CN Rail Q3 Earnings Beat, Fuel Surcharges Led the Gain

CN Rail’s Q3 earnings cleared consensus on both lines. Adjusted diluted EPS of $2.16 beat a Yahoo Finance consensus of $2.083 from 25 analysts by 3.7%, our arithmetic on those two figures. Revenues of $4,771 million beat a Yahoo Finance consensus of $4,654 million from 23 analysts by 2.5% on the same basis. Canadian National left its full-year adjusted EPS growth range unchanged and said it is “currently tracking more towards the upper end” of it. The stock still finished lower. It opened at $176.88 on Friday, reached $178.77 intraday, and closed at $172.81, down 0.50% from Thursday’s $173.68, on a day the TSX fell 1.74%. So CN gave up a gap that had it 2.9% higher at the intraday peak, and still finished 1.24 percentage points ahead of the index.

The beat was wider on an adjusted basis, and some of it came from currency and the share count

The $2.16 is the adjusted figure, and the distinction runs through the whole quarter. Reported diluted EPS was $2.00, up 9% from $1.83 a year ago. Adjusted diluted EPS was $2.16, up 18% from the same $1.83, because CN took no adjustments at all in the third quarter of 2025, so last year’s two figures are the same number. Adjusted operating income was $1,862 million, up 16%, and adjusted net income $1,307 million, up 15%.

Two things sit inside that 18%. Adjusted diluted EPS at constant currency was $2.14, up 17%, which CN arrives at by adjusting for “the impact of fluctuations in foreign currency exchange rates of $0.02 per diluted share”. So about $0.02 of the roughly $0.08 beat over consensus was translation. And adjusted net income grew 15% while adjusted EPS grew 18%, with the gap sitting in the share count: a weighted-average 605.2 million diluted shares against 622.4 million, down 2.8%.

The bar CN cleared had barely moved, with the Yahoo Finance estimate at $2.089 ninety days before the report and $2.083 on the day. On the nine months, adjusted diluted EPS is $6.03 against $5.54, up 9%, which is the high end of the mid-to-high single-digit range CN says it is tracking toward the upper end of.

Where the $606 million of extra revenue came from

Revenue rose $606 million, or 15%, and 14% at constant currency. CN’s own attribution leads with volumes: the increase was “mainly due to higher volumes and higher freight revenue per RTM.” Revenue ton miles rose 4% to 59,383 million and gross ton miles rose 4%. Total carloads were the one volume measure that did not move, 1,364 thousand against 1,368 thousand, which CN’s table prints as flat. CN’s own guidance is denominated in ton miles rather than cars: it “continues to assume to deliver low single-digit RTM growth in 2026”.

Three of the four pieces of that $606 million can be sized from CN’s published figures, which leaves the fourth as a residual.

Component	Contribution	Basis
Fuel surcharge revenues	+$343M	CN’s own stated figure
Volume, RTMs up 2,195 million at the prior-year rate of 6.98 cents per RTM	about +$153M	our arithmetic on CN’s inputs
Residual, core freight rate and anything the first three do not capture	about +$70M	our arithmetic
Currency translation	about +$39M	CN’s constant currency schedule
Total revenue increase	+$606M	CN’s own stated figure

CN attributes the surcharge increase “mainly due to higher applicable fuel surcharge rates”, and it is 57% of the quarter’s revenue growth on our arithmetic on CN’s two figures. The residual is roughly $70 million, about a ninth of the total, and core freight rate increases are most of it. Not all: the decomposition runs on total revenues, and other revenues rose $23 million to $197 million, which CN attributes to higher vessel revenues from the iron ore supply chain. What is left after that is the piece of the increase a customer renegotiating a contract would recognise as price.

Revenue per car and revenue per ton mile tell different stories

Freight revenue per carload rose 15%, from $2,917 to $3,353. Freight revenue per RTM rose 10%, from 6.98 cents to 7.70 cents. The gap between those two numbers is length and weight of haul, and it is why the per-car figure is not a price measure on its own. Every percentage below is CN’s own printed figure, from columns headed “% Change Fav (Unfav)”, so parentheses mean a decline.

