Autonomous freight company Gatik has raised $200 million in Series D financing to expand driverless middle-mile operations across North America. Qatar Investment Authority and Koch Disruptive Technologies led the round, with Millennium Management, ARK Invest, Intact Private Capital and other investors participating.
Gatik says it has more than $600 million in contracted revenue, has completed 85,000 fully driverless orders and maintains a 99 percent on-time delivery rate. These figures come from the company and are not independently audited in the source.
Driverless freight moves beyond fixed short routes
Gatik operates medium-duty trucks between facilities such as warehouses, distribution centers and stores rather than delivering directly to consumers. Its vehicles run on highways and surface streets in Texas, Arizona, Arkansas and Canada without a driver or safety observer aboard.
The company began with short, predetermined routes. Its early Loblaw deployment in Toronto used five vehicles across five fixed pickup and delivery patterns. Gatik now says newer networks can include dozens of locations and routes as long as 400 miles, with stops adjusted as demand and distribution-center activity change.
That progression is central to the funding case. Fixed and repetitive routes limit uncertainty, while dynamic orchestration can make autonomous trucks useful across a larger share of a retailer’s network. It also introduces more variation that the driving system and logistics software must handle reliably.
Gatik currently operates dozens of fully driverless trucks. A spokesperson told Reuters that the company aims to exceed 100 by the end of 2026, then expand to thousands over the coming years. Those are targets rather than completed deployments.
Large customers provide routes and operating evidence
Gatik signed a multiyear agreement with PepsiCo in June 2026. The operation includes 41 fully driverless box trucks moving Frito-Lay products among distribution centers and stores in Dallas, Phoenix and northwest Arkansas, according to reporting cited by AI News. PepsiCo said its work with Gatik began in 2022 before the expanded agreement.
In Canada, Gatik and Loblaw agreed in 2025 to an initial 50-truck deployment serving more than 300 stores in the Greater Toronto Area. The plan called for 20 vehicles by the end of 2025 and another 30 by the end of 2026, with operations transitioning from safety drivers to fully driverless service.
These partnerships provide high-frequency commercial routes, but the source does not disclose profitability, intervention rates or how contracted revenue is recognized. An on-time percentage alone cannot describe safety performance or the operational cost of supporting the fleet.
Simulation and manufacturing become scaling constraints
Gatik uses an internal simulation platform called Arena to reproduce ordinary driving and rare hazardous situations. The system uses Nvidia Cosmos world models to generate synthetic environments for repeated training and validation. Nvidia DRIVE AGX hardware processes sensor data aboard the trucks.
Simulation can expose software to scenarios that are difficult to collect on roads, but synthetic coverage does not replace evidence from the operational design domain in which a Level 4 vehicle is approved to run. Weather, geography, road types and other limits still define where the automated system can take full responsibility.
Manufacturing is another dependency. Isuzu invested $30 million in Gatik in 2024 and is developing a redundant chassis intended to recover from autonomous-system faults. The partners target mass production from 2027.
The new financing gives Gatik capital to enlarge its fleet, infrastructure, workforce and technology program. Its challenge is no longer proving that one truck can follow a repeatable route. It must show that dynamic regional networks can grow from dozens to hundreds of vehicles without weakening safety, reliability or economics.
That transition will be judged across the entire freight operation, not only by the behavior of the driving software.
