ConceptDiscovery layer

Ephemerals Branch Margin Model

Ephemerals lets agencies resell disposable cloud dev environments per Git branch, each running Claude Code. The margin model hinges on the $10/seat/month fee plus metered compute at $0.18 per 2 vCPU/4 GB hour, with heavier add-ons at $1.44 per 16 vCPU/32 GB hour. An agency can charge clients a flat retainer, say $2,500 for a 20-hour setup, then a per-developer monthly fee that covers seat costs and a buffer for compute. Since Ephemerals does not proxy Claude API tokens, the agency must separately bill for agent usage, which can erode margins if not tracked. The model works best when clients run parallel agent-assisted tasks across branches, maximizing the value of isolated environments. Agencies should monitor compute per branch to set retainer tiers that keep margins above 30%.

By InnovaAI ResearchPublished Updated

AI Code Tools

What is Ephemerals Branch Margin Model?

“Per-branch compute cost → client retainer margin”

Compute hours per branch vs. retainer margin

Ephemerals lets agencies resell disposable cloud dev environments per Git branch, each running Claude Code. The margin model hinges on the $10/seat/month fee plus metered compute at $0.18 per 2 vCPU/4 GB hour, with heavier add-ons at $1.44 per 16 vCPU/32 GB hour. An agency can charge clients a flat retainer, say $2,500 for a 20-hour setup, then a per-developer monthly fee that covers seat costs and a buffer for compute. Since Ephemerals does not proxy Claude API tokens, the agency must separately bill for agent usage, which can erode margins if not tracked. The model works best when clients run parallel agent-assisted tasks across branches, maximizing the value of isolated environments. Agencies should monitor compute per branch to set retainer tiers that keep margins above 30%.

ai-code-tools
Open the Ephemerals courseEphemerals decision recordBrowse the Academy

More on Ephemerals

More for AI Code Tools