May 6, 2026 — If you were a time traveler from 2000 coming forward to 2026, you'd think your machine was broken. You'd step onto the shop floor and see the same sight 26 years later — workers at their stations and machines running, using the same tech to manage their work.
If you read anything online, you'd think everyone is out there implementing AI, making obscene amounts of money and increasing their efficiency by thousands of percent. But a new study analyzing U.S. Census Bureau data paints a completely different picture in manufacturing: 87% of U.S. manufacturers have yet to integrate AI into their operations.
The report, published by cloud ERP provider Digit Software, draws on the Census Bureau's Business Trends and Outlook Survey and tracks AI adoption across the manufacturing sector over nearly four years. What it finds is a technology that is genuinely gaining ground, but far more slowly, and far more unevenly, than the headlines suggest.
AI’s presence on the shop floor is increasing — slowly
In September 2023, 1.8% of manufacturers reported using AI in any business function. Fast-forward to February 2026, and that figure has increased to 13.9%.
The trajectory of adoption is impressive when you look at it as a multiple, but considerably less dramatic when you remember that in 2026, nine out of ten manufacturers are yet to implement it in their business, likely due to the fact that most manufacturers simply don’t have the infrastructure in place to incorporate AI.
According to the data, size matters — if there is one variable that predicts AI adoption in manufacturing more than any other, it is the size of the business. Large manufacturers are 2.3 times more likely to use AI than small manufacturers, with an average adoption rate of 28.8% among enterprise facilities.
Manufacturers want AI, but are skeptical
17.4% of manufacturers say they plan to adopt AI, and their desire to adopt outpaces the 12.6% already using it. More manufacturers are waiting to adopt than have already done so, and that gap between intention and action has been consistently present throughout the entire four-year observation period.
But as more manufacturers become aware of AI, uncertainty interestingly rises alongside it. The share of manufacturers answering "do not know" when asked whether they would want to use AI in their workflows also increased from 9.2% to 14.4% over the observation period. The technology is evolving faster than many manufacturers can evaluate it, and for an industry that prizes certainty, that is its own kind of obstacle when deciding whether to invest in this new technology.
But regardless of whether they love or fear AI, over time, AI adoption will likely increase (and maybe reluctantly for those who are uncertain), driven by competitive pressure and the growing integration of AI functionality into standard software tools.
For the 87% still on the sidelines, the trend of AI adoption is only rising, and the question for those who haven't yet adopted AI is whether they want to get ahead of almost the entire industry or play catch-up with the competitors who are riding the wave. Maybe if our intrepid time traveler decides to jump to 2031, they might finally realize that their machine is working as expected and can finally put away that warranty card.
The full report, including sector breakdowns, size comparisons, and state-level data, is available at Digit.
