Operational Excellence (OPEX) Insight – Tuesday - August 10, 2026: Doritos Without a Driver: PepsiCo Goes Fully Autonomous on Its Most Repeatable Lanes.
Góc Nhìn Vận Hành Xuất Sắc – Thứ Ba, Ngày 10/08/2026: Doritos Không Tài Xế: PepsiCo Tự Lái Hoàn Toàn Trên Những Tuyến Lặp Lại Nhất.
Welcome To Operational Excellence (OPEX) Insight Article For The Paid Subscriber-Only Edition.
This is the bilingual post in English and Vietnamese. Vietnamese is below.
Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.
English
PART 1 – OFFICIAL INFORMATION
When one of the world’s largest consumer-goods companies announces it is running driverless trucks hauling snacks on public roads, the reader’s first reflex is to think about the technology: sensors, artificial intelligence, self-driving vehicles. But the most notable detail in PepsiCo’s story this week lies not in how smart the truck is, but in which roads it was set loose on.
According to the disclosure, PepsiCo, together with autonomous-freight company Gatik, is operating 41 box trucks branded Isuzu in North America, distributed as 35 in Arizona, 5 in Texas, and 1 in Arkansas. This fleet hauls Frito-Lay products between plants, warehouses, and retailers such as Walmart and Dollar General. Gatik calls it the largest commercial autonomous-freight deployment to date, and reports an on-time delivery rate above 98% across its operations.
Two more numbers sharpen the picture. First, on safety: PepsiCo said the vehicles have not yet had an accident on a public road. Second, on the rollout path: the program began in 2022 with safety drivers riding along, and only transitioned to fully driverless operations in June 2025. That means the leap to “driverless” was not the flip of a switch, but the result of years of running on the same roads, accumulating data and honing the process.
And here is the core detail, the one easily missed amid the glossy buzzwords: these trucks run on cameras, radar, and lidar over short regional routes with very high repeatability. Not sprawling cross-state journeys full of surprises, but fixed, familiar lanes traveled back and forth every day. PepsiCo did not throw its self-driving trucks into the hardest, most chaotic part of trucking; it chose precisely the most regular, most predictable part to hand to the machine.
In other words, the success story here is not the story of a machine smart enough to handle everything, but the story of choosing the right kind of work to automate. The above-98% on-time rate and the accident-free record did not fall from the sky thanks to an algorithm; they came from placing the technology in an environment already narrowed in variability to the point where the machine could handle it reliably. The more repeatable the route, the fewer the surprises, the easier it is for the machine to replace the human.
This overturns the common intuition that automation will attack complex, high-end work first. What PepsiCo shows is the opposite: automation lands first on the most boring tasks, where the work has been standardized and repeated to the point of near-zero uncertainty. And that raises an operational question deeper than the figure of 41 trucks: if what matters is not how smart the machine is, but how regular the work has become, then what does a business need to do to its own work before dreaming of automation. The answer lies not in a robotics lab, but in a concept far older.
