Contents
- Aggregate EBITDA is the number every PE evaluator asks for first.
- The monthly EBITDA trajectory across 28 months reveals three distinct phases that the aggregate number completely masks.
- The margin trajectory is driven by three structural changes, all enabled by the infrastructure build during 2025.
The Setup
Aggregate EBITDA is the number every PE evaluator asks for first. It is also the number most likely to mislead when applied to a business undergoing a structural transition. A 28-month EBITDA of -$147K tells you the operation lost money. It does not tell you whether the operation is accelerating toward profitability or spiraling toward zero. The trajectory — the month-by-month margin vector — is the signal that aggregate EBITDA obscures.
According to a 2024 Morgan Stanley analysis of private market valuations, PE firms increasingly weight margin trajectory over absolute margin in evaluating early-stage and transitioning portfolio companies. Bain & Company's 2024 research on value creation levers found that portfolio companies demonstrating a 15+ percentage point EBITDA margin improvement during the hold period generated 2.3x higher returns than peers with stable but higher absolute margins. The market has learned that direction matters more than position.
The Stealth Labz operation, run by Michael George Keating, moved from -$57,050 monthly EBITDA (February 2024) to +$2,676 monthly EBITDA (July 2025) — a $59,726 margin swing over 17 months. The mechanism was not revenue growth. Revenue declined 97% over the same period. The margin improvement was driven by infrastructure replacement: eliminating external cost dependencies while building owned operational capacity.
What the Data Shows
The monthly EBITDA trajectory across 28 months reveals three distinct phases that the aggregate number completely masks.
Phase 1 — High Revenue, Deep Losses (February-April 2024). The three highest-revenue months were also the three worst EBITDA months. February 2024: $311,443 in revenue, -$57,050 EBITDA. March 2024: $134,984 in revenue, -$36,286 EBITDA. April 2024: $142,699 in revenue, -$9,339 EBITDA. Combined: $589,126 in revenue generating -$102,675 in EBITDA. The operation was destroying approximately $0.17 for every dollar of revenue at its peak scale. Affiliate payouts exceeded revenue in the first two months, and operating expenses consumed whatever gross profit remained.
Phase 2 — The Trough (May 2024-June 2025). Revenue fell below $30K/month and stayed there for 14 months. EBITDA oscillated between -$7,447 (September 2024) and +$3,330 (June 2024). The average monthly EBITDA during this phase was approximately -$2,400. The losses were smaller in absolute terms but persistent — the business was burning cash slowly rather than quickly. During this period, operating expenses ranged from $2,299 to $15,728 monthly, reflecting ongoing infrastructure investment (particularly contractor costs for PRJ-01 development through CON-02 and CON-03).
Phase 3 — Margin Inflection (July 2025 onward). July 2025 marked the EBITDA inflection: +$2,676 on $9,673 in revenue. This was a 27.7% EBITDA margin — achieved not through revenue growth but through cost structure transformation. Total operating expenses in July 2025 were $2,450, compared to $9,814 in February 2024. The cost reduction was structural: PRJ-01 replaced $19,909 in annual SaaS vendor costs (Konnektive at $583/month, TrackDesk at $499/month, Klaviyo at $60/month, and others), and all contractor dependencies were eliminated.
According to McKinsey's 2024 research on AI-enabled operational efficiency, companies deploying AI for internal operations (rather than as a product feature) achieve median cost reductions of 20-35% in the first 18 months. The Stealth Labz cost trajectory exceeded this benchmark: monthly operating costs fell from a peak of $8,367 (September 2025, during active PRJ-01 development) to approximately $825 by February 2026 — a 90% reduction. The AI-enabled build methodology (total AI tooling cost: $2,664 over the full period) contributed to the speed and cost of the infrastructure build that drove the cost structure change.
The subsequent months confirmed the inflection was structural, not anomalous. December 2025: +$1,346 EBITDA on $3,642 revenue (37.0% EBITDA margin). January 2026: +$541 EBITDA on $663 revenue (81.6% EBITDA margin). The absolute numbers are small. The margin structure is durable.
The full 28-month EBITDA trajectory in quarterly summary:
| Quarter | Revenue | EBITDA | EBITDA Margin |
|---|---|---|---|
| Q1 2024 (Feb-Apr) | $589,126 | -$102,675 | -17.4% |
| Q2 2024 (May-Jul) | $47,839 | $720 | 1.5% |
| Q3 2024 (Aug-Oct) | $9,130 | -$16,086 | -176.2% |
| Q4 2024 (Nov-Jan) | $25,997 | -$6,657 | -25.6% |
| Q1 2025 (Feb-Apr) | $37,225 | -$5,177 | -13.9% |
| Q2 2025 (May-Jul) | $21,903 | -$1,703 | -7.8% |
| Q3 2025 (Aug-Oct) | $127,768 | -$12,844 | -10.1% |
| Q4 2025-Jan 2026 | $9,159 | -$2,361 | -25.8% |
How It Works
The margin trajectory is driven by three structural changes, all enabled by the infrastructure build during 2025.
First, the COGS transformation. During the affiliate-dominated phase, COGS ranged from 97-115% of revenue. As owned traffic (STL channels) replaced external affiliate traffic, COGS dropped to 23-47% of revenue in the owned-traffic months. This is not a negotiation outcome — it is a structural change in the cost of acquiring a customer.
Second, the SaaS displacement. PRJ-01 replaced $1,565/month in SaaS vendor costs (Konnektive CRM, TrackDesk, Klaviyo, SendGrid/Twilio, Socioboard, Sonetel) with owned infrastructure. By January 2026, monthly vendor dependency was $0. The build cost of PRJ-01: $16,800. The annual SaaS displacement: $18,780. The payback period: under 11 months.
Third, the contractor elimination. Pre-CEM contractor costs totaled $62,731 (CON-02: $40,700; CON-03: $22,030). By December 2025, monthly contractor costs were $0. The operator — enabled by AI-assisted development methodology — replaced all external development dependency with direct execution. Total AI tooling cost for the entire build: $2,664. Cost per line of code across the 596,903-line portfolio: $0.06.
What This Means for Decision-Makers
The -$147K aggregate EBITDA is real. It is also the wrong number to evaluate. The February 2024 P&L (-$57K EBITDA on $311K revenue) and the July 2025 P&L (+$2.7K EBITDA on $9.7K revenue) represent two fundamentally different businesses operating under the same entity. The first destroys margin at scale. The second generates margin at any scale.
For PE evaluation, the critical question is: which business is the one going forward? The cost structure as of February 2026 — $825/month operating cost, zero vendor dependencies, zero contractor dependencies, 10 revenue-ready systems, 100% ownership retained — is the current state. Revenue reactivation on this cost base produces margin immediately because the infrastructure cost is already sunk and the operating cost is de minimis. The trajectory from -$57K to +$2.6K monthly EBITDA is not a recovery story. It is a cost structure transformation story — and the transformation is complete.
Related: [C7_S148 — 28 Months of P&L Data] | [C7_S152 — Revenue Distribution Across 18 Business Lines] | [C7_S153 — The 97.7% COGS Problem]
References
- Morgan Stanley (2024). "Private Market Valuations." Research on PE firms weighting margin trajectory over absolute margin in portfolio evaluation.
- Bain & Company (2024). "Value Creation Levers." Analysis of EBITDA margin improvement as a predictor of hold-period returns.
- McKinsey & Company (2024). "AI-Enabled Operational Efficiency." Median cost reduction benchmarks for companies deploying AI for internal operations.
- Keating, M.G. (2026). "The Compounding Execution Method: Complete Technical Documentation." Stealth Labz. Browse papers