Rule: do not quote dollar amounts externally until validated with design partners. This page is internal posture for diligence readers.
Motions
| Motion | Price posture | Notes |
|---|---|---|
| Pilot evaluation | Fixed-scope project fee | One policy or one world-model batch; defined episode budget |
| Repeat / multi-policy batch | Discounted package | Same harness, multiple artifacts or versions |
| Continuous scoring (Phase 5) | Subscription | Release-pipeline integration; re-baseline fees if thresholds change |
Anchors (adjacent, not comps)
| Signal | What it suggests |
|---|---|
| Sim2Real public pilots $499–$2,500/mo | Willingness to pay for sim-to-real tooling exists — different product (training loop, not independent audit) |
| Patronus / enterprise eval vendors | Verification layers can support enterprise ACV once trust is established — not a near-term Haga price point |
| Academic benchmarks (free) | Institutional trust can be free; Haga sells speed + private engagement, not free leaderboard slots |
Use anchors for conversation, not for published rate cards.
Working ASP points used in TAM / SAM / SOM ($12k pilot / $18k near-term blended / $36k mature ACV) are internal hypotheses — update market sizing when design partners lock WTP.
Packaging principles
- Price the report and turnaround, not GPU hours alone
- Always include shown failures and sim-only limitations in scope — no “PASS-only” SKUs
- Keep pilot scopes narrow enough to deliver in days/weeks, not quarters
- Defer continuous-scoring pricing until at least one design partner has paid for a pilot and asked for repeat runs
Validation path
- Priority A outreach + inbound submit-for-eval (no public rate card)
- Record every conversation / WTP / objection on the design-partner pipeline
- Survival lane only: productized public-artifact audit (internal default quote) — dollars quoted only after payment gate; stop after 10 asks with no reply
- Only then publish ranges for investor materials
Related: Evaluation offering · ICP.