SCP-8895
Safe
~
medium confidence
SCP-8895
Expected annual
$1.2M
One-time setup
$3.0M
Annual recurring
$1.2M
Personnel
7
One-time capital costs for secure lab, compute and potential OR buildout dominate (≈$3.0M one-time if OR is built); recurring annual costs for an active research program are ≈$1.20M/yr driven by personnel, insurance, legal/OPSEC and specialized maintenance. A minimalist lock-and-don't-test posture reduces annual costs to a few thousand dollars.
One-Time Capital Costs
Total: $3.0M
Annual Recurring Costs
Total: $1.2M/yr
Cost Scenarios
📊
Baseline
(baseline)
$1.2M/yr
Active research/testing program year with scheduled testing and full maintenance (normal, uneventful year).
scheduled_testing
normal_operations
no_major_incidents
🚨
Minor Incident
$1.3M/yr
Corrupted .adx produces hazardous output or localized data loss requiring investigation, cleanup and limited medical/legal response.
corrupted_adx_hazard
limited_contamination_or_data_loss
small_investigation
🚨
Major Breach
$3.2M/yr
bunbun.exe spreads or there is a large-scale containment/OPSEC failure causing enterprise infection, major legal exposure, and extended recovery.
enterprise_infection
insider_breach
public_exposure
Personnel
7 total
| Role | Count | Notes |
|---|---|---|
| Principal Investigator / Lead Researcher | 1 | [#6] PI overseeing experiments and analysis. |
| Neurotech / Lab Technician | 2 | [#6] Two technicians for handling, sample prep and lab work. |
| Data Scientist / Forensic Analyst | 1 | [#6] Analysis of .adx files and ML/forensic processing. |
| Clinical Psychologist | 1 | [#6, #12] On-call for subject monitoring and psychological safeguards. |
| Administrative Staff | 1 | [#6, #7] Admin support for consent, IRB-equivalent and logistics. |
| Security Officer / MTF Agent | 1 | [#2] Incremental security allocation for access control and physical checks. |
Confidence Notes
Analyst notes include detailed line-item ranges for most categories, but several large items (number of procedures/year, whether an in-house OR is built, cadaveric specimen usage, and severity/frequency of incidents) are highly variable. Estimates are midpoint/conservative choices where ranges existed.