As a freight carrier, how close is driverless trucking to actually cutting our line-haul costs?

Yes—but not yet in the way most carriers mean by “cut line-haul cost.” As of August 2026, driverless Class 8 trucking has crossed from pilot project into actual commercial freight, but the big cost advantage is still mostly an at-scale future-state economics story, not something you can broadly bank on today.

The short version

For a conventional long-haul carrier, I'd think about the opportunity roughly like this:

A 2026 economic study estimates 25–36% marginal cost-per-mile savings for autonomous trucks versus conventional operations, but importantly finds that the result depends heavily on route length, volume, drayage and technical-support costs. At 1,000-mile hub-to-hub operations, it modeled about 31% savings when scale economies are achieved.

The key is that you're not just eliminating the driver

The obvious calculation is:

Driver cost disappears → line-haul gets cheaper. But that's actually the smaller half of the opportunity.

The much bigger economic change is utilization.

A human driver might constrain a tractor to roughly 80,000–100,000 miles/year. An autonomous tractor can theoretically run much closer to continuously. PlusAI, for example, is modeling 240,000 annual miles for an autonomous truck versus 80,000 for a human-driven one—though those are company projections, not industry-wide observed results.

That means the same tractor can potentially produce substantially more revenue before you buy another tractor.

And that matters enormously for a carrier because tractor depreciation, insurance, maintenance and overhead get spread across more loaded miles.

But here's the catch for a carrier

The first generation of driverless trucking isn't going to mean:

“Take our existing tractor, remove the driver, and save $0.70/mile.” You'll have new costs:

• autonomy hardware/software
• autonomous-driving service fees
• remote assistance/operations
• specialized maintenance
• additional insurance/risk costs
• autonomous-system support
• transfer hubs for lanes where the truck can't operate autonomously
• potentially higher capital cost for the tractor

PlusAI's own 2026 materials, for example, show a modeled $0.97/mile human-driver cost versus $0.85/mile for its Driver-as-a-Service model—only about a 12% reduction in that particular component. The bigger benefit comes from dramatically higher utilization.

So I would not build a carrier business case around “driver cost = zero.”

I'd build it around:

driver labor eliminated + dramatically higher tractor utilization + lower cost per productive mile.

And it's becoming real surprisingly quickly

This isn't 2035 science fiction anymore.

Aurora Innovation says it launched its second-generation commercial driverless fleet in 2026 and expects to exit the year with 200 driverless trucks. It has already been running commercial driverless freight and expanding its Sun Belt operating network.

Kodiak AI reported more than 40,000 paid driverless hours in Q2 2026 across 35 trucks, with more than 20,000 revenue-generating loads, although its current economics are still those of a very early-stage operation.

So I'd characterize the market today as:

Technically: real Commercially: real on selected lanes Economically: promising but not yet proven at fleet scale Industry-wide replacement: still several years away

Where I'd expect the economics to work first

For a freight carrier, I'd prioritize:

1. Long, high-volume interstate lanes
2. Predictable origin/destination pairs
3. Minimal urban driving
4. Drop-and-hook operations
5. High tractor utilization
6. Lanes where driver turnover/recruiting is expensive
7. Routes where a human driver's hours-of-service restrictions materially constrain utilization

A 1,000-mile autonomous lane is economically much more interesting than a 200-mile lane because the autonomy system and transfer infrastructure can be amortized over many more productive miles. The recent economic research specifically finds this distance effect.

That's why Texas/Sun Belt corridors are becoming the proving ground. Aurora, for example, has been expanding lanes such as Fort Worth–El Paso and El Paso–Phoenix, including a 1,000+ mile Fort Worth–Phoenix network.

What this could mean for your P&L

Suppose your current all-in line-haul cost is $2.20/mile.

A mature autonomous operation achieving 25% lower cost would get you to approximately:

$2.20 × 0.75 = $1.65/mile

That's $0.55/mile of savings.

At 100 million annual line-haul miles, that's $55 million/year of potential cost reduction.

But I would not put that $55M into a 2026 budget.

For planning purposes, I'd use something more conservative:

• 2026–27: pilot economics; don't assume meaningful fleetwide savings
• 2028–30: perhaps 5–15% savings on specifically selected autonomous lanes
• Early/mid-2030s: 15–30% becomes a reasonable strategic planning range for mature, high-utilization lanes
• Best lanes at scale: potentially >30%

That's an estimate rather than a forecast—the exact timing depends heavily on regulatory deployment, autonomy-system pricing, insurance, utilization and how quickly OEMs scale production.

