Photo provided by DepositPhotos
For most of the digital era, the story was predictable. A new technology arrived, big companies bought in first, and small businesses spent years catching up. Broadband followed that script. So did the personal computer. The gap between the corner shop and the corporate campus was measured in years, sometimes decades.
AI is not following the script.
What the numbers show
Start with the anomaly. Between two federal survey windows six months apart, small-business AI use climbed from 6.3% to 8.8%, according to the U.S. Small Business Administration's Office of Advocacy. That alone is unremarkable, as adoption curves usually go up. The surprise is what large firms did over the same stretch. Their usage did not climb. It slipped from 11.1%, with the SBA noting that part of the shrinking gap came from a decline in large-firm adoption rather than small firms simply running to catch up.
People are also reading…
So the divide is closing from both directions at once. The SBA's read on its own data: Small businesses may be only about a year behind large ones.
Why does that matter? Because the last time a general-purpose technology swept through American business, the gap looked nothing like this.
The broadband comparison
Consider the benchmark the SBA itself reaches for. When broadband was nearly universal among large enterprises in 2004, only 48% of small businesses reported having high-speed access. More than a quarter, 27%, had no internet subscription at all. That was the divide a foundational technology produced: half of small firms locked out of the fast lane, a quarter not on the road at all.
Against that history, a gap of a few percentage points that is actively narrowing is not a minor improvement. It is a different pattern entirely.
Who is actually paying
Survey responses can flatter a trend. People say they use things they have tried once. So the sharper signal comes from money changing hands.
Here the JPMorganChase Institute offers a harder measure. Tracking de-identified business banking transactions rather than self-reported answers, it found that 17.7% of small businesses had started paying for AI services by the end of 2025. Payments are a commitment in a way that a survey checkbox is not. A business that keeps a subscription active is telling you something a questionnaire cannot.
What are they buying? Increasingly, tools that run the parts of a business that used to require dedicated staff: forecasting, planning, inventory replenishment. Software once built for enterprise operations teams now ships at small-business prices, which is how a single manufacturer can reach for the same kind of AI operations tooling that used to sit behind a corporate IT budget. Platforms like Fishbowl's AI for manufacturing sit in that category, aimed at operations work that a small shop previously had to handle by hand or not at all.
That shift raises the next question. If small firms across the board are moving, are some industries moving faster than the headline numbers suggest?
The manufacturing signal
One sector stands out. Reported AI adoption in manufacturing jumped 159% between two iterations of the Federal Reserve's business survey, the steepest climb among the industries the Fed tracked. The figure carries a caveat worth stating plainly: Part of that leap reflects the Census Bureau broadening how it asks the question, from AI used in "producing goods or services" to AI used in "any business function." Widen the definition and you capture more activity.
Even so, manufacturing posted the largest jump of any sector measured, and it is a corner of the economy long assumed to be slow to digitize. When the shop floor starts registering the fastest gains, the "AI is only for tech giants" framing gets harder to defend.
Why the pattern broke
The mechanics behind the reversal are not mysterious. Earlier technologies demanded capital up front: broadband contracts, server rooms, hardware, specialists to run all of it. AI arrived with a subscription button. Entry-level tools that cost a small firm a few hundred dollars a year removed the wall that kept smaller players out during past technology shifts, when the price of admission alone decided who participated.
That does not mean the field is level. Plenty of small businesses still say AI is not relevant to what they do, and adoption remains uneven across regions and sectors. But the direction is set, and it points somewhere the previous playbook did not.
The broader stakes
For years the assumption held that each new wave of technology would widen the distance between large and small, handing scale advantages to whoever could spend the most first. AI is testing that assumption. When the tools that separate a competitive business from a struggling one cost roughly the same for a two-person operation as for a corporation, the advantage of sheer size gets smaller.
The federal data does not prove small firms will pull even. It does show a gap closing at a speed no earlier technology matched, and it shows the corner shop, for once, is not the one waiting to catch up.
