Union Pacific Corporation

Union Pacific owns and runs 32,889 miles of railroad across 23 states in the western two-thirds of the United States. Its trains carry grain from Nebraska to the Gulf ports, chemicals and plastics from the Texas coast, coal from Wyoming's Powder River Basin, cars and car parts to and from Mexico, and, most of all, shipping containers from the ports of Los Angeles and Long Beach to Chicago. A ten-year-old could follow it: if you need to move something heavy a long way over land, a train does it for about a quarter of the fuel a lorry would use, and in the West there are only two railroads to choose from.

In the twelve months to September it took in $26.4bn and earned $7.6bn, a net margin of 29 cents on the dollar, which would be extraordinary for most industrial businesses and is normal for a well-run American railroad. It owns its track, its yards and most of its locomotives, so it is also extremely capital-hungry: it spends about $3.5bn a year, roughly 13% of revenue, simply to keep the network safe and moving.

The essential number in railroading is the operating ratio: the share of every revenue dollar spent to run the railroad. Lower is better. Union Pacific's was 59.1% in the third quarter of 2026, the best of the large American railroads. A railroad with a 59% ratio keeps 41 cents of every dollar before interest and tax; one with a 68% ratio keeps 32. That nine-point gap is the difference between a good business and a great one.

The third quarter of 2026, reported on 22 October, looked spectacular on the surface: revenue up 15% to $7.2bn, earnings per share up 16% to $3.49. Much of it was fuel. Locomotive diesel averaged a record $4.21 a gallon against $2.56 a year earlier, and Union Pacific passes most of that to customers through fuel surcharges, which doubled to $1.2bn. Revenue rises, profit barely notices, and the operating ratio looks worse, by 1.4 points in the quarter. Without fuel, freight revenue grew about 6%, carloads 6% and intermodal 11%, helped by a surge in West Coast imports. That is a strong quarter for a railroad. It is not a 15% growth business.

The productivity record under Jim Vena, chief executive since August 2023, is the real story of the last three years. Trains are longer (about 9,970 feet), cars move further each day (228 miles), each employee moves 5% more car-miles than a year ago, and the headcount is down to about 28,500. Those are the levers of "precision scheduled railroading": fewer, longer, fuller trains on a tighter timetable. The cost of pushing them hard shows in the service measures, which slipped 6–7 points in the third quarter as volumes surged. Shippers notice that, and so does the regulator.

* Includes $234m of land sales (about $0.30 a share). Guidance for 2026: earnings per share up by a "high single digit" percentage, which on 2025's $11.98 implies about $12.80–13.10; capital spending $3.5bn.

Buffett explained the attraction when Berkshire bought BNSF in 2009–10: railroads are essential, they are far more fuel-efficient than lorries, and their networks cannot be duplicated. To build a competing line from the Pacific to the Midwest today you would need tens of thousands of acres of land, decades of permits and tens of billions of dollars, and at the end you would split a market that two incumbents already serve. Nobody will do it. The western United States is, and will remain, a duopoly: Union Pacific and BNSF, two railroads of almost exactly the same size (32,889 and about 32,500 route miles).

The comparison with BNSF is the most useful one an owner has, because it is the same business, on the same geography, run by a different owner: Berkshire. In 2025 BNSF ran at a 65.5% operating ratio and Union Pacific at 59.8%; in the third quarter of 2026, 63.9% against 59.1%. Union Pacific is the better operator. But the gap was eight points in 2024 and is about five now, as BNSF has cut costs. The lesson is that Union Pacific's advantage is management, which can be copied, laid on top of a franchise, which cannot.

On 29 July 2025 Union Pacific agreed to buy Norfolk Southern, the eastern railroad that runs from Chicago and St. Louis to the Atlantic ports. Each Norfolk Southern share would receive one Union Pacific share plus $88.82 in cash, worth $320 when the deal was struck and valuing the company at about $85bn including debt. Union Pacific would issue about 225 million shares, leaving Norfolk Southern's owners with 27% of the combined company, and borrow about $20bn. Jim Vena would run it from Omaha and has committed to stay at least five years.

