Autonomous trucks reach the Laredo border lane as hardware costs fall toward $50,000 a truck

The details

Dallas–Laredo. Kodiak AI and Charger Logistics, an asset-based U.S. and Mexico carrier, are running autonomous trucks between Charger’s Dallas and Laredo terminals, 435 miles apart. The first delivery ran Sept. 8. The trucks haul refrigerated and dry van loads for consumer packaged goods and food and beverage customers, with a safety driver aboard until Kodiak completes its safety case. Kodiak’s readiness measure reached 96% at the end of September, and it plans driverless long-haul service on its Dallas–Houston lane by year-end. Laredo handled more than $350 billion in trade in 2025, nearly 40% of U.S. trade with Mexico, per Kodiak.

Kodiak supplies its system as a service on trucks the carrier owns and operates. Charger also has a separate deal with Aurora on the same lane, in which Aurora holds the operating authority, controls the truck and carries the insurance.

Hardware costs. Houston-based Bot Auto says its full autonomy hardware, including compute and sensors, will cost about $50,000 per truck using a Lenovo controller built on Nvidia’s platform, starting in the first quarter of 2027. Bot Auto says that is three to eight times less than other systems: it cites estimates of $350,000 to $600,000 for a complete autonomous truck, implying $150,000 to $400,000 for autonomy alone. The Lenovo AD1 controller runs on Nvidia’s DRIVE AGX Thor with redundant compute and power, and the companies plan to show it at CES in January. “We made a deliberate choice not to scale our fleet until the compute platform itself is scalable,” CEO Xiaodi Hou said. Aurora, Einride, WeRide and Nuro are also building on Nvidia or Lenovo platforms. The company runs about 350 loads a month in Texas and completed a fully driverless 231-mile commercial run from Houston to the Dallas area in April.

How Kodiak measures readiness. Kodiak’s long-haul Autonomy Readiness Measure, which tracks the share of claims and evidence completed in its safety case, rose from 84% in February to 93% at the end of August and 96% at the end of September, according to FreightWaves. Kodiak’s driverless trucks have been in commercial service in the Permian Basin since December 2024. Charger runs its network on its own transportation management system, and president Andy Khera described the combination with Kodiak as “a nearly fully automated ecosystem.” Aurora announced in late July that Charger would deploy its second-generation driverless trucks on the same lane.

Why the hardware price matters to shippers

Autonomy changes freight costs only when trucks can run without a driver at scale, and that depends on hardware cost as much as on safety approval. A lower kit price lets more carriers, not just the largest fleets, buy driverless trucks, which would spread the technology to more lanes. Bot Auto, which moves about 350 loads a month in Texas and says revenue grew 270% from May to August, is targeting 1 billion autonomous miles by 2031.

What the cost gap implies

If Bot Auto’s $50,000 kit ships as planned, it would cost $100,000 to $350,000 less per truck than the autonomy hardware implied by the estimates it cites, by our calculation. For a carrier converting a fleet of 50 trucks, that is a difference of $5 million to $17.5 million in upfront cost. Hardware isn’t the only cost: software, mapping, remote assistance and insurance add to it, and those figures aren’t public. But hardware has been the main barrier for mid-size carriers, which run much of the capacity procurement teams buy.

How to pilot autonomous capacity

1. Pick one lane between two fixed terminals, with steady daily volume and freight that isn’t time-critical.
2. Run it alongside driven capacity for a set period, with the same service measures.
3. Agree liability in writing: who holds operating authority, who insures the load and what happens on a disengagement.
4. Measure on-time delivery, claims, cost per mile and tender acceptance against conventional carriers on the same lane.
5. Decide on scale only after the provider moves from safety-driver to driverless operation on that lane.

This is our suggested approach.

Where autonomous freight is running in Texas

Questions to ask carriers offering autonomous capacity

1. Who holds the operating authority and insurance on autonomous loads?
2. What happens to a load if the system disengages or a sensor fails mid-route?
3. Are rates different from driven loads, and how will they change as driverless service begins?
4. Which commodities and trailer types qualify, including refrigerated freight?

What procurement teams should do

• Ask your Texas carriers about autonomy. For Dallas, Houston and Laredo lanes, ask whether they are testing autonomous capacity and on what terms.
• Clarify liability before tendering. Contract structures differ: some providers carry the operating authority and insurance, others supply technology to a carrier that does. Your terms should state who is responsible.
• Start with dull, steady lanes. Autonomy works best on predictable, high-frequency lanes between terminals. Those are the lanes to pilot.
• Track service, not novelty. Measure on-time performance, claims and cost per mile against conventional capacity on the same lane.

What to watch next

• Driverless operation. Kodiak plans driverless long-haul on Dallas–Houston by year-end; Laredo would follow once its safety case is complete.
• Bot Auto’s 2027 hardware. If the $50,000 kit ships in the first quarter of 2027, autonomy economics change for mid-size carriers, not just the largest fleets.
• Insurance and liability terms. Different models (Kodiak as a service, Aurora holding operating authority) put liability in different places; watch how carriers price it.

The bigger picture

Autonomy won’t fix this year’s capacity squeeze, but it is becoming a credible piece of long-term freight sourcing on specific corridors. With truckload contract rates at a 52-week high and cross-border driver supply under pressure, new capacity models deserve a place in multi-year freight strategy.