Deployment

Reference deployment: fourteen coordinated facilities.

A low-risk deployment starts by learning the site, validating the twin against measured behaviour and expanding authority only where the evidence supports it.

1reference campus under one operating model

14facilities across manufacturing, compute, energy and logistics

2,400robots and autonomous assets

180,000telemetry streams

35,000digital-twin entities

12AI agents in the mesh

24/7continuous-operation reference design

< 100 mstarget edge response

99.999%target control-plane availability

Reference path: how deployment reduces risk

Weeks 1–3

Connect and map

Map assets, connect data sources and define hard constraints. BASALT observes; existing control remains unchanged.

Weeks 4–8

Validate the twin

Run the twin against live operations and score predictions against measured outcomes. Gaps are corrected before any operational authority is added.

Weeks 9–12

Operate in advisory mode

BASALT proposes plans with evidence and trade-offs. Your team decides what to execute and learns where the model is trustworthy.

Quarter 2

Expand authority selectively

Only domains that have proven their predictions move into supervised execution, inside an agreed envelope with operator hold and rollback.

BASALT

See what BASALT could change at your site.

Start in observation mode. BASALT builds and validates a live twin against real operations before any autonomous action is enabled. Authority expands only where the evidence supports it.

Review your deployment Talk to an engineer