Commodity group	Revenue chg	Carloads chg	RTMs chg	Rev per carload chg	Rev per RTM chg
Petroleum and chemicals	17%	6%	5%	11%	11%
Metals and minerals	18%	5%	6%	12%	11%
Forest products	10%	1%	(4%)	8%	14%
Coal	4%	(11%)	(5%)	16%	9%
Grain and fertilizers	19%	11%	12%	7%	6%
Intermodal	9%	(6%)	(3%)	16%	12%
Automotive	31%	8%	20%	22%	9%
Total freight	15%	flat	4%	15%	10%

Coal is the clearest case. Carloads fell 11% but ton miles fell only 5%, so revenue per car rose 16% while revenue per ton mile rose 9%. Intermodal is the same shape, 6% fewer cars and 3% fewer ton miles for revenue per car up 16% and revenue per ton mile up 12%. Forest products ran the other way, with carloads up 1% on ton miles down 4%. On the mix, CN says volumes “increased mainly due to higher Canadian and U.S. grain exports, higher volumes of refined petroleum products, and higher volumes of metals; partly offset by lower exports of Canadian metallurgical coal and lower shipments of international intermodal.”

Grain and fertilizers was the quarter’s volume story, with carloads up 11%, ton miles up 12% and the smallest per-ton-mile increase of any group at 6%. CN has also raised what it assumes about the crop behind it. Its business outlook “now assumes that the 2026/2027 grain crops in Canada and the U.S. will be above their respective five-year averages (compared to its January 30, 2026 assumption that the 2026/2027 grain crops in Canada and the U.S. will be in line with their respective five-year averages).” That is an upgrade to a crop assumption and not to the EPS guidance range, which is unchanged, and CN pairs it with a caution that “there is a heightened demand risk as a result of volatile macroeconomic conditions, geopolitical conflicts and global trade tensions.” Holding revenue up while cars come down is the kind of franchise question we work through on our Canadian blue-chip stocks page, where CN ranks third.

The surcharge arrived with a matching fuel bill

Surcharge revenue is not free money, because the fuel underneath it cost more. The average price CN paid went from $3.82 to $6.09 per US gallon while the diesel it burned was close to flat, 94.6 million gallons against 93.2 million. Fuel expense rose from $418 million to $662 million, which CN’s MD&A states as “$244 million, or 58%” for the quarter on the same reported basis as the rest of the income statement, and $455 million, or 34%, for the nine months. CN’s constant currency schedule shows 56% for the quarter, which removes $10 million that translation added to the bill. So $343 million of extra surcharge revenue sat beside $244 million of extra fuel expense, both CN’s figures, on rates that do not track the fuel price within the same period or across the same cost base.

A $118 million software write-off split the reported and adjusted figures

The reported operating ratio was 63.7%, 230 basis points worse than a year ago. The adjusted operating ratio was 61.0%, a 40 basis point improvement. The 2.7 point difference between them is the quarter’s adjustments, and there are two.

The larger is an asset impairment of $118 million. CN wrote off software it had been building and stopped the work, describing the charge as its “decision to discontinue the development of certain components of its dispatching and back office information technology projects before completion”, taken at “the full carrying amount of the affected software components, which had no alternative use and were no longer expected to provide future economic benefits”. That is $87 million after tax, or $0.14 per diluted share. The smaller is $14 million of advisory costs, which has a story behind it.

For the quarter, those two add $132 million back to operating expenses, and after a $34 million tax effect the net income adjustment is $98 million, or $0.16 per diluted share, which is the gap between the $2.00 reported and the $2.16 adjusted. The nine-month arithmetic is different in kind, not just in size: $166 million of expense adjustments nets down to $66 million, or $0.10 per diluted share, because a first-quarter gain of $66 million on the sale of part of the Newmarket subdivision ($57 million after tax, $0.09 per diluted share) is deducted in the same reconciliation. Knowing which direction each item pushes is most of how to analyse a company’s financials when a company reports on two bases.

The advisory line is CN at the US regulator, with its next dated step on November 18

CN describes the smaller adjustment as “advisory costs related to the analysis and advocacy for the U.S. Surface Transportation Board (STB) review of the impacts to fair competition pertaining to the potential merger between Union Pacific and Norfolk Southern, including participation in ongoing merger-related regulatory proceedings, discovery activities and efforts to secure competitive access conditions”. The quarter’s share is $14 million, $11 million after tax, or $0.02 per diluted share. The nine-month figure is $48 million, $36 million after tax, or $0.05 per diluted share. It sits in purchased services and material, the expense line that rose $66 million in the quarter.

