Union Pacific Corp UNP

Thesis

Union Pacific is one of two Class I railroads west of the Mississippi (the other being BNSF, owned by Berkshire). It moves bulk goods - grain, coal, chemicals, intermodal containers, autos - on 32,889 miles of right-of-way that no rational competitor would or could rebuild. The economics of this business are well understood by Buffett, who bought BNSF outright in 2010 and has written about rail repeatedly: huge investment in very long-lived, regulated assets, recession-resistant demand, fuel efficiency 4x that of trucking [4], and a regulatory compact that allows reasonable returns on capital provided service is reliable.

The scorer rates UNP composite 83 (profitability 21, balance sheet 20, capital allocation 20, valuation 22). Ten-year average ROIC is 14.05% with 5-year ROIIC of 20.33% - the marginal dollar reinvested has earned more than the average dollar. FCF conversion of 95.6% is high quality. Net debt to EBITDA of -0.06 looks like a data anomaly (the company carries real debt) but interest coverage is sufficient by industry standards. Share count has shrunk 3.5% over a decade; the company is a serial buyer of its own stock and a steady dividend payer. TTM owner earnings of $6.7B against a $266 share price and ~$160B EV imply a 4.2% owner-earnings yield.

IV base of $409 vs. price of $266 = 0.65 P/IV, IV low of $275 - the stock is essentially trading at downside-case fair value. Reverse DCF implies just 4.3% growth, easily achievable for a regulated duopolist with pricing power and an active buyback. At this price, you are paying for the asset with optionality on the Norfolk Southern transcontinental merger thrown in free.

Moat

Union Pacific has one of the deepest moats in any U.S. business, built on three of the five Morningstar moat sources: irreplaceable cost advantage (Class I track), efficient scale (regulated duopoly per region), and intangibles (regulatory franchise built over 160 years).

Cost advantage / efficient scale. UNP's 32,889 route miles connect Pacific and Gulf Coast ports to Midwest gateways across the western two-thirds of the United States. The replacement cost of this network is essentially infinite: you cannot lay new long-haul track through populated America without two decades of eminent-domain fights, environmental review, and tribal/landowner negotiations. Every Class I network was built between roughly 1860 and 1900 and consolidated through a century of mergers - the last major restructuring was the BNSF formation in 1995 [1]. Buffett's framing is exactly right: 'huge investment in very long-lived, regulated assets... earning power that even under terrible economic conditions will far exceed its interest requirements' [6]. BNSF's interest coverage ran 9:1 in 2013 and 6:1 in a recessionary 2010 [2]; UNP runs comparable economics. Rail moves freight at roughly 500 ton-miles per gallon of diesel, four times more efficient than trucking [4] - a thermodynamic moat that no software disruption can erase.

Productivity is staggering and one-way. Class I railroads moved 655B ton-miles in 1947 with 1.35M employees; by 2014 the same industry moved 1.85T ton-miles - a 182% increase - with 187,000 employees, an 86% workforce reduction [1, latticework]. Inflation-adjusted ton-mile cost is down 55% over that span, saving shippers about $90B annually. UNP's 2025 operating ratio of 59.8% (a 10-bps improvement) reflects continuing extraction of productivity from the same fixed asset base. Workforce productivity, locomotive productivity, terminal dwell, train length, and fuel consumption were all at best-ever levels in 2025 per the 10-K.

Pricing power. UNP and BNSF compete with each other and with truck on intermodal lanes, but on captive bulk traffic - coal out of the Powder River Basin, grain out of Nebraska, chemicals out of the Texas Gulf - the railroad is the only economic option. The Surface Transportation Board provides a regulated rate ceiling but also a floor: the regulatory compact assumes the railroads earn their cost of capital so they keep investing [3, 6]. UNP has historically taken core pricing of 2-4% per year above inflation on most book of business.

