> ## Content Index
> Fetch the complete content index at: https://www.siliconsnark.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Gatik Raised $200 Million to Put Doritos in the Driverless Lane
- URL: https://www.siliconsnark.com/gatik-raised-200-million-to-put-doritos-in-the-driverless-lane/
- Published: 2026-08-27T02:17:18.000Z
- Updated: 2026-08-27T02:17:18.000Z
- Description: Gatik raised $200 million to scale driverless box trucks for PepsiCo and other giants. The middle mile is suddenly very real—and very expensive.
- Author: CircuitSmith
- Tags: Startups, AI, Funding, Enterprise, Deals

Somewhere outside Dallas, a 26,000-pound box truck is moving Doritos toward a Walmart with nobody behind the wheel. This is either a meaningful advance in autonomous freight or the opening scene of a Pixar film written by a supply-chain consultant. Possibly both.

On August 25, [Gatik announced a $200 million Series D](https://archive.gatik.ai/news/coverage/gatik-raises-200-million-series-d-led-by-qia-and-kdt-as-demand-for-driverless-commercial-freight-accelerates/?ref=siliconsnark.com) led by Qatar Investment Authority and Koch Disruptive Technologies, with Millennium Management, ARK Invest, Intact Private Capital, and others joining. The company says it has more than $600 million in contracted revenue, has completed 85,000 fully driverless orders, and delivers on time 99% of the time.

Those are company-supplied numbers, so they should not be framed and hung over the loading dock without inspection. But they are unusually concrete numbers for an autonomy startup. Gatik is not promising a universal robot chauffeur that will understand every road, storm, construction cone, teenager on an e-bike, and municipal permitting mood. It is automating regional routes between distribution centers, stores, and commercial facilities.

In startup language, this is the middle mile. In human language, it is the vital stretch where pallets of chips leave a warehouse and begin their journey toward becoming a regrettable gas-station dinner.

## The Robot Did Not Take the Scenic Route

Gatik’s smartest decision may be the one it did not make. It did not start with robotaxis, where the vehicle must serve arbitrary passengers traveling to arbitrary destinations through the full urban obstacle course. It did not begin with coast-to-coast tractor-trailers, where highway speeds, weather, state rules, and long-distance edge cases pile up like an insurance adjuster’s recurring nightmare.

Instead, Gatik chose repeatable commercial freight. Its Isuzu-built box trucks run on highways and surface streets, originally on fixed routes under 10 miles and now on dynamic networks with multiple pickup and drop-off points stretching as far as 400 miles. Customers include Walmart, Loblaw, Kroger, Tyson Foods, and PepsiCo. According to [TechCrunch’s report on the financing](https://techcrunch.com/2026/08/25/self-driving-truck-startup-gatik-raises-200m-following-pepsico-deal/?ref=siliconsnark.com), the PepsiCo deployment alone uses 41 driverless trucks across Dallas, Phoenix, and Northwest Arkansas.

This focused wedge is much less theatrical than a robotaxi gliding past the Golden Gate Bridge. It is also easier for a buyer to value. Retailers know what a late delivery costs. They know the price of an idle truck, a missed replenishment window, and a shelf that contains no snacks during a football weekend. A driverless truck that reliably repeats a route is not an AI lifestyle. It is a spreadsheet cell with tires.

That is why Gatik feels like the practical cousin of [Humble’s cabless electric freight slab](https://www.siliconsnark.com/humble-raised-24-million-for-a-cabless-truck-freight-finally-met-its-midlife-crisis/). Humble starts by redesigning the vehicle around autonomy. Gatik starts by redesigning the route around what autonomy can already do. Both are coherent. One looks like the future arrived as industrial sculpture; the other looks like an Isuzu that learned procurement.

## $200 Million Buys a Lot of Trucks and Exactly Zero Forgiveness

The new money will fund more vehicles, more customers, more cities, engineering and operations hires, and eventually expansion beyond North America. Gatik currently has dozens of driverless trucks operating commercially and says it wants thousands in the years ahead. Its valuation was not disclosed, which leaves us unable to perform the traditional late-stage ritual of dividing aspiration by revenue and calling the answer “momentum.”

Still, the capital requirement is obvious. This is not software that can add 10,000 users because someone increased a cloud limit. Every new market requires trucks, sensors, mapping, maintenance, depot operations, safety validation, customer integration, regulatory work, and people who know why vehicle number 38 is making a noise described as “metallic but judgmental.”

We saw the same three-dimensional tax when [Gravis Robotics raised $200 million for autonomous excavators](https://www.siliconsnark.com/gravis-robotics-raised-200-million-to-make-excavators-work-like-employees/). Physical AI can generate software-like excitement, but it still fails in rain, dust, traffic, loading bays, hydraulic systems, and all the places where reality has declined to provide an API. The money is not merely buying growth. It is buying redundancy, field service, and the right to learn expensive lessons at fleet scale.

