US Manufacturing Trends: Demand, Inventory, and AI Over the Past 3 Years
Over the past three years, US manufacturers have dealt with shifting demand, inventory levels that don’t always match what’s needed, and AI tools becoming part of everyday work. We looked at three years of BTOS survey data to see how these trends have actually played out across manufacturers of different sizes.
Key takeaways
- Demand has been negative across the board for three years, but a turnaround is expected everywhere.
- Larger manufacturers are more likely to have optimal inventory, but also far more likely to have too much of it.
- Every size group expects to move closer to optimal inventory, and the expected improvement grows with company size.
- AI adoption is climbing fastest for mid-sized manufacturers, not the largest ones.
Every 2 weeks, the US Census Bureau surveys 200,000 businesses.
The survey, known as the Business Trends and Outlook Survey (BTOS), asks US companies about their current and expected business conditions. Survey participants include businesses of all sizes and from a range of industries, including construction, retail, and manufacturing.
The BTOS has been taking place since September 2023. Its large sample size and bi-weekly polling frequency make it a reliable indicator of what’s actually been happening with US manufacturers over the past three years.
In our research, we analyzed three years of BTOS survey data to understand the strongest trends in US manufacturing when it comes to inventory, demand, and AI use. Here is what we observed.
Demand on the rebound
As part of the survey, the BTOS asks manufacturers the following questions about demand:
- In the last two weeks, how did demand for your goods or services change?
- How do you expect this to change in six months?
Survey responses came from manufacturers across every company size group, from the smallest businesses (1 to 4 employees) up to the largest (250 or more employees).
We averaged three years of survey responses, then subtracted the share reporting falling demand from the share reporting rising demand. This gives each group’s “net” demand, both now and six months from now. Here’s what we saw:
Averaged across the survey period, more manufacturers reported falling demand than rising demand, regardless of company size. The current net was negative for every group.
The size of this gap varied widely, though.
For the smallest manufacturers (1 to 4 employees), the share reporting falling demand far outweighed the share reporting rising demand. The gap shrank steadily with company size, and was smallest for manufacturers with 250+ employees.
On the other hand, when manufacturers described what they expect six months ahead, the outlook was positive.
Manufacturers from every size group expected demand to rise, rather than fall. The following chart shows the difference between currently reported and expected demand (averaged over the 3-year period).
Demand has cratered, but a turnaround is coming
Net share reporting rising vs. falling demand (Increased − Decreased), now vs. expected in 6 months, by company size
- Current net – last 2 weeks
- Expected net – 6 months out
Inventory trending in the right direction
When looking at how manufacturers currently describe their inventories, there’s a clear pattern:
The share describing their inventory as “oversized” rises steadily with company size. The “optimal” share follows a similar upward pattern overall, though it dips slightly for manufacturers with 100 to 249 employees before climbing again at the largest size group.
At one end of the spectrum, 36.9% of very small manufacturers (1 to 4 employees) describe their inventory as optimal. Just 6.6% say it’s oversized.
57.3% of the largest manufacturers in the survey (250+ employees) say their inventory is optimal. Meanwhile, 30.9% say it’s oversized.
This shows that larger manufacturers are more likely to have optimal inventory levels, but they’re also far more likely to have too much inventory.
Inventory ratings by company size
Share currently describing inventory as “optimal” vs. “oversized”
| Company size | Optimal | Oversized |
|---|---|---|
| A1–4 | 36.9% | 6.6% |
| B5–9 | 46.6% | 7.9% |
| C10–19 | 50.6% | 10.9% |
| D20–49 | 54.9% | 15.9% |
| E50–99 | 56.2% | 23.2% |
| F100–249 | 55.6% | 29.2% |
| G250+ | 57.3% | 30.9% |
Looking ahead, the share of manufacturers expecting “optimal” inventory levels six months out also increased. This was true for manufacturers of every size group, with the difference growing steadily with company size.
- For the smallest manufacturers (1 to 4 employees), 41.9% expect optimal inventory in 6 months, compared to 36.9% currently reporting optimal inventory.
- For manufacturers with 100 to 249 employees, 70.6% expect optimal inventory in 6 months, compared to 55.6% currently reporting optimal inventory.
- For manufacturers with 250+ employees, 73.1% expect optimal inventory in 6 months, compared to 57.3% currently reporting optimal inventory.
This gap makes sense in light of where each group starts. Larger manufacturers currently carry the most excess inventory: nearly a third of the largest manufacturers describe their inventory as oversized.
That gives them more room to improve, since many expect to bring that excess back down to an optimal level. Smaller manufacturers report far less oversized inventory to begin with, so they have less room to correct on that front.
The optimal share keeps climbing
Share describing inventory as “optimal,” current vs. expected in 6 months
| Company size | Current | Expected |
|---|---|---|
| A1–4 | 36.9% | 41.9% |
| B5–9 | 46.6% | 52.3% |
| C10–19 | 50.6% | 57.3% |
| D20–49 | 54.9% | 63.3% |
| E50–99 | 56.2% | 67.6% |
| F100–249 | 55.6% | 70.6% |
| G250+ | 57.3% | 73.1% |
AI adoption among US manufacturers
The survey also asks businesses if they’ve used AI in the last two weeks, and whether they expect to be using AI at all in six months. The BTOS first asked about AI use in 2023, but broadened the question’s wording in December 2025 to capture AI use across more business functions.
Our analysis uses only data collected under that current wording, so the comparison is still based on a limited number of survey cycles.
Regardless of the small sample size, the survey responses provide insight into AI use among small to mid-sized manufacturers in the US.
AI use is climbing across the board
We looked at the 3 earliest survey cycles with AI-related questions, then compared the responses with the 3 latest cycles. Our analysis showed that actual AI use has been climbing for manufacturers of most sizes:
- The smallest manufacturers (1 to 4 employees) show a strong increase in AI use, from 10.2% to 19.3%.
- Manufacturers with 20 to 49 employees show a smaller rise, from 13.9% to 20.6%.
- Manufacturers with 50 to 99 employees show one of the sharpest rises, with AI use climbing from 14.4% to 30.0%.
- AI use for manufacturers with 100 to 249 employees rose from 19.2% to 27.7%.
One group in particular showed growing uncertainty about the future. For manufacturers with 20 to 49 employees, “don’t know” responses to the six-month AI-use question rose from 22.5% to 28.7%, even as their own current-use number climbed. This shows that manufacturers in this size group are using AI more, but also less clear about future AI use.
AI doesn’t need a big budget
AI use rose sharply for manufacturers with 50 to 99 employees. But this increase wasn’t limited to one group.
Almost every other size bracket shows a real, statistically reliable rise in the use of AI. Only manufacturers with 250+ employees fail to show a reliable trend in either direction.
This suggests AI adoption doesn’t follow the “bigger companies move first” logic. It also shows how accessible a lot of AI tools have become. A small manufacturer doesn’t need a dedicated IT team or large budget to start using AI, as it would for larger-scale automation or equipment investment.
A similar shift already played out with manufacturing ERP software. Cloud-based platforms like MRPeasy removed much of the infrastructure cost and licensing overhead that once made ERP systems practical mainly for larger manufacturers, putting the same kind of planning and production tools within reach of much smaller shops.
You might also like: Hiring Trends for US Manufacturers: What 3 Years of Census Data Shows
Get manufacturing know-how delivered to your inbox!