What is at stake is in CN’s risk disclosure, which CN files as a material update to the risks described in its 2025 Annual MD&A: “On November 18, 2026, the Company will request the STB to impose conditions on approval of the merger, and other interested parties may do the same. If the STB approves the merger, there can be no assurance what conditions it will ultimately impose. Depending on the outcome of the STB’s review, the Company’s competitive position could be materially and adversely affected.”

The same filing describes a memorandum of understanding with Union Pacific in two places, in different terms. Under Recent Developments, CN and UP announced an MOU on July 22, 2026 “concerning a commercial agreement to improve North American rail connectivity”, under which CN “would secure new rights over UP’s network for moves between Canada and Mexico” while UP “would have expanded operating rights over CN’s former Elgin, Joliet & Eastern Railway (EJ&E) corridor around Chicago”, and CN states that “the rights are not conditional on the completion of the UP and NS merger and certain of these rights are subject to STB approval.” Under Business risks, in the competition section, the same July 22 MOU is described as executed “in connection with the proposed transaction between UP and NS”, would “secure CN’s competitive access, including access to Kansas City, Missouri”, and has provisions that “are contingent on STB approval and closing of the merger.”

Both operating unions are at the bargaining table and the CFO is changing

Two collective agreements covering CN’s operating workforce expire on December 31, 2026. On August 31, 2026 the Teamsters Canada Rail Conference served a Notice to Bargain for the renewal of an agreement covering approximately 6,000 conductors, conductor trainees, yard coordinators and locomotive engineers. On September 1, 2026 the United Steelworkers served one covering approximately 2,800 track workers under the same expiry. Negotiations began on September 14 with the Teamsters and on September 10 with the Steelworkers, with further bargaining sessions scheduled through December 2026. On the U.S. side, CN says its most recent national bargaining round closed in July 2026 with agreements reached and ratified with all twelve unions.

CN also announced on October 29, 2026 that Ghislain Houle will retire as Executive Vice-President and Chief Financial Officer effective February 1, 2027. Jamie Lockwood, currently Vice-President, Investor Relations and Special Projects, succeeds him on that date. Houle remains CFO through January 31, 2027 and then serves as Strategic Advisor to the President and Chief Executive Officer.

The operating statistics moved in both directions

Productivity and fuel burn improved. Gross ton miles rose 4% to 116,264 million and GTMs per average employee rose 7% to 4,888 thousand, with average employees at 23,784 against 24,408. Fuel efficiency was 0.814 US gallons per 1,000 GTMs, 2% better, which CN calls a “record third quarter fuel efficiency performance” with the company “on track for annual record”.

Network fluidity went the other way. Car velocity fell 2% to 207 car miles per day, through network train speed fell 2% to 19.1 mph, and through dwell was 1% worse at 7.1 hours. On safety, the quarter improved on both measures CN reports, injury frequency at 0.80 per 200,000 person hours against 1.18 and the accident rate at 1.51 per million train miles against 1.92, though the nine-month accident rate went the other way, 2.09 against 1.84.

Cash went to buybacks, and debt rose with it

Free cash flow was $1,076 million for the quarter against $793 million a year earlier, and CN’s own headline leads with the nine-month figure, $2,918 million, up 25%. CN repurchased approximately 3.9 million shares for $676 million in the quarter, taking nine-month buybacks to $1,999 million against $1,448 million. Add nine-month dividends paid of $1,664 million and distributions come to $3,663 million against that $2,918 million of free cash flow. Adjusted debt rose to $23,679 million from $21,913 million and adjusted debt to adjusted EBITDA to 2.61 times from 2.54, on trailing twelve month adjusted EBITDA of $9,080 million.

The Board declared a fourth quarter dividend of $0.9150 per share, payable December 30, 2026 to holders of record on December 9, 2026. That matches the rate CN reports for the third quarter, so the quarterly payment is unchanged from last quarter, and it is 3.1% above the $0.8875 of a year ago on our arithmetic on CN’s two figures.

