Industry NewsAUR · JBHT · WERN•13 min read

Aurora Just Put 2028 on Driver-Out Trucking — Every Long-Haul Carrier Is Now a Margin Bet

Aurora Innovation said on Sep. 22, 2026 that driverless trucks will reach positive free-cash-flow by 2028 with per-mile economics around $0.85 versus ~$1.30 for human drivers — a spread that instantly re-prices J.B. Hunt, Werner, Schneider National, Knight-Swift, PACCAR and Ryder System, while NVIDIA sits on the silicon side and PACCAR's own CEO publicly says his trucks won't drop the driver.

Published Sep 23, 2026Updated Sep 23, 2026

Aurora Driver-as-a-Service target rate

$0.85/mile

Bank of America estimate cited on CNBC, Sep. 22, 2026

Human driver wages + benefits cost

$1.30/mile

Bank of America estimate of all-in driver cost, Sep. 22, 2026

Implied per-mile profitability

~7.5× human

Morgan Stanley estimate cited on CNBC, Sep. 22, 2026

U.S. trucking gross freight revenues

$906B (2024)

American Trucking Associations industry tracker

Event

Aurora's 2028 free-cash-flow claim lands — and the dollar-per-mile math is the freight trade

On a ride-along in Texas on Sep. 22, 2026, Aurora Innovation CEO Chris Urmson said the company is targeting positive free-cash-flow by 2028 — driven by scaling driverless truck counts on its Sun Belt freight lanes, not by chasing robotaxi miles in cities. The claim sits on a simple cost wedge the publicly listed freight industry has not had to price before.

Aurora Driver-as-a-Service target rate

$0.85/mile

Bank of America estimate cited on CNBC, Sep. 22, 2026

Human driver wages + benefits cost

$1.30/mile

Bank of America estimate of all-in driver cost, Sep. 22, 2026

Implied per-mile profitability

~7.5× human

Morgan Stanley estimate cited on CNBC, Sep. 22, 2026

U.S. trucking gross freight revenues

$906B (2024)

American Trucking Associations industry tracker

Aurora Innovation slates positive free-cash-flow for 2028 — and the entire J.B. Hunt, Werner, Schneider National, Knight-Swift and PACCAR coverage list just got a 2028 clock. The drivers are fleet count, hardware cost, and a roughly 30% gap between autonomous and human per-mile economics.

Until now the autonomous trucking narrative has been either a Tesla/Waymo robotaxi story or a science-project pilot. Aurora's Sept. 22 economics push the discussion into freight P&L lines: which listed carrier eats the per-mile cost reset, which OEM trucks run which corridors, and where the capital plan sits for the next 24 months.

Cash path

From 20 driverless trucks today to a fleet big enough to clear $1.2B of cash by 2028

Aurora ended Q2 FY2026 with $2M in quarterly revenue, a $266M operating loss (including $60M of stock-based comp) and roughly $1.2B of cash and short-term investments. CFO David Maday guided to an average quarterly cash burn of $190–$220M through 2026 and full-year 2026 revenue of $14–$16M — numbers that make 2028 free-cash-flow depend on a step-change in per-truck contribution once the Driver-as-a-Service (DaaS) migration begins in 2027.

Aurora driverless fleet build plan vs. implied cash breakeven

Build trajectory implied by Aurora disclosures (CB at ~5,000–7,000 driverless trucks assuming DaaS rate of ~$0.85/mile and ~$760–880M annualized 2026 burn).

