The technology underneath AI must be adequate for the use cases, systems and models the FI intends to run. AI systems carry the same technology risks as any other system, availability, resilience, safety and cyber among them, so the FI confirms that processing units, network, memory and secure data pipelines meet the performance, scalability and resilience that each use case demands, guided by the MAS technology risk guidelines and notices and by industry frameworks such as the infrastructure elements of the NIST AI Risk Management Framework.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.