Onfleet

Somewhere, right now, a driver's phone just buzzed

It is a Tuesday. A grocery van pulls away from a warehouse in one city while a pharmacy courier idles at a red light in another. Neither driver has looked at a paper manifest. Neither dispatcher is squinting at a whiteboard. Both routes were drawn, ranked, and assigned by the same piece of software running quietly in the background: Onfleet.

Onfleet is a San Francisco company that builds a last-mile delivery platform. That phrase does a lot of unglamorous work. Strip it down and it means this: when a business needs to get a physical thing from a building to a doorstep, Onfleet decides who carries it, which way they drive, when it will arrive, and how everyone finds out. The company has done this more than 250 million times, across 90-plus countries, for roughly a thousand businesses.

You have almost certainly been on the receiving end without knowing it. The wine that showed up in an hour. The prescription that arrived without a trip to the counter. The text with a live map and a name. A lot of that runs on rails Onfleet laid.

The last mile is short. It is also the hard part.

Move a shipping container across an ocean and the cost per item is almost nothing. Move that same item the final mile to a house, and the economics fall apart. The last mile is the slowest, most expensive, most failure-prone leg of any delivery. Wrong addresses. Missed windows. Drivers calling dispatch. Customers calling support. A delivery that costs pennies to ship can cost dollars to actually deliver.

For years, the businesses doing this ran on a patchwork: spreadsheets, group texts, a dispatcher with a good memory, and a lot of hope. It worked, in the way that juggling works until someone adds a fourth ball. The moment a small delivery operation tried to grow, the whole improvised system buckled.

That gap - between the simplicity of one delivery and the chaos of many - is the tension Onfleet was built to resolve. Not the romantic part of commerce. The part that breaks.

Three engineers, one stubborn idea

Onfleet traces back to 2012 and a group of engineers connected to Stanford who had spent time building location-based tools for last-mile delivery. They noticed something that should have been obvious and somehow was not: the companies actually doing deliveries had almost no good software to run them. The big carriers had built private fortresses of logistics tech. Everyone else got spreadsheets.

The bet, placed by co-founders Khaled Naim, David Vetrano, and Mikel Carmenes Cavia, was that delivery software did not need to be a custom fortress. It could be a product - something a coffee roaster, a grocer, and a pharmacy could all switch on in an afternoon. They launched commercially in 2015 and grew, at first, on angel money and customers rather than a war chest.

It is a modest-sounding bet. It also happened to land exactly as the world decided it wanted everything delivered. By the time on-demand delivery became a habit rather than a novelty, Onfleet already had the product running.

One platform, from the dispatch board to the doorstep

Onfleet's pitch is that the entire messy middle of delivery should live in one place. A dispatcher uploads the day's orders; the software plans the routes. Drivers get a mobile app instead of a phone call. Customers get a live map and a predictive ETA instead of a four-hour window. And at the door, a photo or signature closes the loop.

The API is the quiet power move. A business does not have to use Onfleet's screens at all - it can wire the routing and tracking engine straight into its own checkout and app. That is why Onfleet shows up in places where you would never see its name.

Names you know, running on software you don't

The case for Onfleet is not a slogan; it is a customer list. Kroger. Total Wine & More. Drizly. Foxtrot Markets. Zumiez. GrubMarket. These are not delivery startups testing an idea - they are established operators that decided building their own dispatch stack was not worth it.

The reach skews across industries that, on paper, share nothing: food and beverage, grocery, pharmacy, cannabis, retail, and e-commerce. What they share is the same Tuesday-afternoon problem - a lot of things that need to reach a lot of doors, on time, with proof.

Investors noticed even if customers didn't. Onfleet has raised north of $40 million across its rounds, with the 2022 Series B led by Kayne Partners alongside Almanac Insights, TB Ventures, and Savant Growth. The team - around 90 people spread across a dozen-plus countries - was once named the #2 best small place to work in the Bay Area, which is a strange honor for a company that mostly thinks about other people's traffic.

Make moving goods efficient - and, oddly, delightful

Onfleet describes its purpose as powering the future of commerce by helping businesses of all sizes move goods more efficiently and delightfully. That last word, delightful, sits awkwardly next to fleet routing and proof-of-delivery scans. Logistics is not a field that throws the word around.

But it is the whole bet, restated. The argument is that delivery does not have to be a black box that ends in a shrug and a missed package. It can be a clean, tracked, predictable experience - for the business dispatching it and the person waiting on it. Efficiency is the engineering. Delight is the point.

The doorstep is not getting quieter

Local commerce keeps moving onto wheels. Groceries, medication, hot food, retail returns - more of it expects to arrive at a door, soon, with a tracking link. Every business that joins that shift inherits the same hard last mile, and most of them have no interest in building delivery software from scratch. That is the long tailwind underneath Onfleet.

The competition is real - Bringg, Routific, DispatchTrack, and the ever-present temptation to just build it in-house. Onfleet's wager is that a focused, API-first platform beats a homemade whiteboard every time, and beats a bloated suite most of the time.

Back to that Tuesday. The grocery van has finished its route; the pharmacy courier has dropped the last package and snapped the photo. Two drivers, two cities, one platform - and two customers who got a text that turned out to be true. A decade ago that whole scene ran on guesswork and group chats. Now it runs on software you will never see, which is precisely how Onfleet wants it.