what that cycle costs a fleet as the fleet grows. The question in the AGENESIS-2 campaign was whether the answer depends on how the monitoring budget is set.
The lab simulated 256 runs across fleets of 57 to 787 agents over 18 months. Under a monitoring budget that stays flat while the fleet grows, cost grows as fleet size to the power 1.4665, with a 95% interval of 1.4101 to 1.5230. Each doubling of the fleet multiplies cost by 2.76. Where every agent carries its own check and review grows with the fleet, the exponent is 0.9001, or 1.87 per doubling.
The mechanism shows in the detection lag. With a fixed inspection budget, the time from an agent going bad to someone noticing runs from 2.78 weeks at 57 agents to 22.49 weeks at 787. Review capacity did nothing under diluted monitoring: in 55 of 64 matched cases, changing it left the simulation bit-identical, because nothing had been flagged for the reviewers to see. Detection sits upstream of review.
Independent verification found 95 duplicate runs. One verdict moved from supported to refuted and is published as refuted. The 1.47 awaits a confirmatory study.