Competitor stress test. Suppose a competitor had $10B and five years to take share from UNP in the western U.S. They cannot build parallel track. They cannot meaningfully take coal or grain share via truck (the unit economics are catastrophic: 1 train = 280 trucks, four times the fuel). They could build out short-line feeder rail, but short lines depend on Class I interchange. They could build pipelines for crude or chemicals, which has happened and has compressed crude-by-rail volumes - but that is a 20-year erosion, not a five-year disruption. The realistic answer is that the $10B is wasted.

Erosion risks (real). (1) Coal volumes are in secular decline as utilities retire plants - UNP's Powder River franchise was once the most profitable single book of business in U.S. railroading and is now a fading annuity. (2) Trucking technology - autonomous long-haul trucks - could shift the rail/truck breakeven distance and pull intermodal share. This is real but slow; even autonomous trucks burn 4x the fuel. (3) Regulation could turn unfavorable: re-regulation of rates, open-access mandates (forcing UNP to lease its track to competitors), or labor-driven margin compression. The 2022 PEB-250 forced wage settlement showed political risk is non-zero. (4) The Norfolk Southern merger introduces several years of integration risk and creates a transcontinental Class I that competitors and regulators may attack on antitrust grounds.

Buffett's own admission that BNSF lost share to UP in 2014 because of service problems despite spending more capex [3] shows the moat is contestable on execution; UP has historically been the better-run operator and is currently extending that lead (best-ever safety, best-ever operational metrics in 2025).

Moat verdict: WIDE

Management & Capital Allocation

UNP's capital allocation playbook is a textbook regulated-utility-with-some-pricing-power approach: spend roughly depreciation-plus on the network, run a high-50s operating ratio, return everything else to shareholders via dividends and buybacks, occasionally lever up to do so.

Reinvest in the network. The 2025 10-K reports $3.5B of capital invested 'to harden our infrastructure, modernize older locomotives, grow our business, improve service, and embed new technologies.' Specific projects called out: Kansas City, Inland Empire, and Lathrop intermodal terminals; Texas Gulf Coast manifest terminals; Pacific Northwest and Southwest main lines; modernized transportation planning systems. This is the right kind of capex - capacity expansion at chokepoints, not greenfield empire-building. UNP capex as a percentage of revenue is roughly 14-15%, structurally below BNSF's stated 16-17% [3] and well below BNSF's 26%-of-revenue super-cycle in 2015. UNP earns more per dollar invested than BNSF historically, which Buffett himself acknowledged in the 2014 letter when he said 'U.P.'s earnings beat ours by a record amount' [3].

Acquisitions. The big move is the announced Norfolk Southern merger (agreed July 28, 2025). This is genuinely strategic: a true U.S. transcontinental Class I would eliminate interchange friction at Chicago/Memphis/New Orleans and unlock single-line service from L.A. and Long Beach to East Coast ports. The economic prize is real. The risks are also real: the STB will scrutinize this for years, may impose service obligations, divestitures, or open-access concessions; the deal will distract management; the price paid (terms not fully detailed in this filing excerpt but expected to be a meaningful equity component) will dilute UNP holders; and integration of two large unionized rail networks is historically difficult (look at the multi-year struggles of the Canadian Pacific / Kansas City Southern combination). On balance, the asymmetric upside justifies the bet, but it shifts UNP's risk profile materially for the next 3-5 years.

Debt. UNP runs with meaningful long-term debt - the scorecard's net debt / EBITDA of -0.06 is a calculation artifact (likely a sign error or a cash-classification issue); the actual leverage is more like 2-2.5x EBITDA, which is normal for a Class I. Interest coverage is comfortable. Buffett's framing applies: the credit quality flows from the asset base, not the holding company. Debt is used to smooth share repurchases, not to fund speculative acquisitions.

Buybacks. Share count has fallen 3.46% over 10 years - modest pace, but at a stock that has compounded for shareholders, the dollars retired are real. The right test, per Buffett, is what price was paid relative to intrinsic value. With current P/IV of 0.65, today's buybacks are highly accretive. Historically, UNP has bought back stock at varying P/IV ratios - it has not been disciplined the way Berkshire is (Buffett famously refuses to buy above 1.2x book), but neither has it overpaid catastrophically.

