Insights

Reading the Numbers the Industry Publishes

Class I railroads disclose a great deal — fuel efficiency, emissions targets, network velocity. Very little of it is analysed against the operating reality behind it. These are the write-ups I publish for the suppliers and carriers who have to act on those numbers, reproduced here in full with the underlying charts.

33,200+ impressions LinkedIn analytics
17,985 people reached LinkedIn analytics
283 reactions visible on the posts
61 comments visible on the posts
22 saves all activity, 90 days

Across six in-depth analyses published since July 2026. Impressions and reach are from LinkedIn's own post analytics; reach counts distinct members, 82% of them outside my own network. Saves are LinkedIn's 90-day total across all my activity, not the six analyses alone. Reaction and comment totals as of 31 August 2026.

Series · 3 parts

Class I Railroads vs. SBTi Targets

A three-part read of how North America's Class I railroads are tracking against their Science Based Targets commitments — and why the type of target changes the entire conversation.

Part 1 of 324 August 2026

Intensity targets: fuel efficiency is a half-strength lever

Four Class I railroads carry emissions-intensity targets measured per gross ton-mile. Fuel efficiency and emissions intensity look like twins over 2019–2026E — 9.9% versus 9.7% cumulative improvement, an r of 0.95 on levels. Year over year, the relationship falls apart. The regression slope says a 1% fuel-efficiency gain returns only about 0.47% of intensity reduction, because Scope 2 sits in the intensity numerator and never appears in a fuel number, and biofuel cuts reported emissions while acting as a headwind to pure efficiency. A savings case built on gallons travels only half the distance toward the target the customer is actually measured on.

Four intensity-target Class I railroads, 2019–2026E. Year-over-year changes are not statistically significant (r = 0.63, p = 0.13).

Join the discussion on LinkedIn

Part 2 of 326 August 2026

Absolute targets: when commercial success moves the goalposts

Union Pacific and BNSF carry absolute reduction targets rather than intensity targets, and absolute behaves nothing like intensity. Their combined emissions track traffic volume almost one for one — r = 0.96, with a regression slope of 1.11 — so volume, not operating performance, is the dominant variable. Most of the reduction those two can claim today came from the 2018–2020 volume collapse, and combined 2026E volume is on pace for the highest since 2021. That creates a real tension: every ton won from the highway lowers society's emissions and raises the railroad's reported number. Rail is roughly three to four times more fuel efficient per ton-mile, which is precisely why the conflict matters.

The two absolute-target Class I railroads. A 1% change in gross ton-miles moves reported emissions about 1.1%.

Join the discussion on LinkedIn

Part 3 of 328 August 2026

Tracking to 2030: how both groups are really doing

Both camps miss, and by similar margins. Modelling each carrier's trend forward with a Monte Carlo on efficiency, biofuel blend ramp and traffic volume puts the intensity group's 2030 outcome about 23% above its own goal and the absolute group about 25% above. The arithmetic is the obstacle: hitting these targets needs roughly 8% annual reduction over four years against a delivered run rate of 1.5 to 2% a year — a four- to fivefold sustained acceleration, with no volume downturn to help. One carrier has better than a one-in-ten chance, BNSF, and about two thirds of that edge traces to target design rather than performance: a 30% absolute goal against Union Pacific's 50.4%, set off a 2018 base year that was a cyclical volume peak. Higher biofuel blends are the cheapest ton of CO₂ avoided by a wide margin, but closing these gaps on biofuel alone implies fleetwide B25 to B45 by 2030 against OEM approval that today tops out at a conditional B20. That is the commercial opening: the gap is real, the levers are known, and the qualification work is not done.

Monte Carlo simulation, 200,000 draws, on operating-efficiency trend, biofuel-blend ramp and traffic volume. Published goals and base years; the 2026 estimates, forward simulation and probabilities are my own.

Join the discussion on LinkedIn

12 August 2026

Class I fuel efficiency is moving again after four flat years

For decades the industry eked out roughly 1% a year in gallons per 1,000 gross ton-miles, mostly by buying newer locomotives. PSR and energy management systems changed the slope, then 2021 through 2025 went essentially flat — with one carrier quietly improving close to 3% a year and closing a gap that had persisted for two decades. Regression on first-half results puts 2026 average improvement above 2%, more than double the historical trend. One caveat stated plainly: the metric is sensitive to traffic mix, not just efficiency, so losing dense coal tonnage makes the number look worse even when operations improve.

Seven Class I carriers, 2000–2026E. The 2026 figures are Kennedy Consulting estimates, not company-reported.

Join the discussion on LinkedIn

Series · 2 parts

Network Velocity and the Fuel Bill

What happens to locomotive fuel efficiency when a railroad absorbs a demand surge faster than it can resource it.

Part 1 of 220 July 2026

The network is slowing down — and burning fuel doing it

Class I composite velocity fell from 22.25 mph in January to 20.54 mph by July, down almost 8% in seven months while traffic climbed. Rising volume is good news, but slower networks mean more stops, more starts and more idling — the three biggest drains on locomotive fuel efficiency. Congested trains burn diesel going nowhere. The recoverable part matters most: idle reduction, smarter train make-up and throttle discipline claw back real gallons without waiting for speeds to rebound.

Class I composite velocity, January to July 2026.

Join the discussion on LinkedIn

Part 2 of 227 July 2026

The evidence behind the slowdown

Plotting U.S. weekly rail traffic against Class I composite velocity from January 2025 through June 2026 puts numbers to the claim. Volume rose 11.7%, from 467k to 522k average weekly units, while velocity fell 6.8%, from 22.25 to 20.74 mph. Across 18 monthly points the Pearson correlation is −0.62, and the regression says every additional 10,000 weekly units is associated with roughly a 0.19 mph drop in network velocity. The 2026 months sit systematically high-volume and low-velocity, below and to the right of the 2025 cloud — the fingerprint of a network absorbing a demand surge it is not fully resourced to move. Every 1 mph lost adds stops, restrictive-signal time and idle hours.

18 monthly points, January 2025 – June 2026. Pearson r = −0.62.

Join the discussion on LinkedIn

Want this analysis pointed at your own numbers?

The same methods behind these posts — regression on reported results, benchmarking against peer carriers, and separating genuine efficiency from traffic mix — are what I bring to client work. If a fuel or emissions number on your side is not behaving the way it should, that is a conversation worth having.

Start a conversation Follow on LinkedIn