DenseNet: Infrastructure Failure Prediction

If an entrepreneur "adopted" this product today, here's the realistic math.

Fermi summary

If you close 5 utility contracts at $65k/yr by month 18, that's $325k ARR - but each deal costs ~$25k to pursue and takes over a year, your sales cycle will outlast your runway, and there's roughly a 6% chance you get there inside 12 months.

Market size (TAM)

$420.0M

~3,200 medium/large US water+electric utilities + ~800 large industrial operators × $90k avg annual software contract for predictive infrastructure monitoring

Year-1 ARR range

$75k - $1.1M

midpoint $320k

Gross margin

63%

Investment to production

$165k

Dev ($65k): sensor data pipeline hardening, multi-tenant architecture, API integrations with SCADA/historian systems. Sales ($55k): dedicate

Probability of success

P(reaching mid case in 12 months)

Expected take-home Y1

$-152840

probability-weighted, after investment

Go-to-market motion

Outbound to infrastructure asset managers at mid-size utilities via LinkedIn + conference networking → pilot agreement (often unpaid or deeply discounted) → 12-18 month sales cycle to full contract.

Key risks

Sensor heterogeneity hell: most infrastructure operators run 15-30 year old SCADA systems with proprietary protocols (Modbus, DNP3, OPC-UA variants) - integration alone can consume the entire Year 1 engineering budget on a single customer
Liability exposure for missed predictions: if the AI says 'no imminent failure' and a water main bursts or a substation catches fire, the vendor is named in the lawsuit - utilities will demand indemnification clauses that kill margins
Pilot purgatory: utility procurement requires board approval, CAPEX budgeting cycles, and union coordination - 60% of pilots run 18+ months before a PO is cut, if ever
Federal critical infrastructure regulations (NERC CIP, AWIA 2018 for water) require security certifications that take 9-12 months and $40-80k to obtain, blocking sales to most serious buyers until completed

Generated by the Wishdeal Factory financial-analysis agent. Numbers are honest Fermi estimates, not guarantees. Real outcomes depend on the operator. The studio is bullish on the engineering quality, agnostic on the business outcome.