PILOT WAVE
The investment thesis

The models are ready. The operations are not.

AI reached production just as North American infrastructure ran out of capacity. The companies that close that gap cannot do it alone. We buy them and do it with them.

01Adoption

Real AI spend is concentrated in tech. The rest of the economy has barely started.

Ramp AI Index, November 2026 (U.S. businesses on Ramp)

Why now
Production scaleAT&T runs 45 billion tokens a day and has cut AI costs by up to 90% by routing each task to the cheapest model that can do it.
Cost trajectoryThe same performance costs 10x less every year. Work that was uneconomic two years ago pays now.
Diffusion speedAt the current pace, AI use evens out across U.S. states in two to five years, about ten times faster than the major technologies of the last century.

Sources: AT&T, The Tokenomics Equation, July 2026; Andreessen Horowitz, Welcome to LLMflation, November 2024; Anthropic Economic Index, January 2026

02 · The deployment gap

Most AI pilots fail.

95% of organizations report no measurable return from generative AI. The failure is organizational, not technical. The people who can build have no authority. The people with authority have no time and every reason to wait.

MIT NANDA, The GenAI Divide: State of AI in Business 2025, July 2025

03Verticals

Utilities and grid

The contractors who build and maintain grid infrastructure.

Industrial manufacturing

Producers reshoring work to North America. We cannot win on labor cost, so technology has to close the gap.

Retail, e-commerce, and distribution

Operators competing on forecasting, inventory, and fulfillment.

04What we bring

Capital

We keep strong management teams in place and fund a value creation plan: new revenue, operational excellence, and add-on acquisitions.

Partners

Every partner brings sector-specific experience, relationships, and an operating record to the deal.

Technology

In-house capability that lifts margin, wins customers, and turns into new lines of business.

05The floor

Structure sets the floor. AI is the upside.

We buy at lower-middle-market multiples with conservative leverage. Businesses that grow into the middle market have historically been valued higher on size alone. We underwrite to that floor, and build on it.

TEV / EBITDA by segment · GF Data, M&A Report, H1 2025; PitchBook LCD via Cambridge Associates, Q4 2025

Large U.S. buyouts, 2025~11.0x
$100M to $250M TEV10.0x
$10M to $250M TEV, average7.2x
Under $25M TEV6.3x to 6.9x
06The seller
6M

small and mid-sized businesses facing an ownership transition by 2035

1M+

of them viable candidates for sale

$5T

of enterprise value in those businesses

Most have no succession plan. Our structure is built for that owner, and leaves them a stake in what the business becomes.

McKinsey Institute for Economic Mobility, as reported by Fortune, February 2026

07The model
01 / Buy

Conservative control

Majority ownership of cash-flowing businesses, leverage below market norms. The data, the workflows, and the customer relationships are already there. Technology is what makes them worth more.

02 / Build

Automation into margin

Our engineers build production systems from the company’s own data. Each deployment makes the next one faster and cheaper.

No guessing at product-market fit. The customers are already ours and the workflow is in front of us.

03 / Amplify

Incubation into new companies

Sometimes what gets built inside becomes a new company.

08Three answers
Financialization

Private equity’s answer. Multiple arbitrage, leverage, cost out, roll up, wait for a liquidity window. The value is engineered in the capital structure.

Disruption

Venture’s answer. Fund the replacement and let the incumbent lose. It worked on bits. Atoms are slower: factories, fleets, permits, supply chains, and customers who already have a vendor.

Amplification

Pilot Wave’s answer. Buy an established company and build AI inside it. Conservative structure protects the principal. The operators stay, the seller keeps a stake, and the transformation is the upside.

Everyone calls themselves an AI rollup now. Most are private equity firms adding an AI story, or software companies buying services businesses. The first has the structure without the build. The second has the build without the structure. It takes both, and the empathy to do it with the people who run the company.

09 · Proof

Proven from acquisition through exit.

A utility infrastructure contractor in Atlantic Canada, rebuilt around one system of record and sold to a public strategic acquirer. The AI company built inside it was not part of the sale.

Inquiries

If you invest in this market, get in touch.

Afsheen Afshar founded the firm. J.P. Morgan made him Wall Street’s first Chief Data Science Officer. Cerberus made him private equity’s first Chief AI Officer. Goldman Sachs before that. MD and PhD, Stanford.

Seven transactions and three realized exits, 2020 to 2026 · details under NDA for qualified investors

Thank you. We will reply directly.