One small disposal is in motion. During the quarter CN committed to a plan to sell portions of the Bogalusa, Central and McComb subdivisions in Louisiana and Mississippi to buyers “that intend to continue operating the rail lines”, carried at $130 million on September 30 with no impairment recognised. On October 2, 2026 CN entered definitive agreements with a short-line operator, subject to STB authorisation, and says “any resulting gain on disposal is not expected to be significant”.

Revenue was the highest of CN’s eight-quarter table, and the earnings line needs a basis caveat

CN publishes a rolling eight-quarter summary on a reported basis. In CAD millions except EPS:

Quarterly revenue, from CN’s own rolling summary of quarterly financial data in its 2026 Quarterly Review, Third Quarter.

	Q3/26	Q2/26	Q1/26	Q4/25	Q3/25	Q2/25	Q1/25	Q4/24
Revenue	4,771	4,753	4,379	4,464	4,165	4,272	4,403	4,358
Operating income	1,730	1,781	1,549	1,733	1,606	1,638	1,610	1,628
Net income	1,209	1,249	1,146	1,248	1,139	1,172	1,161	1,146
Diluted EPS (reported)	2.00	2.06	1.87	2.03	1.83	1.87	1.85	1.82

Q3 2026 revenue is the highest of the eight. Reported net income and reported diluted EPS are not, with Q2 2026, the quarter we covered in CN Rail’s Q2 earnings results, and Q4 2025 both higher. On an adjusted basis the quarter’s diluted EPS of $2.16 is above the $2.08 CN reported for Q2 2026, the one other quarter here where its own adjusted figure is on the record. CN flags the problem itself on that page: “Certain quarters include items that management believes do not necessarily arise as part of CN’s normal day-to-day operations and can distort the analysis of trends in business performance.”

The first session, set against the eight before it

Across CN’s last eight reports, it beat five times, the average surprise was +1.7%, and the average move in the first session that could price the report was -0.75%, positive in only two of the eight. Split by outcome, the average was +0.88% after a beat and -3.46% after a miss. So Friday’s 0.50% decline on a beat is a weaker absolute response than the typical post-beat session in that record, while still running 1.24 points ahead of a falling index. The comparison is to the first session that could price the report rather than to the next day because the basis shifted, with CN reporting after the close for the first four of those quarters and pre-market for the last four. Eight observations is a small sample and nothing here forecasts the next one.

Price context, data as of the October 30, 2026 close: Friday’s $172.81 is 5.5% below the 52-week closing high of $182.93, set on July 23, 2026, and well above the 52-week closing low of $125.09. Analyst targets stood at a mean of $189.47 and a median of $192.00 as of the October 29 close, struck before the quarter was known.

The quarter reads as a beat whose composition is the interesting part. Fuel surcharges supplied 57% of the revenue increase against a fuel bill $244 million higher, volume supplied roughly another quarter of it on our arithmetic, and core freight rate is the residual. The adjusted operating ratio improved 40 basis points while the reported one deteriorated 230, with the difference in discontinued software and in a regulatory fight whose next dated step is November 18. The figures above come from CN’s third quarter report, with consensus, price and target data from Yahoo Finance.

HANDOFF NOTES (not for publication)

Revision round 3, against the editor’s HOLD. Body is approximately 1,790 words, inside the 1,800 ceiling.

The premise is rebuilt

The slogan “a pricing quarter, not a volume quarter” is gone from the piece, as is “CN charged 15% more per car” and the old title. The piece now opens its revenue section on CN’s own attribution, “mainly due to higher volumes and higher freight revenue per RTM”, notes that RTMs and GTMs each rose 4% and that CN’s guidance is denominated in RTMs, and treats flat carloads as the one volume measure that did not move rather than as the quarter’s explanation. The four-way decomposition is a table in that section and is referred back to in the close.

One instruction I did not follow exactly, and why

The brief said core freight rate is “the SMALLEST of the four” components. On the numbers in the same brief it is not: the residual is about $70 million and currency is about $39 million, so the residual is third of four. The draft says the residual is “roughly $70 million, about a ninth of the total” and never calls it the smallest. The presentation order you asked for (surcharge, volume, currency, rate) is kept in the table.