Unit: driverless trucks (units)

Q2 2026 (on-road)

Q4 2026 (target)

200

Roush capacity (Oct ’26)

1,000

Hirschbach order (2027+)

500

Implied FCF breakeven

6,000

Roush has begun upfitting and is on track for an annual run-rate of 1,000 trucks/year starting October 2026, with Aurora's stated goal of clearing 1,000 trucks of cumulative production by year-end 2026
Second-generation Aurora hardware kit, designed for a one-million-mile service life, shrinks the per-truck hardware bill by more than 50% versus first-gen hardware
Aurora's second-generation driverless trucks — built on the International LT Series (Navistar/Volkswagen Traton) — launched commercially on July 22, 2026, the third Sun Belt corridor after Dallas-Houston and Fort Worth-El Paso
The next-generation compute platform pairs NVIDIA's DRIVE Thor with Continental's mass-produced sensor kit, with mass production targeted for 2027 — that's when the per-truck economics fully unlock for DaaS fleets

At Aurora Innovation's current $190–220M quarterly burn, the 2028 FCF target only holds if the driverless fleet reaches roughly 5,000–7,000 trucks by mid-2028 — a 250× increase from today's 20-vehicle fleet, riding on Roush's 1,000-truck annual capacity and partner deliveries.

Downstream

Listed carriers most exposed to the long-haul cost reset

The per-mile economics matter most where drivers are. Aurora's $0.85/mile DaaS rate compares to a 2025 ATRI industry-average all-in operating cost of $2.336/mile, of which driver wages ($0.818) and benefits ($0.210) total $1.028/mile. Bank of America's $1.30/mile all-in human estimate adds indirect labor and insurance on top, but the directional point is the same: a 17–35% per-mile cost wedge sits inside the largest single line item on every long-haul carrier's income statement.

Q2 FY2026 carrier scale vs. long-haul exposure
Carrier (US listing)	Q2 2026 revenue	Q2 2026 op. margin (TTM)	Aurora partnership?	AV-read-through
J.B. Hunt Transport Services	$3.50B	7.4%	None disclosed	Mixed — intermodal hinges on drayage, not long-haul
Werner Enterprises	$0.93B	2.6%	Named customer	Bearish near-term — lowest cushion vs. freight-rate reset
Schneider National	$1.57B	4.5%	Named partner	Watch — pilot data pending
Knight-Swift Transportation	$2.10B	5.2%	Not named	Bullish — LTL/intermodal cushion
PACCAR	$7.06B (Q2 est.)	12.0%	OEM partner	Mixed — CEO publicly cautious on driver removal
Ryder System	$3.25B TTM	8.5%	Named partner	Mixed — fleet management hedges truckload exposure

Werner Enterprises leaves the smallest margin cushion against the AV squeeze — a TTM operating margin of 2.6% on $3.25B of revenue means the company's $17M of quarterly operating income is roughly what a single 200-driverless-truck fleet run-rate could re-allocate.

Supply chain

Upstream: PACCAR, Volvo, NVIDIA, International — and a Peterbilt CEO who publicly disagrees

Driverless trucks need three things upstream of Aurora's software: an OEM chassis built to integrate redundant steering and braking, a compute platform capable of running SAE Level 4 inference, and a manufacturing partner that can hit the volumes Aurora's plan implies. Each has a publicly listed (or privately disclosed) counterpart in the build, and one of them has a CEO who publicly pushed back on the timeline on July 28, 2026.

PACCAR (Peterbilt, Kenworth, DAF) is a strategic Aurora partner and has been testing Aurora-powered trucks at its facilities since 2021 — but CEO Preston Feight said on PACCAR's Q2 2026 earnings call that PACCAR \"has no plans to take the driver out of its autonomous trucks at this point in time\"
NVIDIA supplies the Aurora Driver compute platform via DRIVE Thor and DriveOS; production samples began shipping in H1 2025 and the chip is part of the second-gen chassis that is 40% smaller than the first-generation unit
International LT Series (Navistar, a Volkswagen Traton subsidiary) is the chassis for Aurora's second-generation commercial driverless fleet; Continental will mass-produce the sensor kit starting 2027
Volvo Trucks' Autonomous Solutions unit has a parallel Aurora Driver integration program on the Volvo VNL Autonomous; a separate Texas corridor with DSV launched in June 2026, and a Volvo-Aurora run to Oklahoma City began in May 2026

We're really happy with the progress. That said, we have no plans to take the driver out of our autonomous trucks at this point in time.
Preston Feight, CEO of PACCAR, Q2 FY2026 earnings call, July 28, 2026 (per Transport Topics)

Risks

Five ways the 2028 free-cash-flow timeline slips

The 2028 FCF target is conditional, not promised. Each prong of the Aurora plan (fleet count, hardware cost, regulatory clearance, DaaS subscription take-up) has a public disagreement already recorded against it. Investors who want to underwrite the 2028 clock should anchor to the catalysts below, not to the headline number.