The strategic implication for a carrier is pretty big: I'd be less worried about “Will autonomous trucks replace our drivers?” and more worried about “Which of our lanes become structurally cheaper for an autonomous competitor first?”

If you give me your average loaded miles/year, driver pay, fuel cost, tractor payment/depreciation, maintenance, insurance, and average revenue/mile, I can build you a carrier-specific driverless-vs-human cost model and show the break-even autonomy price per mile and the fleet size at which it starts making financial sense.

That's $0.55/mile of savings.

At 100 million annual line-haul miles, that's $55 million/year of potential cost reduction.

But I would not put that $55M into a 2026 budget.

For planning purposes, I'd use something more conservative:

• 2026–27: pilot economics; don't assume meaningful fleetwide savings
• 2028–30: perhaps 5–15% savings on specifically selected autonomous lanes
• Early/mid-2030s: 15–30% becomes a reasonable strategic planning range for mature, high-utilization lanes
• Best lanes at scale: potentially >30%

That's an estimate rather than a forecast—the exact timing depends heavily on regulatory deployment, autonomy-system pricing, insurance, utilization and how quickly OEMs scale production.

The strategic implication for a carrier is pretty big: I'd be less worried about “Will autonomous trucks replace our drivers?” and more worried about “Which of our lanes become structurally cheaper for an autonomous competitor first?”

If you give me your average loaded miles/year, driver pay, fuel cost, tractor payment/depreciation, maintenance, insurance, and average revenue/mile, I can build you a carrier-specific driverless-vs-human cost model and show the break-even autonomy price per mile and the fleet size at which it starts making financial sense.

For a freight carrier looking at line-haul optimization, driverless trucking has officially moved out of the speculative R&D phase and into early commercial reality, but the impact on your balance sheet depends heavily on how you deploy it.

Commercial driverless operations are scaling across high-density freight corridors (primarily the Sun Belt, such as the 1,000-mile Fort Worth-to-Phoenix lane). Here is how close autonomous tech is to actually slashing your line-haul costs, broken down by the operational realities:

• Asset Utilization (𝟐× multiplier): This is the immediate win. Human drivers are legally capped by Hours-of-Service (HOS) rules, meaning traditional long-haul trucks sit idle up to 6 0% to 7 0% of the time. Driverless trucks operate nearly 20 to 24 hours a day, slashing transit times on long corridors (e.g., a 2-day human run drops to roughly 15 hours). You get double the productivity out of a single physical asset.
• Direct Labor Elimination: Eliminating the driver cuts the single largest variable cost in line-haul (traditionally accounting for 3 0% to 4 0% of per-mile operating costs). However, these savings are currently offset by high upfront hardware costs, redundant safety systems, centralized remote-monitoring supervision, and steep maintenance overhead.
• The Hub-to-Hub Bottleneck: True driverless operation is currently optimized for highway-to-highway line-haul. First-mile and last-mile drayage still require human intervention or terminal transfer hubs. Until hub-transfer infrastructure is ubiquitous across your network, you are running a hybrid model rather than a pure end-to-end cost cut.
• Insurance and Safety Tipping Point: While regulatory frameworks like the NHTSA Automated Vehicle Framework are clearing paths for commercial deployment, insurance markets are still pricing risk conservatively for autonomous fleets. Net realization of savings (projected industry-wide at scale to save billions) is gated by how fast liability costs adjust as safety data eclipses human error metrics.

Current deployment updates and scaling strategies can be tracked via industry analysis on platforms like Omdia.

To see how this fits your specific operation, tell me:

• What are your primary high-density lanes (e.g., Sun Belt vs. weather-heavy northern corridors)?
• Are you looking to buy/lease autonomous trucks directly, or partner with AV-as-a-service providers?

As of mid-2026, driverless trucking has officially moved out of the R&D and "safety driver" pilot phase and into early commercial execution on specific southern Sun Belt corridors—but realizing a bottom-line cut to your everyday line-haul costs is still a near-horizon reality (12 to 36 months away) for most general carriers, rather than an instant plug-and-play fix.

The technology has crossed major thresholds, but economic scaling depends on specific operational realities.

Where the Tech Stands Right Now (Commercial Reality)

• Active Driverless Lanes: Companies like Aurora Innovation are running commercial driverless freight on core southern lanes (like Dallas–Houston and Fort Worth–Phoenix), having surpassed hundreds of thousands of truly driverless miles.
• Hub-to-Hub Constraints: Fully autonomous operations (without safety drivers or onboard observers) are heavily concentrated on predictable, high-volume interstate corridors (primarily Texas and the Southwest).
• Hardware Cost Reductions: Next-generation autonomous hardware kits are seeing steep cost reductions (with suppliers and tech firms targeting mass-production readiness heading into 2027), but the initial capital expenditure (CapEx) for retrofitting or buying autonomous-ready trucks remains high compared to standard diesel tractors.