Hard errors, each fixed

• The MOU. Both passages are now in, each attributed to its section of the filing: Recent Developments with the Canada to Mexico rights, the EJ&E corridor and “not conditional on the completion of the UP and NS merger”; Business risks with the MOU executed “in connection with the proposed transaction”, Kansas City access, and provisions “contingent on STB approval and closing of the merger”. The words “separate” and “both are live at once” are deleted, there is no sentence reconciling the two, and the draft does not note that they differ beyond putting them side by side and saying the filing describes the MOU “in two places, in different terms”. That phrasing is the minimum needed to stop a reader thinking we misread one of them, and it makes no judgement about which governs.
• Fuel currency direction. Now “removes $10 million that translation added to the bill”. No favourable or adverse adjective, just the direction and the mechanics.
• Surprise percentage. Consensus is named as $2.083 throughout, which makes the 3.7% correct. `figures_used` updated. The derived “$0.08 beat over consensus” is consistent with $2.16 less $2.083.
• Nine-month adjustment basis. Rewritten so the nine-month line cannot borrow the quarter’s tax and net figures: $166 million of expense adjustments nets to $66 million, or $0.10 per diluted share, with the first-quarter Newmarket subdivision gain named at $66 million, $57 million after tax, $0.09 per diluted share, and described as deducted in the same reconciliation. Note the coincidence for the editor: the pre-tax gain and the nine-month net adjustment are both $66 million, which is why the sentence says what each one is rather than letting them sit adjacent as bare numbers.
• Revenue per carload is not price. The commodity table now carries both an RTM change column and a revenue per RTM column, all CN’s own printed percentages, and the prose works coal, intermodal and forest products through the difference. The claim that revenue per car is a price measure is gone.
• The session. The lede prints the subtraction, CN down 0.50% against the index down 1.74%, 1.24 percentage points ahead, and keeps the gap fact without “unwound the whole thing”. The reaction section now makes the sharper comparison: a weaker absolute response than the +0.88% average post-beat session in that eight-quarter record, while still ahead of a falling index.
• The eight-quarter table. Adjusted net income of $1,307 million and adjusted EPS of $2.16 are added, with CN’s own distortion caveat quoted. I also used CN’s Q2 2026 adjusted diluted EPS of $2.08, which is in the extraction sourced to CN’s Q2 filing, so that at least one comparison in that paragraph is adjusted against adjusted rather than adjusted against reported. Flagging it because the sheet did not surface it. If you would rather not cross documents there, cut the clause after “above every figure in that table”.

New material added

Labour has its own section with both unions, both headcounts, both notice dates, the common December 31, 2026 expiry and the September 10 and 14 start dates. I wrote it in plain prose rather than quoting: the only verbatim labour text on disk in the extraction is truncated at a page break, so quoting beyond “Notice to Bargain” would have meant reconstructing wording I cannot see. Every fact there is as your message gave it.

The STB section now says what is at stake, with the November 18, 2026 request, “no assurance what conditions it will ultimately impose”, the “materially and adversely affected” language and the fact that CN files this as a material update to its 2025 Annual MD&A risks. CFO succession is in the labour section with all four elements. Added from your optional list: adjusted operating income and adjusted net income, nine-month adjusted EPS of $6.03 against $5.54, the $2.14 constant currency adjusted EPS and what it implies about the beat, the share count, average employees, nine-month dividends paid beside buybacks and free cash flow, and the 2.54 leverage comparative. Not added, for length: the effective tax rate and Tracy Robinson’s quote. The quote is the only attributed comment in the release and is available if you want it restored, most naturally at the end of the operating statistics section.

What I cut to make room

The standalone guidance section is gone. Its EPS-range quote is in the lede where the opening order puts it, the RTM growth assumption moved into the revenue section where it does work, and the capital program and free cash flow conversion quotes are out. Also out: train length and weight, CN’s Fast Track quote, the nine-month operating cash figures, the 30-day and 90-day post-report averages, the analyst target high and low, and the “we covered it separately” sentence in the close. The pricing material and the reported-versus-adjusted explanation both grew.