The case for it is a single-line railroad from coast to coast. Today a container from Los Angeles to Atlanta changes railroads in Chicago or Memphis, which adds days, cost and a handover where things go wrong. A combined railroad could offer a through service and take lorries off the interstate: the companies talk of converting 10,000 interline routes into single-line ones and shifting more than two million truckloads a year. They promise $2.75bn a year of extra EBITDA by the third year, $1.75bn of it from winning new traffic and $1.0bn from cost savings, and say the deal will add to earnings per share from its second full year, rising to a high-single-digit percentage.

The regulator. The deal needs approval from the Surface Transportation Board under its 2001 rules for major mergers, which have never been applied. They require the railroads to show that the combination enhances competition, not merely that it does not harm it. The first application, filed in December 2025, was rejected as incomplete in January 2026 because it lacked projections of market share and the full merger agreement, including the clause that lets Union Pacific walk away if the conditions are too burdensome. A revised application was filed in April and accepted in May; the Board set its schedule in August and in September unanimously refused to throw the case out. Comments, protests and requests for conditions are due on 18 November; the railroads' rebuttal on 16 February 2027; final briefs on 28 May 2027; a decision within 90 days after that. Union Pacific's own target for completion has moved from "early 2027" to "the first half of 2027" to "the second half of 2027".

The opposition is broad. BNSF, Berkshire's railroad, and CPKC lead a coalition against the deal; CSX joined a motion to dismiss it; agricultural and chemical shipper groups have lobbied the White House; seven state attorneys general have asked the Board to block it; some rail unions oppose it while others have signed jobs agreements. CN, which had objected, settled in July in exchange for terminal rights in St. Louis.

The remark. At a shippers' conference in March 2026, Vena spoke of raising rates on opponents of the merger "to teach you a lesson". The company says it was a joke and that it does not retaliate. On 4 November the Board, citing "serious concerns" and "zero tolerance for retaliation", ordered the video and related documents released; Baird cut the shares to neutral the next day, and they fell 3.4%. It is the most damaging moment in the case so far, because it hands opponents exactly the story they want to tell, about a railroad with too much power over captive shippers, two weeks before their comments are due.

What it does to an owner. Three things worry us more than the headlines. First, Norfolk Southern has got worse since the deal was signed: its operating ratio for the first nine months of 2026 was 68.5% against 62.9% a year earlier. Union Pacific would be paying, at today's prices, about 25 times Norfolk Southern's trailing operating profit for a railroad that earns much less on each dollar than its own. Second, 64% of the promised synergies depend on winning new traffic, the least reliable kind of synergy, especially if the Board imposes conditions that open the network to rivals. Third, history: large rail mergers have a habit of breaking service for a year or more, and Union Pacific's own 1996 merger with Southern Pacific is the textbook case.

The arbitrage gap. At yesterday's prices the offer is worth $393.02 per Norfolk Southern share, while the shares trade at $333.24: a gap of 17.9%. Markets rarely leave that much on the table for a deal they expect to close soon on the agreed terms. The gap says the odds are well short of certain, or that conditions may be heavy. If the Board says no, Union Pacific pays a $2.5bn break fee (about $4.20 a share, 1.4% of its market value), writes off its costs and, we assume, restarts the buybacks. If it says yes, the owners get a bigger, more indebted, lower-margin railroad, with the promise of synergies.

As an operator, Vena has delivered. The operating ratio improved from 62.3% in 2023 to about 59–60%, productivity rose every year, and the third quarter's volume surge was carried with fewer people. That is what the board hired Vena to do, and the numbers show it was done.

As a capital allocator, the record is about to be decided by one transaction. The 2024 investor day promised buybacks of $4–5bn a year from 2025. Those were stopped when the merger was announced, and Union Pacific has bought almost no shares in 2026. The merger itself is Vena's: the bet, the timing and the price. If it works, it will be remembered as the deal that built America's first transcontinental railroad in modern form. If it fails or goes badly, owners will have given up two years of buybacks and paid a break fee for it.

The remark matters more than its words. A railroad with captive customers lives by a social contract: it is allowed high returns because it does not abuse the customers who have no alternative. A chief executive who jokes in public about raising rates on opponents gives the regulator reason to doubt that contract, at the moment the regulator is deciding whether to let the railroad grow larger. We do not doubt Vena's ability. We doubt the judgment of that sentence, and the Board has said it will look at the documents.

