Pre-seed · Investor Overview

A serious wedge into
a $70B industry.

Trash collection runs on 1960s infrastructure. The biggest inefficiency is not hauling; it's stopping. Robotrash is building the constrained first version of a system that eliminates most of those stops.

$70B

US municipal waste collection market, annually

300–600

Individual stops per residential truck route, every day

$6–8

Estimated labor and fuel cost per stop (SWANA, 2023)

The problem

The cost isn't hauling trash.
It's stopping for it.

Today's model sends large trucks from stop to stop, house to house, day after day. Each stop adds labor, fuel, time, emissions, and route complexity. Nothing about this has changed structurally since the 1960s.

300–600

Stops per route, every day

The truck idles, the driver dismounts, lifts, empties, and repeats. Hundreds of times per route. That sequence is the cost center. It has never been automated.

$6–8

Cost per stop, not per ton

Each stop costs operators an estimated $6–8 in labor and fuel. Across millions of routes, that's tens of billions in annual inefficiency with no structural fix in sight.

SWANA benchmarking report, 2023

40%

Annual driver turnover

Sanitation driver turnover averages 40% annually. Routes get combined, pickups get missed, operators face contract penalties. The workforce problem compounds every year.

SWANA Workforce Survey, 2023

The thesis

The architecture is wrong.
Replacing the driver doesn't fix it.

The real problem is that the truck goes to the waste. The waste should come to a container. Robotrash changes the architecture, not just the labor input.

What's broken

Trash systems don't talk to each other. Fill levels aren't tracked in real time. Routes run on fixed schedules regardless of demand. A residential route is 300–600 stops of the same mechanical sequence.

The truck goes to the waste. The waste should come to a container.

Replacing the driver doesn't fix this. The collection architecture is fundamentally wrong.

Why now

LiDAR: $75,000 in 2017, under $500 in 2024. Sidewalk-capable electric drivetrains are now cost-competitive at sub-1-ton scale. Computer vision for pedestrian environments is solved at the foundation model level.

The hardware to build this finally costs less than a used car.

Three cost curves bottomed out simultaneously. That's the opening.

What we're building

The first version does not require perfect city-wide autonomy. Fixed routes, remote assist, bagged trash only, controlled environments. The navigation stack is well-understood.

Start constrained. Prove the model. Expand the environment.

The active engineering challenge is the waste acceptance mechanism: the specific problem no delivery robot addresses today.

Why now

Three structural shifts
converged in 2023–2024.

This window didn't exist three years ago. The company that builds the first deployable version now controls the physical infrastructure, and that infrastructure isn't software-switchable.

−90%

Hardware costs collapsed

LiDAR dropped from $75,000 to under $500 per unit. Electric drivetrains are now cost-competitive at sub-1-ton scale. The hardware to build a deployable rover finally costs less than a used car.

+34%

Labor costs keep climbing

Sanitation wages have risen 34% since 2020 (BLS). Driver retention is at historic lows. Every year the status quo remains, the economic case for automation strengthens.

2035

Regulation is forcing the shift

The EPA's Clean Trucks Plan (Phase 3) sets strict zero-emission targets for vocational fleets through 2032. State mandates require 75% of new garbage trucks to be zero-emission by 2035. Operators are already budgeting for this transition.

What we believe

Three non-obvious bets.

01
The incumbents will not build this. Republic Services and Waste Management are optimized around the truck. Automating last-mile collection cannibalizes their largest workforce cost center before they have a replacement revenue model. They will buy this eventually; they won't build it.
02
The real buyer is not the resident. It's the HOA property manager who writes one service contract and fields every missed-pickup complaint. Their incentive to switch is structural. They don't need to love the product; they need one fewer budget problem.
03
Zone station placement is the moat. Whoever places the first containers in a controlled environment controls the physical collection point. That infrastructure isn't software-switchable. First mover wins the zone for years.

The founder

One founder.
Very specific
obsession.

Babatunde Fashola

Founder & CEO

Built large-scale systems used by millions at Twitter/X. Now applying software-first infrastructure thinking to one of the most overlooked physical systems: waste collection.

The gap between how trash collection works today and how it should work is too specific and too large to leave alone. The first version doesn't need to be perfect. It needs to be deployable.

IBM Watson Twitter / X Affirm

Long-term vision

On-demand waste
infrastructure.

Robotrash is not just a robot. It's a new collection model: local robotic pickup, neighborhood aggregation, and dramatically fewer truck stops. Starting constrained. Scaling from there.

Local pickup

Rovers handle the last-meter problem inside controlled zones. No truck needed at each door. Collection happens on demand or on schedule, not on a fixed route tied to driver availability.

Neighborhood aggregation

Zone stations consolidate waste from dozens or hundreds of units into one pickup point. The infrastructure sits at the neighborhood level, not at each home.

Fewer truck stops

The truck still hauls. But instead of stopping at every house, it stops at a small number of zone stations. Same waste. Far less labor. Better unit economics for operators at every scale.

A serious wedge intoa $70B industry.

The cost isn't hauling trash.It's stopping for it.