The Line-Haul Cost Equation: Promise vs. Reality

What This Means for Your Fleet Today

If you are running dedicated, repetitive hub-to-hub freight down predictable southern corridors (e.g., DFW to Houston/Phoenix), you can partner with autonomous-as-a-service providers right now to shave line-haul variable costs on those specific lanes.

However, if you operate a general-commodity or spot-market network across unpredictable weather zones, complex urban interchanges, or non-Sun Belt states, mass-market cost relief is still a couple of years away. Full commercial scaling relies heavily on broader regulatory harmonization (anticipated around the 2027 legislative cycles) and the roll-out of mass-production vehicle hardware by major truck OEMs (like International and Isuzu paired with firms like Aurora and Gatik).

If you'd like to dive deeper, let me know:

• What specific lanes or corridors your fleet runs most frequently?
• Are you looking at this from a leasing/partner perspective or general long-term fleet planning?

I can share more details on how hub-to-hub transfer models might integrate with your current terminal network.

If you're a freight carrier looking at this as a line-haul cost question rather than a technology question, we're getting fairly close—but the headline savings are ahead of what most carriers can capture today.

The latest data suggests driverless Class 8 trucking is moving from pilot economics toward real commercial economics, especially on long, repetitive Sun Belt lanes.

The short answer

I'd model it this way:

That's line-haul operating cost, not necessarily the freight rate you'll be able to charge customers. The distinction is crucial.

ATRI puts the average U.S. trucking operating cost at $2.336/mile for 2025, with driver compensation alone at about $1.03/mile.

So, in principle, eliminating the driver is enormous. But you're not simply taking $1.03 off the bill.

What replaces the driver?

You'll have some combination of:

• autonomy-system fees
• expensive sensors/compute and specialized equipment
• remote assistance/supervision
• additional maintenance
• higher insurance initially
• autonomous-fleet operations staff
• transfer/drayage operations at autonomous hubs
• potentially higher financing/depreciation costs

That's why the near-term carrier saving is much smaller than the gross driver-cost elimination.

For example, Aurora is currently targeting a driver-as-a-service price of $0.85+/mile. That compares with roughly $1.03/mile of conventional driver compensation—but the carrier still pays fuel, truck, maintenance, insurance, etc.

Where the really big saving comes from

The bigger economic lever isn't actually labor. It's truck utilization.

A human-driven truck is constrained by hours-of-service. A driverless truck can potentially run close to continuously.

Reuters recently described the fundamental advantage as a truck potentially operating around the clock rather than stopping for driver rest, effectively allowing much greater utilization of the capital asset.

That matters enormously for a carrier.

Imagine your current tractor runs:

10 hours/day × 250 days = 2,500 productive hours/year

An autonomous tractor doesn't necessarily need to run twice as many hours—but if you can get materially closer to 20 hours/day, you can move the same freight with fewer tractors and trailers.

That can reduce:

• tractor capital per unit of freight
• depreciation per mile
• fixed insurance
• parking/facility requirements
• equipment sitting idle
• driver recruiting/retention expense
• schedule variability

And it increases your revenue-producing miles per tractor.

The interesting real-world number

Aurora has already published a comparison on a roughly 1,000-mile Fort Worth–Phoenix lane. Its analysis estimated human-driven operating cost at $1.99/mile versus $1.84/mile for its autonomous truck—only about a 7.5% cost reduction in that particular comparison. But the autonomous truck was able to make substantially more trips because it wasn't constrained by driver hours.

That's an important warning against the simplistic:

"Remove the driver = save $1/mile." You probably won't see that $1/mile drop in your P&L.

Instead, you'll potentially get a combination of:

lower cost/mile + dramatically higher miles/tractor + more consistent utilization.

Where I'd expect the economics to break

A 2026 economic study found potential 25–36% marginal cost-per-mile savings for autonomous trucks versus conventional trucks, but the result was highly dependent on route length, volume, drayage distance and technical-support costs. In hub-to-hub operations, its modeled savings ranged from only 13% on 200-mile routes to 31% on 1,000-mile routes when scale economies were achieved.

That is probably the most useful way to think about this.

Autonomy is fundamentally a long-haul network technology, not a universal trucking replacement.

A 150-mile route with complicated city driving, low utilization and expensive transfer operations may be a lousy autonomous lane.

A 700–1,200 mile, high-volume, mostly interstate lane between two autonomous hubs is much more compelling.

My carrier-side forecast

If I were building your fleet plan today, I wouldn't put a 30% cost reduction into the budget.

