What Does the Full P&L Look Like for a Solo AI-Operated Business?

The audited 28-month P&L (February 2024 through January 2026) shows $868,147 in net revenue, $848,031 in COGS (affiliate payouts), and a -$146,783 EBITDA — a -16.9% margin on an affiliate-heavy revenue model. These numbers are QB-verified and reflect a business that survived a 99.9% revenue collapse and rebuilt on owned infrastructure. The P&L tells a story of structural transformation, not steady-state profitability.

The revenue structure. Total net revenue across 28 months was $868,147. PRJ-12 generated $652,495 (75.2% of total), driven primarily by PRD-01 at $509,821 in net revenue. The revenue was heavily affiliate-sourced: COGS consisted almost entirely of affiliate commission payouts, leaving a gross margin of just 2.3% ($20,116). This is the structural weakness the operator model was designed to address. Six revenue-generating projects contributed: PRJ-12 ($652,495), PRJ-13 ($113,009), PRJ-16 ($48,342), PRJ-05 ($29,686), PRJ-14 ($14,419), and PRJ-15 ($10,197). Ten additional projects are pre-revenue, representing the infrastructure portfolio built during the October 2025 through January 2026 acceleration.

The cost structure. Total operating expenses were $166,899 (19.2% of revenue), broken into three categories: project shared costs of $96,802 (QB-verified contractor and infrastructure spend), operating expenses of $61,721, and chargeback management of $8,376. The total build cost for all 10 production systems was $65,394 — comprising $62,731 in contractor costs (CON-02: $40,700, CON-03: $22,030) and $2,664 in AI tooling (Anthropic, OpenAI, Leonardo.AI). According to Pilot's 2024 Startup Benchmarks Report, the median early-stage SaaS company spends 55-65% of revenue on engineering alone. This operation spent 7.5% of revenue on total build costs across all 10 systems.

The cost trajectory tells the real story. Monthly operating costs peaked at $8,367 in September 2025 and fell to $0 by January 2026 as PRJ-01 displaced 6 SaaS vendors ($1,565/month eliminated) and contractor dependency dropped to zero. The current run rate is approximately $825/month — hosting and AI tools only. SaaS displacement saved $19,909 over the 28-month period, with $62,731 in annualized contractor costs eliminated permanently.

Capital structure. Total capital deployed was $733,099 — comprising $572,506 in operator reinvestment and $160,593 in outside capital. All outside capital is structured as debt, not equity. Ownership: 100% held by the operator, Michael George Keating. No equity sold. No investor dilution. Current liabilities total $139,100 across personal loans and credit obligations. Total distributions over the 28-month period were $54,659, resulting in a net-after-distributions figure of -$201,442.

The asset side. The 10 production systems (596,903 LOC, 2,561 commits) carry a market replacement value of $1.4M to $2.9M against a build cost of $65,394 — a 22x to 44x multiple on infrastructure investment. For PE evaluators, the negative EBITDA reflects the investment phase of a business that chose to build owned infrastructure instead of paying ongoing vendor and contractor costs. The question is not whether this P&L shows current profitability — it does not. The question is whether $1.4M to $2.9M in owned, unencumbered infrastructure assets, running at $825/month with zero external dependencies, represents a defensible position for future revenue generation.

Related: C7 FAQ #157 — ROI of an operator vs. a team | CS08 — The Cost Inversion

References

1. Pilot (2024). "Startup Benchmarks Report." Median early-stage SaaS engineering spend as a percentage of revenue.
2. Keating, M.G. (2026). "The Compounding Execution Method: Complete Technical Documentation." Stealth Labz. Browse papers