Risk Transfer

Capacity Stack

Answer how much modeled exposure needs to be retained, transferred, layered, or reviewed. Insurance capacity should be based on modeled loss distribution and desired risk transfer, not total AI investment.

TOTAL AI INVESTMENT ≠ REQUIRED INSURANCE CAPACITY

AI Investment -> Insurable Interest -> AI Value at Risk -> Loss Distribution -> Controls -> Retention -> Transferred Risk -> Required Capacity

Layer inputs

Underwriter assumptions are preserved as user assumptions, not replacements for the model.

RetentionPrimary limitExcess limitAdditional layerAggregate limit

AI Value at Risk

$106,657,792

What economic value could be affected

Control-Adjusted Exposure

$67,514,382

Modeled exposure after controls

Maximum Plausible Loss

$6,440,000

Selected scenario MPL

Transferred Risk

$4,940,000

Subject to aggregate limit

Capacity stack

AI Value at Risk -> Control-Adjusted Exposure -> MPL -> Retention -> Primary -> Excess -> Alternative capital -> Residual retained risk.

Insured retention$1,500,000

Primary layer$5,000,000

Excess layer$7,500,000

Reinsurance / alternative capital$3,000,000

Residual retained risk$61,074,382

Model provider concentration feeding capacity

Portfolio correlation view for AICR and capacity analysis.

OpenAI

Excessive concentration

3 systems · $160,615,487 AI Value at Risk · $99,995,753 Control-Adjusted Exposure

Anthropic

Normal

1 systems · $21,438,857 AI Value at Risk · $9,969,069 Control-Adjusted Exposure

Google

Normal

1 systems · $10,327,200 AI Value at Risk · $4,543,968 Control-Adjusted Exposure

Cloud provider concentration feeding capacity

Portfolio correlation view for AICR and capacity analysis.

Azure

Excessive concentration

2 systems · $138,520,706 AI Value at Risk · $85,899,283 Control-Adjusted Exposure

AWS

Elevated

2 systems · $43,533,638 AI Value at Risk · $24,065,539 Control-Adjusted Exposure

Google Cloud

Normal

1 systems · $10,327,200 AI Value at Risk · $4,543,968 Control-Adjusted Exposure