I'd underwrite something like:

2026–27: 5–10% demonstrable savings on suitable lanes 2028–30: 10–20% as equipment and operations scale 2030+: 20–30%+ on highly optimized autonomous corridors

There is meaningful evidence that the industry is approaching the second phase. Kodiak, for example, reported more than 40,000 paid driverless hours and 20,000 revenue-generating loads in Q2 2026, while planning expansion into long-haul highway operations.

And Goldman Sachs' forecast cited this year puts autonomous trucks becoming cheaper per mile than human-driven trucks around 2028.

So I'd characterize the technology as commercially real but not yet economically mature.

The strategic implication for a carrier

The question I'd be asking isn't:

"When can I buy a driverless truck?" It's:

"Which 10–20% of my network could be converted to autonomous hub-to-hub operation first?" Those lanes could produce disproportionately large savings because you're getting all three benefits simultaneously:

driver cost ↓ + utilization ↑ + equipment required per load ↓

If you give me your average loaded miles, annual miles/tractor, driver cost per mile, fuel cost, tractor payment/depreciation, and typical line-haul lane lengths, I can run a carrier-style human vs. driverless cost-per-mile and cost-per-load model and show you the break-even autonomy fee you'd need from an Aurora/Kodiak-type provider.

Here are top web results for exploring this topic:

Kodiak AI·https://kodiak.ai Total Cost of Ownership for Autonomous Trucks - Kodiak AI The business case for self-driving trucks is most often discussed in broad strokes. The technology is advancing. The industry is changing. The opportunity is significant. What gets far less attention SEKO Logistics·https://www.sekologistics.com The Impact of Autonomous Vehicles on Freight Transportation Autonomous Vehicles Cut Operational Costs. Speaking of expenses, autonomous trucking cuts costs through: Less fuel usage. Lower driver compensation. Fewer accidents. According to McKinsey, autonomous Reddit·https://www.reddit.com How much money will actually be saved by self driving trucks? - Reddit To me, that doesn't seem like a lot in the grand scheme of things due to all the other costs associated with trucking. Truck. The average cost of semi is around 75k + 5-10k in self driving tech (Pulle MarketWise·https://marketwise.com How Autonomous Trucking Could Drive a $600 Billion Freight ...Research from McKinsey shows that autonomous trucking will reduce costs per mile by 42%. Those savings come from reduced driver salaries, less fuel consumption, and fewer projected accidents. (Fewer a Earth.com·https://www.earth.com**Self-driving** trucks could upend America's freight economy - Earth.com Self-driving trucks could upend America's freight economy. Autonomous trucks could slash freight costs by 35 percent while reshaping trade, traffic, jobs, and economies across the United States. Eric LinkedIn·https://www.linkedin.com**Autonomous Freight** : The End of Brokers, Carriers , and Class 8 ...I continue to research the autonomous trucking industry, and the more that I do, the more I'm convinced that most people don't understand that autonomous trucks won't just make drivers obsolete and un Facebook·https://www.facebook.com**Autonomous trucking** to cut costs by $9 billion annually - Facebook Who going to buy products and services?? You people keep posting this automation stuff are really stupid. Not going to save anything in the long run. Been driving 39 years. Those trucks are not better HERE Technologies·https://www.here.com Debate worth having: will truckers be replaced by self-driving trucks?... shipping container containers supply chain fleet management 0. Read more: The logistics landscape in 2024 — resilience is crucial. “These benefits are crucial for strengthening our brittle supply NACFE·https://nacfe.org The Financial Implications of Autonomous Trucking - NACFE Initial Investment and Technology Costs. The adoption of AVs in trucking necessitates a substantial initial investment. · Vehicle Acquisition Costs. Autonomous trucks are more expensive than tradition Delaware Valley Contractors·https://www.delawarevalleycontractors.com How Are Driverless Trucks Affecting the Trucking Industry?How Are Driverless Trucks Affecting the Trucking Industry? DVC • May 20, 2025. The trucking industry is on the verge of a technological revolution, and at the forefront of this transformation are driv

If you mean true Level 4/driver-out trucking on interstate line-haul, rather than today's driver-assist systems, the economics are getting close—but I would not underwrite a 30–40% line-haul cost reduction today.

The short version

For a carrier, I'd think about the opportunity in three stages:

Those are my estimates rather than guarantees. A 2026 economic study found potential marginal cost-per-mile savings of 25–36%, but importantly, its hub-to-hub modeling found only 13% savings on 200-mile routes versus 31% on 1,000-mile routes when scale economies are achieved. Without sufficient scale, shorter routes can actually become more expensive.