Regulatory gaps: federal guidelines today still require a human safety operator in autonomous commercial vehicles, and Aurora itself flags insurance coverage for driverless operations as an open item — Federal Motor Carrier Safety Administration frameworks aren't expected to loosen materially until 2027 pilot programs expand
Hardware cost durability: the >50% hardware-bill reduction comes from second-generation kit designed for one million miles of service life; if duty cycles wear sensors faster than modeled, the cost curve reverts and DaaS margins compress
PACCAR public posture: Peterbilt's Peterbilt-brand counterpart isn't pulling the driver out — if a meaningful share of U.S. Class 8 OEM capacity stays in driver-in-cab mode through 2027, Aurora's 2028 truck count is capped by the third-party platforms it's working with
DaaS migration friction: Aurora plans to move each customer from Transportation-as-a-Service to Driver-as-a-Service in 2027, but carriers (especially Werner, Schneider National) must replace existing owned-truck balance-sheet assets with Aurora Driver subscriptions — capex cycle re-opens for the carriers, not Aurora
Driver shortage ceiling: American Trucking Associations estimates the U.S. driver shortfall at roughly 80,000 today, growing to 160,000 by 2030 — if labor tightness forces the wage line up faster than Aurora's DaaS rate compresses, the per-mile cost wedge narrows

Timing

Two clocks investors should set

Investors who treat Aurora's 2028 FCF claim as a freight thesis rather than an autonomous-vehicle headline should track two distinct arcs: a 24-month mechanical build (Q4 2026 truck count, 2027 DaaS transition, 2027 FMCSA pilot expansion), and a 3-year structural reset (carrier net-cost-of-freight, OEM chassis mix, compute attach rate on Class 8).

Short-term (days–quarters): Q4 2026 driverless fleet count target of 200 trucks is the first mechanical test; Hirschbach's 500-truck purchase order (delivery starting 2027) is the first DaaS subscription volume; PACCAR's next earnings call will reveal whether PACCAR's posture on driver removal has shifted
Medium-term (1–2 years): 2027 DaaS migration cadence is the margin unlock — every TaaS customer that converts to DaaS moves Aurora from owning asset risk to running software subscriptions; FMCSA's 2027 framework expansion is the regulatory unlock that broadens the lane map beyond Texas
Long-term (1–3 years): 2028 free-cash-flow inflection is the equity-debt refinancing catalyst for Aurora Innovation; carrier margin compression (or expansion) is the structural read for J.B. Hunt, Werner, Schneider National; OEM platform share is the structural read for PACCAR and International (Navistar/Volkswagen Traton)

Stocks this event actually touches

AAurora InnovationAUR--

--Vol --

Bullish

Aurora Innovation earmarks positive free-cash-flow for 2028 on per-mile economics of roughly $0.85 versus ~$1.30 for human drivers — the first time autonomous trucking has been dated to a specific cash breakeven
Q2 FY2026 ended with $1.2B of liquidity against a $190–220M quarterly burn, so 2028 FCF depends on the fleet reaching roughly 5,000–7,000 driverless units by mid-2028
Second-generation hardware kit cuts the per-truck bill by more than 50% and a 1,000-truck/year Roush facility comes online in October 2026
Key 2027 watchpoints: DaaS migration of existing TaaS customers and Federal Motor Carrier Safety Administration pilot expansion