## Safety Claims Have Entered the Revenue Phase

The difficult part is no longer proving that a box truck can drive itself during a polished demonstration. It is proving that hundreds of them can operate every day without turning one weird reflection, sensor fault, road worker, or unusual trailer movement into national news.

Gatik says its safety approach has undergone independent review. In 2024, [TÜV SÜD assessed portions of the company’s safety case and functional-safety process](https://archive.gatik.ai/news/press-releases/gatiks-safety-case-and-functional-safety-approaches-underwent-independent-assessment-from-tuv-sud-as-part-of-an-industry-first-third-party-review/?ref=siliconsnark.com), finding its approach aligned with the evaluator’s framework while describing the work as a progress checkpoint and pathway toward future compliance. That distinction matters. An independent review of a safety process is meaningful; it is not divine certification that a robot truck has defeated weather.

Gatik’s constrained operating model helps. Repeatable commercial routes create familiar conditions, known facilities, predictable demand, and a manageable operational domain. This is the same lesson lurking beneath [the robotaxi industry’s long reality check](https://www.siliconsnark.com/deep-dive-the-robotaxi-industrys-2026-reality-check/): autonomy becomes commercially useful by narrowing the world until the system can handle it reliably, then expanding with adult supervision.

But scale changes the risk. Going from dozens of trucks to thousands multiplies exposure to rare events. A failure that appears once every million miles is a research footnote in a small pilot and an operations meeting when the fleet becomes infrastructure. Gatik will need transparent performance data, disciplined incident reporting, robust fallback systems, cyber security, state-by-state regulatory competence, and the humility to pause a deployment when the evidence says pause.

There is also labor, the subject autonomy companies prefer to describe as “capacity.” Gatik argues that its trucks address freight demand and constrained driver supply. That may be true, particularly on repetitive regional routes. It is also true that removing the driver is central to the economics. The honest version is not that nobody’s job changes. It is that logistics work shifts toward fleet operations, maintenance, remote assistance, loading, and exception handling while some driving roles shrink. Industrial transformation rarely arrives without an org chart quietly updating in another tab.

## PepsiCo Is the Customer Logo That Changes the Sentence

The PepsiCo agreement matters because it moves Gatik beyond pilot theater. The [multi-year deployment announced in June](https://archive.gatik.ai/news/coverage/pepsico-and-gatik-announce-multi-year-agreement-to-deploy-autonomous-freight-in-north-america/?ref=siliconsnark.com) places autonomous trucks inside one of the world’s largest food-and-beverage supply chains. That is not a proof of concept wandering around a fenced corporate campus. It is a buyer trusting robots with replenishment, route performance, and the sacred American obligation to keep Cheetos geographically available.

Contracted revenue is not recognized revenue, and a large agreement is not the same thing as profitable delivery. Hardware depreciation, maintenance, insurance, support staff, compute, and vehicle utilization will decide whether Gatik’s economics become attractive. Investors are betting that removing the driver and improving asset use will eventually overpower the cost of the autonomy stack. That is plausible. It is not automatic.

The competition will not remain politely outside the loading dock. Aurora, Kodiak, Waabi, Einride, Torc, Bot Auto, and others are pursuing autonomous trucking. Gatik’s middle-mile lead is useful, but the focused wedge must become a defensible operating system, not merely a sensible place to begin.

## Verdict: A Serious Breakout Hauling a Capital Furnace

Gatik looks like a serious breakout, not a beautiful overreach wearing lidar. The company chose a tractable slice of autonomy, built around measurable customer pain, pulled safety drivers from commercial routes, and signed the sort of customers that make a financing story feel less like venture cosplay. Eighty-five thousand driverless orders and $600 million in contracted revenue are exactly the traction figures a late-stage physical-AI company should be bringing to the loading bay.

It is also absolutely a capital furnace. Trucks are expensive. Fleet operations are expensive. Safety is expensive. Scaling across jurisdictions is expensive. The furnace is forgivable because the output is tangible: freight moving on public roads for paying customers, not a humanoid folding one towel while investors applaud the concept of laundry.

I remain cautiously impressed. Gatik has reduced the grand philosophical question “Can vehicles drive themselves?” to the more useful question “Can this box truck move Fritos between these facilities at 2:13 a.m. without causing trouble?” That is how enormous technologies often become real: not through universal intelligence, but through a boring route, repeated thousands of times, until the future has a delivery schedule.

The driverless revolution may not arrive in a sleek robotaxi. It may arrive behind a convenience store, reverse carefully into a loading bay, and unload enough Doritos to justify a sovereign wealth fund’s term sheet. Honestly, that feels right.