This month, Google committed $50 million to prepare more than 300,000 American workers for skilled-trade careers across more than 20 states. The goal is to scale apprenticeships, fund training, and modernize coursework for various other trades. But Google’s not the only one. Meta also recently announced a $115 million workforce academy .

Hidden behind the marketing copy and press releases is one realization: labor, not chips or capital, is most likely to become the constraint that could slow the entire AI buildout. With 2.1 million skilled-trade jobs projected to go unfilled by 2030 , are doing everything they can to help bridge the gap.

That’s the headline story, and it’s real. But it points at something larger and far less discussed. The most consequential thing AI is about to do is not invent a superintelligent agent or place a server farm in space. AI is rapidly accelerating to support industries that have historically been overlooked despite representing the backbone of our global economy.

The Industries Software Forgot

In 2011, Mark Andreesen famously wrote his essay “ Why Software Is Eating The World ” and it seems the industry never looked back. Despite the rapid development of software across every industry over the last 15 years, the trades were largely left out. Some of this due to lack of understanding from the owners of trades businesses, some of it from the bubble effect of Silicon Valley where most engineers are never exposed to the dirty, paper-filled back office of a services business.

More recently that vantage has seemed to change. The children of service business owners also grew up, learned how to code, and wanted to make an impact on their family business. A handful of early apps took off and then investors realized the U.S. construction and home-services market is roughly $2.1 trillion a year yet they run on almost no modern software, let alone AI. Together, this combination has created a wave of innovation that is only accelerating with the dawn of AI capabilities.

The Leap From Paper And Pencil To AI Native

Let’s start with answering the phone. A restaurant that misses a call loses a $30 takeout order; a home-service business that misses one can lose a $30,000 install. That asymmetry gave Avoca its opening. The startup raised $125 million to build AI voice agents that answer calls. The service quotes, handles objections, books the visit, and chases the estimates that never closed. Its founders are clear about the line: the technician is the main character, and no AI wave is replacing the person who crawls into the attic. But its plenty capable of managing or augmenting administrative work.

The phone is only the first chore getting handed to software; together a handful of companies are assembling a full tech stack for the trades.

Balto rides a live call, feeding a rep real-time info to help support the customer or close the deal, auto-scoring the conversation afterward to identify room for improvement. Rilla records sales pitches in-person and on the phone, and turns those recordings into analytics that show where a deal was lost. Driive , a scheduling tool with drive-time in mind, built a scheduling engine that bakes drive time into every booking so technicians roll onto jobs that fit the route rather than one that prioritizes the customers’ schedule in isolation. CompanyCam , recently valued at $2 billion , flipped the process of documenting and planning of jobs on its head years ago by turning the photos and videos a crew was already snapping into organized records and AI-generated reports that make managing the job and closing deals easier.

Beyond the various point solutions there are also platforms trying to become the whole operating system.

ServiceTitan , the heavyweight, runs on data from more than $80 billion in annual transaction volume. Housecall Pro , serving 45,000-plus businesses, hit a $1.1 billion valuation building the same for smaller organizations; Jobber serves the mid-market; Procore does it for commercial construction. And the opportunity is only growing with investments into the trades from companies like Google, Meta and others who rely on the trades for more efficient infrastructure development.

Trade Labor Will Grow, Augmented By Artificial Intelligence

There’s a reason AI shows up throughout the trade industries as a tool, not a threat, and it’s the same reason workers are flooding toward these jobs. The trades require physical presence: someone still has to install the equipment, replace the valve, and understand old plumbing systems that aren’t mapped on Reddit waiting to be crawled for training data.

Geoffrey Hinton, the Nobel laureate often called the godfather of AI, has compressed his career advice to one blunt insight: train to be a plumber. The labor math is durable, too. Every year more tradespeople retire than enter the field. And around half a million skilled-trade jobs sit empty nationwide. The software’s value rises precisely because the human work is so secure: the platforms aren’t automating the business out of existence, they’re freeing a permanent, in-demand workforce to spend its hours on what customers actually pay for and what they enjoy doing.

The most durable winners of the AI era may not be only the model builders and chipmakers competing for headlines. They may be the deeply ordinary service businesses that keep the physical world running, finally handed software smart enough to run the office while humans do everything software can’t. This is what technology leaders need to keep an eye on: The next great software company may be more likely to be answering a plumber's phone than chasing the next shiny object out of Silicon Valley.