JJ.B. Hunt Transport ServicesJBHT--

--Vol --

Mixed

J.B. Hunt Transport Services's Q2 FY2026 operating margin of 7.4% (TTM) sits at the top of the listed long-haul group, giving it the most cushion against an AV cost reset — but only for the truckload (JBT) segment
Intermodal (JBI) routes through East Coast and Midwest terminals aren't on Aurora's Sun Belt lanes yet, so near-term AV read-through is muted
Long-term 1–3 year lever: if autonomous Class 8 holds, J.B. Hunt Transport Services's dedicated contract business (DCS) can theoretically shed 15–25% of variable cost

WWerner EnterprisesWERN--

--Vol --

Bearish

Werner Enterprises sits at the lowest margin cushion in the list — TTM operating margin of 2.6% and a $0.76 LTM diluted EPS make the company the most leveraged name to a freight-rate reset
Werner Enterprises is already a named Aurora customer, so fleet substitution risk is asymmetric: short-term (days–quarters) market re-prices the long-haul split against AV capacity coming online
Watch Q3 FY2026 results on October 30, 2026 for signs that autonomous-truck lane availability is compressing their traditional trucking segment pricing

SSchneider NationalSNDR--

--Vol --

Watch

Schneider National is a named Aurora partner with a broader AV pilot footprint than most pure-play truckload names — Q4 FY2026 pilot data is the near-term catalyst
Truckload segment 4.5% TTM operating margin exposes Schneider National to a per-mile cost reset, though its Intermodal and Logistics segments hedge the long-haul exposure
P/E of 50.5× trailing on $0.64 LTM diluted EPS limits margin for execution missteps

KKnight-Swift TransportationKNX--

--Vol --

Bullish

Knight-Swift Transportation is the most diversified truckload exposure with 18,019 tractors plus LTL and intermodal — Q2 FY2026 revenue $2.10B and operating margin 5.2% TTM provide a wider cushion than Werner Enterprises or Schneider National
Near-term (days–quarters): Knight-Swift Transportation has read-across benefits from Q2 FY2026's announced structural recovery narrative; AV lanes add optionality without requiring fleet churn
Long-term (1–3 years): if autonomous Class 8 holds on the long-haul lane and Knight-Swift re-allocates tractors to LTL/dedicated, the company's mixed-business model is structurally defensive

PPACCARPCAR--

--Vol --

Mixed

PACCAR's Q2 FY2026 segment is the critical OEM counterweight — its Peterbilt and Kenworth brands are the standard U.S. Class 8 platforms, with a $7B Q2 FY2026 revenue run rate and TTM operating margin of 12.0%
CEO Preston Feight publicly said on July 28, 2026 that PACCAR has no plans to remove drivers from its autonomous trucks yet — that posture puts Aurora's 2028 truck-count target partially at PACCAR's discretion
Long-term (1–3 years): if autonomous Class 8 reaches commercial scale on PACCAR's platforms, parts and financial-services segments capture attach-rate upside even if truck volume softens

RRyder SystemR--

--Vol --

Mixed

Ryder System is a named Aurora partner and runs a Fleet Management Solutions lease/maintenance business that re-rates positively if autonomous fleets shift to DaaS subscriptions (lessees buy the kit, Ryder System maintains it)
Q2 FY2026 TTM operating margin of 8.5% on $12.8B revenue, but a debt-to-equity ratio of 2.94× (TTM) means Ryder System is levered to a freight-volume slowdown
Hirschbach's 500-truck Aurora-powered order (delivery starting 2027) is the first major frozen/refrigerated fleet to commit to the Ryder System-style maintenance model

NNVIDIANVDA--

--Vol --

Bullish

NVIDIA's DRIVE Thor compute platform is the silicon in Aurora's second-generation driverless chassis (40% smaller than first-gen, per Aurora's July 2026 platform disclosures)
Mass-production with Continental is targeted for 2027 — that's the volume inflection: every Aurora Class 8 unit carries multiple NVIDIA compute modules plus DriveOS
Near-term (days–quarters): Aurora Innovation's Q4 FY2026 driverless truck count of 200 is small relative to data-center volumes, but structurally 2027–2028 mass production would attach NVIDIA to long-haul freight the way it attached to passenger-EV robots in 2024