The reason is that removing the driver isn't the whole economic story.

Where the savings actually come from

Today's U.S. truck operating cost is about $2.34/mile according to ATRI's 2026 report.

A driverless truck potentially eliminates/reduces:

• Driver wages, benefits, recruiting and turnover
• HOS-induced downtime
• Some driver-related deadhead
• The need to park/stop the truck for overnight rest
• Some utilization constraints

But you add:

• Autonomous-driving hardware/software
• Remote assistance/supervision
• Higher technology depreciation
• Additional maintenance complexity
• Insurance/risk costs
• Operational hubs and transfers
• Potentially expensive autonomy-service fees

And the biggest prize may actually be utilization, not labor.

A human-driven tractor might realistically run ~80,000 miles/year. PlusAI, for example, models an autonomous truck at 240,000 miles/year under its assumptions. That's a huge increase in revenue-producing capacity per tractor.

Reuters recently described the same basic economic mechanism: a truck that can operate around 20 hours/day instead of 10 roughly doubles the productivity of the capital invested in it.

The number I'd put in a carrier model

If your current line-haul cost is, say, $2.20/mile, I wouldn't initially budget:

$2.20 → $1.40 I'd model something more like:

$2.20 → $1.90–$2.05 initially → $1.70–$1.85 at scale → potentially $1.45–$1.65 in a mature network The latter range is consistent with some industry projections. Volvo's 2026 modeling, for example, puts a mature autonomous truck around $1.15–$1.35/mile versus roughly $2.25–$2.30 today, although it explicitly notes that these are at-scale economics and that today's hardware/operating costs are substantially higher.

I would treat that $1.15–$1.35 figure as an upside case, not a carrier budget assumption.

The really interesting part for a carrier

Suppose you're running a 1,000-mile lane.

A conventional solo driver can make the trip, but the truck then sits during mandatory rest periods. A driverless truck can potentially run essentially continuously.

That's why the economics become much more compelling on long, predictable, high-volume lanes.

An industry case study presented in 2025 modeled a 1,000-mile Phoenix–Fort Worth lane at:

• Human: $1.99/mile
• Autonomous: $1.84/mile
• But trips/week: 3 → 8

So the headline cost/mile improvement was only ~8%, while the modeled revenue/profit improvement was dramatically larger because the truck could do substantially more work.

That's the distinction I'd emphasize to a CFO: autonomy isn't merely a labor-substitution project. It's potentially an asset-utilization project.

Where I'd start as a carrier

The first lanes I'd automate aren't necessarily the ones with the highest driver cost. I'd target lanes with:

1. 500–1,000+ highway miles
2. Very predictable origin/destination
3. High weekly volume
4. Minimal urban driving
5. Easy terminal/hub access
6. Little special handling
7. Consistent freight
8. High tractor utilization
9. Good autonomous operating-domain coverage
10. Enough density to justify dedicated autonomous hubs

That's why the industry is initially concentrating on places such as Texas and other favorable Sunbelt corridors rather than trying to make every truck everywhere driverless immediately. Commercial driverless freight is already operating in Texas, while deployments elsewhere are still being staged.

My bottom line

Driverless trucking is probably 2–4 years away from being genuinely compelling on the right long-haul lanes, but we're not yet at "replace the driver and immediately save 40%" economics.

For a carrier planning capital today, I'd use roughly 10% near-term cost reduction and 20–25% mature-network reduction as conservative planning cases, with 30%+ as an upside scenario.

And I'd separately model the revenue/asset-utilization gain. That could ultimately matter more than the direct line-haul cost reduction.

If you give me your average line-haul length, current cost/mile, annual tractor miles, driver compensation, and percentage of miles on repeat lanes, I can build a carrier-level driverless-vs-human unit economics model and show the break-even mileage and likely savings per tractor.

Suppose you're running a 1,000-mile lane.

A conventional solo driver can make the trip, but the truck then sits during mandatory rest periods. A driverless truck can potentially run essentially continuously.

That's why the economics become much more compelling on long, predictable, high-volume lanes.

An industry case study presented in 2025 modeled a 1,000-mile Phoenix–Fort Worth lane at:

Driverless trucking is no longer a futuristic science experiment confined to closed test tracks—it is actively operating on public commercial corridors , meaning the economic reckoning for line-haul operations has arrived. However, cutting actual line-haul costs at scale depends heavily on the specific lanes you run and your willingness to adopt a hub-to-hub transfer model.

Here is how close the industry is to impacting your bottom line, broken down by cost drivers, operational realities, and timelines.

