OSHA's warehouse program just got longer and narrower at the same time

If you run a distribution center, the enforcement window you are operating inside just got two years longer.

OSHA reissued its National Emphasis Program for Warehousing and Distribution Center Operations as directive CPL-03-00-026. It was signed on 6 July 2026 and took effect on 31 July. It now expires 31 July 2031. The notice that circulated in early August was OSHA publicising something that was already in force, not announcing a fresh decision.

Several write-ups described this as OSHA expanding its warehouse crackdown. That is not quite what the document says, and if you are the one being inspected the distinction is worth ten minutes of your time.

Longer, but also smaller

The program did get longer. The original ran three years from July 2023. The renewal runs five.

It also got narrower. The 2023 version covered a set of high injury rate retail establishments alongside warehousing proper: home centers, hardware stores, other building material dealers, supermarkets, and warehouse clubs and supercenters. The renewal removes those entirely. What remains is seven NAICS codes covering postal processing and distribution, couriers and express delivery, local messengers and delivery, and general, refrigerated, farm product and other warehousing and storage.

The renewal also removed the mandatory screening requirements for ergonomic and heat hazards. That reads like a softening until you notice that OSHA now runs a separate emphasis program dedicated to heat, so heat enforcement moved rather than disappeared. Reading the warehousing change as OSHA easing off heat would be a mistake.

What did not change

The core hazard list is intact, and it is the part that should shape how you think about your floor:

• Powered industrial vehicles, which in practice means forklifts
• Material handling and storage
• Walking and working surfaces
• Means of egress
• Fire protection

Inspections remain comprehensive. Area Offices also now have clearer written authority to expand an inspection based on fatalities and catastrophes, complaints, or referrals tied to establishments inside the covered NAICS codes. In plain terms: one triggering event can widen into a fuller look at the site.

The numbers OSHA chose to publish

The directive includes its own justification, and those figures are more interesting than the enforcement mechanics.

In the first eighteen months of the original program, OSHA reports identifying more than 1,700 violations and removing roughly 37,410 workers from hazards.

The injury data in the directive's own table, drawn from BLS five year averages for 2020 to 2024, puts general warehousing and storage at a recordable case rate of 5.2 and a DART rate of 4.4. Across all private industry the equivalents are 2.6 and 1.6. Couriers and express delivery run higher still, at 8.6 recordable and 6.9 DART.

And the sector kept growing through all of it. OSHA cites warehousing and storage employment rising from 882,100 at the end of 2015 to 1,836,200 at the end of 2025.

More people, in a category that injures people at roughly triple the national DART rate, for five more years of focused enforcement.

Nobody is short of information

Here is the part that has always struck me about warehouse safety, and it is not a technology observation.

None of this happens because operators do not know forklifts are dangerous. Everyone knows. Every site has the training, the high visibility vests, the painted walkways, the toolbox talks. The hazard list in this directive would not surprise a single EHS manager in the country, because it has been roughly the same list for decades.

The gap is not knowledge. The gap is attention, distributed across time and space.

An unsafe moment happens in one aisle, at one second, usually when the person who could have intervened is somewhere else entirely. A forklift and a pedestrian arrive at the same blind corner. Someone goes down between racks. A fire door is propped open with a pallet because it is genuinely easier that way. Each of those is fine a hundred times and catastrophic on the hundred and first.

And in almost every case, there is already a camera pointed straight at it, recording faithfully, for later.

Turning retrospective footage into something that acts

That gap between something happening and somebody knowing is the specific problem DHI is built to close.

DHI runs on the cameras a site already has. It watches every feed simultaneously and flags the forklift closing on a pedestrian, the person who has gone down, the blocked exit, the smoke, while there is still time to do something about it. Inference happens on a box on site, so the video never leaves the building and the only thing that goes out is the event.

Two honest boundaries. It is not a compliance product and it will not fill in your OSHA 300 logs. It also does nothing for ergonomics, which is a real hazard category on this list that computer vision does not address.

What it changes is timing. Whether a hazard is discovered by an inspector in six months, or by a supervisor in six seconds.

Five years is long enough to change something

Enforcement windows tend to produce one of two responses. Sites either brace for inspections, or they use the window as cover to fix things they already wanted to fix.

Five years is long enough for the second option to be realistic. It is long enough to run something on one camera, see what it catches, and expand only if it earns the spot. That is a considerably better use of the window than waiting to find out which aisle the inspector walks down first.