Gatik raises $200M: the middle mile automates before the last one

Gatik raised $200 million on August 25, two months after signing a multiyear deal with PepsiCo. The number that matters is not the round — it's the $600 million in contracted revenue behind it, earned on the least glamorous leg of the supply chain. Middle-mile automation is showing up in commercial retail freight before the last mile, and the reason should reshape how you pick your own automation targets.

What actually happened

Per TechCrunch, the round was led by Qatar Investment Authority and Koch Disruptive Technologies, with Millennium Management, ARK Invest, and Intact Private Capital participating. It's Gatik's largest raise; the Santa Clara company has taken in roughly $500 million since leaving stealth in 2019.

The PepsiCo work is the flagship: 41 driverless box trucks moving Frito-Lay product from distribution centers to stores across Dallas, Phoenix, and Northwest Arkansas. Not highways full of 18-wheelers — short, fixed, repeated runs between two known points. Gatik also names Walmart, Kroger, Tyson Foods, and Loblaws as customers, and says it has completed 85,000 fully driverless orders. Safety drivers came off commercial routes within the past year. The company reports $600 million in contracted revenue.

Why the middle mile matters for your business

Middle-mile routes automated first because they are boring in exactly the right way. Same origin, same destination, same time window, run again tomorrow. High volume, low variance, a clear success condition, and a human process that was already documented because a dispatcher had to schedule it. That is the profile of work machines finish, and it has nothing to do with trucks.

Your version of the middle mile is the repeated internal handoff nobody brags about. The nightly export from your POS into your accounting system. The purchase order that gets re-keyed from a supplier email. The delivery confirmation that turns into an invoice. Same shape every time, high volume, obvious when it's wrong. Meanwhile most teams point their first automation project at the customer-facing thing — the chat widget, the personalized recommendation — where variance is highest and failure is most visible.

Flip the order. Automate the boring repeated interior route first, where you can measure error rate against a known-correct answer and a bad run costs you a re-run instead of a customer.

The other lesson is in Gatik's headline metric. They led with contracted revenue, not miles driven or a demo video. When you evaluate an automation vendor, ask the same question: what is under contract and running in production, not what's in the pilot. And note that Phoenix is one of the three live markets — this is running in our backyard, not in a lab.

Sources: TechCrunch, Gatik.