1. The Financial Reality: Where the Savings Hit

• The Labor Replacement Equation: Driver compensation typically accounts for 40–45% of traditional long-haul operating costs . Fully driverless operations eliminate per-mile driver wages and per diem costs on the highway portion of a route. Early commercial metrics (such as pilot runs by autonomous developers on corridors like Houston-to-Dallas) put the all-in cost for humanless runs at under$2.0 0 per mile, undercutting human-operated trips averaging closer to$2.5 5+ per mile.
• Asset Utilization (The Real Game-Changer): Federal hours-of-service rules cap human drivers at 11 hours of driving within a 14-hour window. Autonomous trucks running on permissive highway infrastructure face no such rest breaks, operating up to 20–22 hours a day. A 1,000-mile stretch (e.g., Fort Worth to Phoenix) that takes a human team or multi-day relay nearly 3 days can be completed autonomously in roughly 15 to 20 hours, more than doubling equipment utilization.
• Fuel and Maintenance Efficiency: AI-driven speed control, braking, and terrain management cut fuel waste significantly, with projected long-term efficiencies reducing fuel consumption and waste by up to 32%.

2. The Bottleneck: The "Transfer Hub" and Edge Cases

You cannot yet load freight at a dock in downtown Chicago and have an empty truck navigate downtown Miami. Driverless tech relies on a hub-to-hub transfer model:

• The Middle Mile is Solved: Companies like Aurora Innovation and Kodiak Robotics have scaled driverless operations across structured Sun Belt highway lanes (Texas, Arizona, New Mexico). Aurora's commercial scaling phase features fleets running without safety drivers or in-cab observers on major highways.
• The Endpoints Still Need Humans: You still need human drivers or terminal staff at transfer facilities on the outskirts of cities to manage dock-level backing, local street navigation, and complex yard environments. This means you aren’t entirely erasing labor from your P&L; you are shifting it to a lower-cost, localized hub model.
• Weather and Environment Limits: Current Level 4 systems reliably handle roughly 85% of standard weather conditions. Heavy downpours, dense fog, snow, or ice still trigger remote intervention or temporary suspensions, creating occasional schedule unpredictability.

3. Timeline to Your Fleet

• Right Now (2026): If your freight runs high-density, predictable southern/Sun Belt interstate corridors (like I-10 or I-20), commercial capacity is already live. Industrial applications (such as closed-loop or regional aggregate hauling) have already proven 24/7 profitability. Mass production of second-generation, cost-reduced hardware kits (slashing sensor suite costs by over 50%) is hitting the market, and recent regulatory openings—such as updated heavy-duty autonomous testing frameworks on California highways—signal that geographic expansion is accelerating.
• 2027–2028 (The Tipping Point): Major truck OEMs (like Volvo Autonomous Solutions) are scheduled to roll out factory-produced autonomous platforms at an industrial scale in partnership with software drivers, targeting hundreds to thousands of active units. By 2028, industry consensus suggests autonomous per-mile operational costs will consistently beat human-driven models across a much broader cross-section of the interstate network.

If you'd like to narrow this down for your business, tell me:

• What are your primary freight corridors or lanes?
• Are you primarily operating dry van, reefer, or specialized freight?

I can give you a better idea of how a hub-to-hub partnership or software integration fits your current network.

Yes—but the economics are getting close faster than the headline “remove the driver” suggests. For a freight carrier, I’d think about it in three stages:

The key number today: about $1/mile of driver cost

ATRI's latest 2026 report puts average trucking operating cost at $2.336/mile in 2025. truckingresearch.org Driver wages and benefits are roughly $1/mile, so eliminating the human driver doesn't mean eliminating half your cost—it creates a theoretical ceiling of roughly 40–45% before considering the cost of the autonomous system.

And that's exactly what the early commercial pricing is showing.

Aurora is currently targeting roughly $0.85/mile for its Driver-as-a-Service, versus roughly $1/mile for driver wages and benefits. So its direct labor saving is only about $0.15/mile, or roughly 7% of today's $2.34/mile total cost.

That's why I wouldn't build a carrier business case around “we save $1/mile.”

The bigger prize is utilization

This is where driverless becomes genuinely disruptive.

A human driver is constrained by hours-of-service rules and typically doesn't generate anywhere near 24 hours of productive movement. Autonomous trucks can potentially run essentially continuously, subject to charging, fueling, maintenance, loading and operational constraints. Reuters notes that this can potentially roughly double truck productivity.

Aurora's own carrier example illustrates the point: on a Phoenix–Fort Worth 1,000-mile lane, its model assumes 3 trips/week today versus 8 with autonomy, with cost per mile falling from $1.992 to $1.843. The resulting modeled weekly profit rises from $185 to $1,695 per truck. That's an 816% increase in weekly profit, despite only about a 7.5% reduction in modeled CPM.