The finance chief is leaving in the middle of the most complex financing and regulatory process in the company's recent history. A good successor can be found; the timing is still unhelpful.

The buyback machine. From 2016 to 2025 Union Pacific spent about $43bn buying back its own shares, more than twice what it paid in dividends, and reduced its share count by 29%, from 835 million to 596 million. Much of it was financed with debt, which rose from $15bn to $32bn over the same decade. That was a deliberate policy: borrow cheaply against a stable railroad and return the cash. It worked while rates were low. The 2025 purchases averaged $227 a share, well below today's price. Since the merger was announced, the machine has been off.

The balance sheet today is solid for a railroad. At 30 September Union Pacific had $30.3bn of debt and $3.1bn of cash and short-term investments; its own measure of adjusted debt to adjusted EBITDA was 2.4×, down from 2.7× at the end of 2025. Interest is covered about 8 times by operating profit. The pension is overfunded by about $700m. Ratings at the last bond issue were A3/A-/A-. The debt is cheap: its market value is $7.5bn below its book value, because most of it was issued at low coupons and runs to as late as 2072.

After the merger the picture changes. About $20bn of new debt would take leverage to around 3.3–3.6 times EBITDA at closing, by the companies' own guidance and our arithmetic. Fitch put the rating on watch positive in 2025, expecting leverage to fall below 3 within 18 months; the 10-K warns that agencies may take "negative actions". A combined company would still be investment grade, but the era of borrowing to buy back shares would be over for several years.

The financial statements of a railroad are simple to read once three adjustments are made: fuel surcharges inflate revenue when diesel is expensive and shrink it when it is cheap; land sales and one-off charges move the earnings line from quarter to quarter; and capital spending, about $3.5bn a year, is mostly maintenance, not growth. On that basis Union Pacific earned about $12.83 a share in the last twelve months and turned about $11.26 a share into free cash flow (operating cash of $6.69bn after capital spending, before the dividend).

Union Pacific has paid a dividend for 127 consecutive years and raised its annual dividend for 20. The quarterly rate rose 3% in July to $1.42, $5.68 a year, a yield of 1.9%. The company aims to pay out about 45% of earnings, and does.

The dividend is safe and modest. At 1.9% it is not the reason to own Union Pacific; the reason is the railroad. For most of the last decade the larger return to owners came through buybacks, and those are paused. An income investor looking at this stock today is being paid less than 2% to wait for a regulator.

The model. We value Union Pacific on its own, as a railroad, and treat the merger as a risk to the range rather than a source of value: the market already gives Norfolk Southern a value close to the offer, and the synergies are promises. We start from 2026 earnings per share of about $12.95, the midpoint of the company's guidance, grow them to 2030, apply an exit multiple, discount 4.1 years at 10% and add the dividends collected on the way, about $23 a share.

The middle case moves to about $230 at 18 times and $276 at 22 times; to $237 at 5% growth and $271 at 9%; to $262 at a 9% discount rate and $245 at 11%.

What is priced in? At $304.20, on a 20 times exit multiple, Union Pacific needs earnings per share to grow about 12.5% a year to 2030. That is above the top of the range the company itself set at its 2024 investor day ("high-single to low-double digit" over three years), and it is what a successful merger would have to deliver on top of a strong railroad. The price is close to our "everything goes right" case.

Our rule for wonderful businesses. We do not demand a bargain for a business of this quality: we would be content to buy it at around our central estimate of value. Union Pacific is not there. It is about 20% above it, at the top of its own five-year range of multiples (18.5× to 25.3×), with the buyback switched off and a merger outcome that could move the value either way. We would start to buy at about $240: close to our central value, about 18.5 times this year's earnings, with a 2.4% yield.

The striking thing about this table is that no row makes Union Pacific's owners much richer in the near term. The best outcome swaps a simpler, very profitable railroad for a larger and more indebted one with the hope of synergies; the worst pays a fee and goes back to where it was. Buffett's own railroad is fighting the deal, which tells you BNSF fears a stronger competitor. That is not the same as telling you Union Pacific's owners will gain. An owner who already likes the railroad loses little by waiting for the regulator's decision, and an owner thinking of buying at 24 times earnings is paying in advance for the best outcome.