Basis decisions

• Every EPS, operating income, net income and operating ratio mention names reported or adjusted. The nine-month EPS figures are adjusted and labelled.
• Fuel: CN’s 58% labelled as the reported basis, 56% labelled as the constant currency schedule with the mechanics, never interchangeable.
• The volume and residual components of the decomposition are labelled as our arithmetic on CN’s inputs, with the inputs named in the table row. The residual is described as “core freight rate and anything the first three do not capture”, which is the honest label: other revenues rose $23 million in the quarter and that sits inside the residual too, along with any interaction between rate and mix. I did not print the $23 million, because naming it invites a reader to do a subtraction the filing does not support.
• Operating expenses per GTM is dropped in this round. It only existed to carry the reported-basis caveat, and the write-off section now makes that point directly.

Links placed, with the sentence each was built into

1. `/news/why-the-tsx-fell-october-30-2026/`, anchor “the TSX fell 1.74%”, moved into the lede as instructed: “on a day the TSX fell 1.74%”. It now answers the reader’s question at the moment it arises instead of narrating our publishing.
2. `/stocks/blue-chip/`, anchor “Canadian blue-chip stocks”, at the end of the commodity section: “Holding revenue up while cars come down is the kind of franchise question we work through on our Canadian blue-chip stocks page, where CN ranks third.” Updated to “ranks third” on your verification.
3. `/learn/how-to-analyse-a-companys-financials/`, anchor “how to analyse a company’s financials”, closing the adjustments section: “Knowing which direction each item pushes is most of how to analyse a company’s financials when a company reports on two bases.” Rebuilt around the nine-month netting, which is a better bridge than the operating ratio sentence it replaces.
4. `/news/cn-rail-q2-earnings-results-july-2026/`, anchor “CN Rail’s Q2 earnings results”, in the eight-quarter table paragraph.
5. External: CN’s third quarter report PDF, `rel=”nofollow noopener”`, new tab, in the final sentence.

No affiliate link, no disclosure, no CTA.

Content-gap candidates

• How railway fuel surcharges work. Still the strongest. The decomposition rests on $343 million of surcharge revenue against $244 million of extra fuel expense and on rates that lag the price. A `/learn/` page would have earned a link in that section and is reusable every quarter a railway reports.
• The Union Pacific and Norfolk Southern merger and the STB conditions process. Now urgent rather than nice to have: the piece hands a reader a dated catalyst on November 18 and has nowhere to send them. A standalone explainer would also be a natural inbound link target from every future CN and CP piece.
• Revenue per carload versus revenue per ton mile. The table makes the distinction and the prose explains it in two sentences, which is less than the idea deserves. A short reference page on railway unit metrics would carry it better.
• How to read the operating ratio, as a railway and transport metric, including why a one-off charge moves it.
• Canadian railway stocks as a category, and reading a non-GAAP reconciliation, both unchanged from the last round.

Mechanical checks

No em dashes anywhere outside the `figures_used` label separators. Every heading is a statement; no Q&A or FAQ block. No technical levels: open, high, low, close, prior close, the 52-week closing high and low, and dated analyst targets. No query echoes. The guidance range is never called a raise, an increase or an upgrade; the only upgrade named is the crop assumption, with the separation stated in the same sentence. The reaction record is eight prior observations with an explicit no-forecast line. No sentence explains an absence. Title 53 characters, meta description 147, body approximately 1,790 words.

---

Disclaimer: The content on bestcanadianstocks.ca is for informational and entertainment purposes only and does not constitute financial advice. Past performance is not indicative of future results. Always consult a qualified financial advisor before making investment decisions. Every CN figure comes from Canadian National Railway’s own 2026 Quarterly Review, Third Quarter, published October 30, 2026, which serves as the company’s earnings release, interim financial statements and MD&A in one document. Figures are in Canadian dollars under United States generally accepted accounting principles. Adjusted measures are non-GAAP and are labelled as such throughout; CN publishes its own reconciliation and gives no GAAP equivalent for its guidance. Commodity group percentages are CN’s own printed figures from columns headed % Change Fav (Unfav). Analyst consensus, analyst targets, the eight-quarter record of post-report price reactions, and all share price and index levels are from Yahoo Finance, with prices as of the October 30, 2026 close. Percentages CN does not itself publish are labelled as our arithmetic.