That's the crucial distinction:

Autonomy's first economic win may be dramatically higher revenue/profit per tractor, rather than dramatically lower cost per loaded mile.

But route design matters enormously

For a carrier, I would be particularly bullish on:

• 500–1,500+ mile interstate lanes
• High-volume, repetitive origin/destination pairs
• Freight moving between your own/customer facilities
• Lanes where you currently have driver availability problems
• Operations where you can keep the autonomous truck moving overnight
• Temperature-controlled or time-sensitive freight where extra utilization has real value

I would be much less excited about short-haul freight. McKinsey's 2024 analysis found that hub-to-hub autonomy becomes substantially more attractive as distance increases, with 1,500+ mile routes potentially seeing ~42% lower TCO per mile once the technology and operating model mature. Short routes get killed by the fixed costs of hubs, transfers and manual first/last-mile work.

Where I'd put the realistic carrier expectation

If you're budgeting 2026–2028, I'd model something like:

$2.34 current cost/mile

→ ~$2.15–$2.25 early autonomous operation
→ ~$1.75–$2.00 at meaningful scale
→ ~$1.40–$1.70 once the operating model is mature

Those aren't guaranteed industry numbers—they're a planning range based on the current economics and published projections, not a promised autonomous-truck price.

The really interesting part is that the truck doesn't necessarily have to become dramatically cheaper per mile to be transformational. If a tractor goes from, say, 100k miles/year toward 200k+ miles/year, you're spreading the tractor, autonomy hardware and fixed overhead across vastly more productive miles. Plus, you're replacing the labor constraint.

One thing I'd watch very closely

Don't evaluate an autonomous vendor on its advertised $/autonomous mile.

Evaluate it on:

fully loaded cost per delivered load = truck + autonomy fee + fuel + maintenance + insurance + remote operations + terminals/hubs + repositioning + manual drayage + downtime

That last group can eat the apparent labor savings.

The industry is actually moving into that test now. Aurora says it was already fully allocated for 200 driverless trucks operating by year-end 2026, with carriers such as Value Truck and Charger Logistics putting the technology onto Dallas–Laredo and Fort Worth–Phoenix lanes.

My bottom line: if you're a line-haul carrier, I'd say driverless trucking is commercially real now, but the big cost advantage is probably 3–7 years away rather than already sitting in today's quoted rates. The near-term opportunity is less “cut our CPM by 40%” and more “get substantially more revenue-producing miles out of each tractor while taking labor out of the cost stack.”

If you give me your current loaded CPM, driver CPM, average annual miles/truck, average length of haul, and typical line-haul rate, I can build a carrier-specific driverless-vs-human P&L and show you the break-even mileage and likely margin improvement.

The short answer is: We are at the inflection point, but the impact is corridor-specific and gated by scale.

Fully driverless, commercial line-haul operations are no longer science fiction or perpetual R&D pilots. Companies like Aurora Innovation, Kodiak, and Bot Auto are actively running revenue-generating, driver-out freight across high-priority Sun Belt corridors (such as Dallas–Houston and Fort Worth–Phoenix).

However, realizing those cost cuts across your entire network depends on navigating a hybrid reality over the next 18 to 36 months. Here is how close the economics are to actually hitting your bottom line:

• The Core Savings Potential: Industry analyses (including recent 2026 logistics economic models) project that mature Level 4 autonomous trucking will ultimately slash per-mile operating costs by 30% to 45% on applicable highway lanes.
• The 20-to-22 Hour Duty Cycle: The single biggest near-term cost cutter isn't just zero wages for that leg—it’s asset utilization . Bound by Hours of Service (HOS) rules, a human-driven rig sits idle or moves slowly relative to a 24/7 machine. A 1,000-mile lane that takes a human team 2+ days can be compressed dramatically, effectively doubling or tripling truck productivity.
• The Fuel and Safety Dividend: Autonomous acceleration and speed-profiling algorithms yield measurable fuel efficiency gains (projected up to 5.7÷ savings industry-wide over the next decade), and cleaner safety records point toward structurally lower insurance premiums.

The Bottlenecks Delaying Instant Network-Wide Savings

• The Transfer Hub Model Tax: You cannot simply send a driverless semi from a shipper's messy dock in Chicago to a tight receiver in downtown Miami. Autonomy is restricted to hub-to-hub interstate highways. You still incur costs for local human drivers, terminal real estate, and trailer handoffs at origin and destination points.
• Weather and Geography Constraints: Current driverless software safely manages about 8 5% of weather conditions, meaning rain, heavy fog, snow, or construction-heavy detours still require temporary suspension or remote intervention.
• The CapEx vs. "Trucking-as-a-Service" (TaaS) Equation: Upfront hardware, LiDAR suites, and software subscription models ($0.12 to $0.25+ per autonomous mile) eat into early gross margins. True break-even for mid-to-large fleets on optimized lanes sits in the 2.5 to 3-year range before pure margin expansion kicks in.

