While You Were Watching AI...

Turn on CNBC any morning, and you’ll hear about AI, software multiples, and whether the Fed just saved or killed Silicon Valley’s next unicorn. On Twitter (now X), founders are debating whether their SaaS GTM motion is product-led or sales-led.

Meanwhile, ~160 million Americans are going to work.

Not at OpenAI. Not at Silicon Valley portfolio companies. They’re driving trucks, pulling wire, fixing HVAC units, pouring concrete, running restaurants, and maintaining the physical infrastructure that keeps the country from collapsing.

This is the real American economy, and almost nobody writing about it is the same person writing about venture capital.

The U.S. Bureau of Labor Statistics tracks employment across every industry in America. When you actually look at the data, the numbers are staggering.

Six times as many people work in construction, manufacturing, transportation, and food services combined as in all of tech, finance, and professional services combined. Yet one of those groups commands nearly all of the cultural attention, the venture capital, and the media narrative.

This is not a political statement. It’s an arithmetic one.

In 2025, AI startups alone attracted roughly a third of all venture capital deployed globally. By Q4, AI companies captured more than 50 cents of every venture dollar invested.

That’s fine. AI is genuinely transformative.

But consider the math: the industries that employ 125 million Americans, that build every home and road and power line in this country, received a rounding error of that capital. The 81% of GDP those industries represent is competing for pocket change after the model training runs are paid for.

The pitch that almost never gets made at Sand Hill Road is the one about the plumbing company that’s been operating in the same three counties for forty years, throwing off $2M in EBITDA, with a founder who’s 63 and has no succession plan. Not flashy. Not fundable. But very, very real.

Here’s where it gets alarming.

Each year, American employers post nearly 2.9 million job openings across skilled trades. Trade schools, apprenticeship programs, and community colleges collectively produce about 1.25 million qualified workers. That’s roughly four trained workers for every ten available jobs, a structural shortfall of 1.7 million workers annually.

These aren’t abstract labor market statistics. They are the reason you waited four months to get a roofer last summer. They are the reason a single experienced HVAC technician can now charge $200 an hour in a major metro and still have a six-week backlog. And why more businesses are using companies like Classet for hiring.

The infrastructure to train a tradesperson hasn’t evolved in a century. You need four to five years to become a licensed plumber or electrician. The pandemic cratered enrollment right when the retirement wave hit. And now companies are racing each other for the same dwindling pool of people who actually know how to do things with their hands.

There is a category of business that never makes headlines, never gets a TechCrunch post, never raises a Series B, and has been quietly compounding for decades.

The pest control company is servicing 3,000 routes a week on subscription revenue. The commercial HVAC contractor who has every property management company in a mid-sized city on retainer. The waste management business with exclusive municipal contracts.

These businesses share a few characteristics worth noting.

Recurring, sticky revenue. Their customers aren’t churning to a competitor with a slicker UI. Relationships in physical services are built over years and are hard to replicate.

Pricing power nobody talks about. When there’s a shortage of 550,000 plumbers, the plumbers who remain don’t compete on price. They compete on availability. (unless you are in the Houston area using: Goodsmith).

Real barriers to entry. Anyone can spin up a SaaS product. Not everyone can walk into a market and build 40 years of regional reputation, a trained workforce, and municipal relationships overnight.

Resilience. These businesses survived 2008, COVID, and every economic cycle in between. Their customers need them regardless of what the Fed does.

The irony is that the very qualities that make these businesses unattractive to venture capital, slow, capital-intensive, low margin, unsexy, are the same qualities that make them remarkably durable.

81% of the economy that early-stage capital ignores is not a charity case. It is the largest pool of underpriced, under-institutionalized businesses in the United States.

The financial and technology infrastructure built for modern tech is now mature enough to be deployed across industries that have never had access to it. An AI-enabled operator in one of these sectors looks nothing like the regional founder who built the business in 1987. The underlying asset is the same. The potential is completely different.

Most people writing about the economy are writing about the 20 million. The opportunity is in the other 125.

Alex Kirshenbaum is an investor at The 81 Collection, a high-growth equity fund that invests minority positions in businesses at the intersection of cash flow and technology