This article was originally published on LinkedIn.
A theater sustainment area, somewhere forward.
Sixty soldiers running forklifts around the clock, breaking down container loads of ammunition, repair parts, and medical supplies. Reconfiguring loads for three dispersed brigades.
The node has been stationary for fourteen hours. It needs another six to finish processing. Adversary ISR has been watching for the last four.
The problem is not the trucks waiting to carry supplies forward.
The problem is the time those supplies are spending at the node.
The contested logistics problem hiding in plain sight
We spend a lot of time thinking about how to move supplies between nodes.
We spend far less time thinking about how to compress what happens inside them.
Yet in a contested environment, the throughput node often becomes the most targetable part of the system. It concentrates people, equipment, and materiel. It generates signature. It sits long enough to be found.
Why I believe autonomous forklifts belong in this conversation
I recently authored a whitepaper on how human-in-the-loop autonomous forklifts and remote labor pooling are transforming commercial warehouse operations.
The research focused on cross-docking environments. Pallets arrive inbound, get sorted, and move outbound with minimal storage. The operating model combines autonomous material handling with remotely pooled teleoperators.
The core effects are straightforward:
- Faster throughput time
- More consistent performance under surge conditions
- Better inventory location accuracy
- Less labor volatility and less need for standby labor
Those same effects translate directly to contested sustainment. Not as warehouse efficiency, but as survivability.
What human-in-the-loop autonomy actually looks like
This is not lights-out automation.
The system operates in three modes:
- Autonomous mode for routine, repetitive shuttle moves Pallet from inbound staging to outbound consolidation lane.
- Remote teleoperation for exceptions An obstruction, an irregular load, an ambiguous placement decision. The vehicle hands control to a remote operator who resolves the situation and returns it to autonomy.
- Local manual control when needed For truly complex or unusual circumstances.
The key point is simple. The machine handles predictable volume. A human handles judgment calls. That human does not have to be physically present at the node.
Why commercial benefits become battlefield advantages
In a commercial warehouse, this model improves throughput consistency and reduces on-site labor requirements. It also fixes a persistent data integrity problem. Because the system continuously records where freight is picked up and set down, it reduces location errors created when humans move fast and skip scans.
In a contested sustainment node, every one of those commercial benefits becomes a survivability advantage.
Smaller footprint, less signature, fewer exposed personnel
Fewer soldiers operating forklifts at the node means less life support infrastructure, less predictable activity, and fewer people exposed when fires arrive.
Faster processing, faster displacement
Continuous autonomous throughput without fatigue, shift changes, or degraded performance means the node can process faster and displace sooner. Time stationary is time targetable.
Remote labor pooling is not a staffing trick. It is resilience.
In the whitepaper, I showed that pooling teleoperators across a network can reduce the standby labor each site must carry.
In theater, the same portfolio effect applies across sustainment nodes.
When one node surges, a mass ammunition push before an operation or a medical spike after a mass casualty event, teleoperators redirect virtually.
No convoy required to move the labor. No new exposure. No delay.
Surge capacity already exists in the network. It just gets allocated to where demand appears.
Lead-time compression changes the inventory posture of the force
This is where autonomous forklifts have strategic implications beyond the node.
When autonomous material handling compresses throughput time at intermediate nodes, it compresses the effective lead time experienced by forward formations waiting on resupply.
Shorter effective lead time reduces required forward safety stock.
Less safety stock forward means lighter formations, fewer resupply convoys, less signature, and greater operational reach.
Autonomy at the node does not just help the node. It changes the inventory posture of every unit the node feeds.
What has to change for contested environments
None of this works if we transplant commercial technology directly into a contested environment without adaptation.
A contested design needs:
- Graceful degradation under denied communications Continue autonomous operations on pre-assigned missions when the teleoperation link drops, then reconnect when communications return.
- EW-hardened communications architecture With redundancy and rapid reconfiguration.
- Expeditionary deployment Containerized, rapidly deployable, and operational at an expedient site quickly.
- Power realism Designs that work with tactical generation constraints, not only warehouse power assumptions.
- Kinetic resilience If a forklift is destroyed, the teleoperator is alive at a remote operations center. A replacement vehicle can be activated and the same operator can continue immediately.
These are not footnotes. They are the design requirements that determine whether the concept survives contact.
The bottom line
The value of autonomy at a sustainment node is not labor savings.
It is time.
Less time stationary. Less time exposed. Less time searching for supplies that should already be on a truck moving forward.
In contested logistics, every hour of dwell is an hour of targeting.
Human-in-the-loop autonomous forklifts, combined with remote human judgment and network-wide labor pooling, compress that risk in ways that manual operations cannot match.
If we are serious about sustaining dispersed formations under fire, we need to design the nodes as carefully as we design the platforms that move between them.
The throughput problem is a survivability problem.
It is time we treated it like one.