If you are evaluating your capital allocation or network planning right now, the technology is ready today if your freight is concentrated on long, straight, predictable southern interstate corridors. Scaling it to generalized, weather-proof, nationwide network-wide savings is a 2027–2028 reality as OEM factory lines ramp up production.

To narrow this down for your specific fleet, what are your primary high-density lanes or average length-of-haul distances , and are you looking at capital investment in autonomous tech or outsourcing via TaaS partnerships?

Pretty close—but not yet to the point where you should assume a dramatic line-haul cost reduction across your whole network. As of August 2026, driverless Class 8 trucking has crossed from pilot projects into actual commercial freight, but the economics are still lane- and utilization-dependent.

What the savings could look like

For a carrier, I'd think about it in three stages:

The long-term numbers aren't fantasy. McKinsey has estimated that full autonomy could eventually reduce trucking total cost of ownership by roughly 45%, with labor being the largest source of savings.

But your actual savings won't equal 45% of today's line-haul rate. That's the important distinction.

Where the money actually comes from

The big prize is driver productivity, not fuel.

A driverless truck can potentially run much closer to continuously. Aurora says its driverless trucks are already averaging 4,000+ miles/week, and it expects autonomous trucks to more than double utilization/revenue per truck.

That creates several levers:

• Driver wages/benefits disappear from the autonomous segment
• More miles per tractor per year
• Less equipment sitting idle because of HOS constraints
• Potentially fewer tractors required for the same freight volume
• Better consistency and potentially lower accident/insurance costs
• Some fuel savings from smoother driving

Aurora's current commercial model is particularly interesting because it has reported an approximately $0.85/mile operating-cost target, which gives carriers a real benchmark rather than a purely theoretical one.

But there's a catch: autonomy isn't free

You're replacing a driver with:

• autonomous-driving hardware
• software/service fees
• additional maintenance
• remote assistance/operations
• higher initial truck/equipment costs
• autonomous-system depreciation
• transfer hubs or specialized operating procedures
• potentially higher insurance initially

And that's why lane selection is everything.

McKinsey's more recent analysis finds the economics become substantially more attractive as route length increases. Short routes can actually be uneconomic because of the overhead associated with transferring between autonomous and human-driven segments; routes over roughly 1,500 miles are where the TCO advantage becomes much more compelling.

The really interesting part for a carrier

I'd stop thinking of this as:

"Driverless truck = eliminate driver cost."

and instead model it as:

"Driverless truck = turn a tractor from a ~10-hour/day asset into a potentially 20+ hour/day asset."

That's potentially much more valuable.

For example, suppose you currently have a tractor doing 110,000–130,000 miles/year. If autonomy ultimately lets the same equipment generate something approaching 200,000–250,000 miles/year, your tractor, terminal, and depreciation costs get spread over dramatically more revenue miles.

That's why autonomous trucking could eventually change the economics of the entire network—not merely reduce payroll.

Aurora is already pursuing this exact model: it launched a second-generation driverless fleet in 2026, expects 200 driverless trucks operating by year-end, and says its new autonomy hardware should cost more than 50% less than the previous generation.

My take for a freight carrier in 2026

I'd put it this way:

Driverless trucking is probably 1–3 years away from being economically compelling on a meaningful subset of predictable long-haul lanes, but probably 5–10+ years away from being a generalized replacement for conventional OTR driving.

The sweet spot is likely:

fixed lanes + highway-heavy mileage + predictable terminals + high utilization + relatively little urban driving.

If you have, say, Dallas–Laredo, Fort Worth–Phoenix, Houston–Dallas, or similar repetitive lanes, you're much closer to the economic inflection point than if your network is predominantly irregular truckload with complicated pickups and deliveries.

And there's evidence this is already moving beyond experiments: Aurora says it has commercial customers operating driverless freight and is targeting 200 trucks this year, while Volvo Autonomous Solutions plans driverless operations using the Aurora Driver beginning in Q1 2027.

If you give me your approximate current cost/mile (or line-haul rate), annual miles per tractor, driver compensation, fuel MPG, and typical lane lengths, I can build a carrier-specific model showing what driverless would do to your cost/mile at 10%, 25%, and 50% fleet penetration.