Dividends. UNP pays a steady, growing dividend, currently around 2.0% yield. This is appropriate for a regulated cash-cow business; raising the payout ratio further would crowd out buybacks and reinvestment.

Communication. The 2025 letter has the right priority order - Safety, Service, Operational Excellence, Growth - and the metrics back it up: best-ever safety performance, intermodal SPI at 99%, manifest SPI at 100%. The CEO speaks like an operator, not a financial engineer. Median employee compensation is disclosed at $107,889 (2025), and the proxy gives shareholders adequate visibility into incentive design. Buffett's standard for railroad management is that they 'must think today of what the country will need far down the road' [3]; UNP's recent operational record meets that standard.

Concerns. (1) The Norfolk Southern deal introduces execution and price-paid risk on the largest decision a UNP CEO will ever make. (2) Pre-2020 management briefly chased Precision Scheduled Railroading too aggressively and damaged customer relationships - that history is recent enough to remember. (3) Regulatory and labor relations will dominate the next five years.

Capital allocator: B

Industry Structure

Threat of new entrants: NEGLIGIBLE. Class I railroads are the textbook example of an unrebuildable industry. The track was laid 1860-1920, consolidated to four players (UNP, BNSF, CSX, NSC) plus two Canadians, and is now legally and physically impossible to replicate. New track requires environmental review, eminent domain, and >$10M/mile construction cost; nobody has built a long-haul Class I from scratch in a century [4, 6].

Bargaining power of suppliers: MODERATE. Two key supplier categories: locomotive manufacturers (effectively a duopoly in Wabtec / GE Transportation legacy and Progress Rail / Caterpillar) and labor (12+ unions covering operating crafts, working under federal Railway Labor Act). Locomotive supply is concentrated but UNP's fleet is long-lived and the cost is amortized over decades. Labor is the more important risk - the 2022 federal intervention to prevent a national rail strike showed that wage settlements are increasingly subject to political pressure rather than pure negotiation, and that political pressure now favors workers more than it did 15 years ago.

Bargaining power of buyers: LOW for captive freight, HIGH for intermodal. For coal, grain, ethanol, plastics, and many chemicals, the railroad is the only economic option for a given lane - the shipper has no alternative. UNP's three commodity groups generated $23.2B of freight revenue in 2025; Bulk and Industrial together are mostly captive. Premium (intermodal + autos) competes directly with truck and is sensitive to fuel price and trucking capacity. Major intermodal customers (UPS, J.B. Hunt, Schneider, the steamship lines) negotiate hard, but they cannot move container volumes by truck without making the math collapse on long-haul lanes.

Threat of substitutes: MODERATE and slowly rising. Substitutes are (a) trucks, (b) pipelines, (c) coastal shipping/barge, (d) demand destruction (coal). Trucks have taken share where lane lengths are short and time-sensitive. Pipelines have permanently displaced crude-by-rail and some chemical volumes. Coal is in secular decline as the U.S. retires coal-fired generation - UNP's Powder River Basin franchise will continue to shrink. Autonomous trucking is a 10-20-year tail risk that could shift the rail/truck breakeven distance from ~750 miles down to ~500 miles, eroding intermodal share. None of these are existential; rail's 4-to-1 fuel efficiency advantage [4] is a thermodynamic floor.

Rivalry among competitors: LOW for captive, MODERATE for intermodal. The western U.S. is a duopoly: UNP and BNSF. They compete on overlapping intermodal lanes but each has captive franchise on its own bulk commodities. They do not compete on price for captive traffic - they compete on service quality, equipment availability, and capacity. Buffett's 2014 letter explicitly notes that BNSF lost share to UNP because of service problems [3] - rivalry happens, but at the operating-ratio margin, not in price wars. The pending NSC merger would consolidate the eastern duopoly with UNP, leaving effectively two transcontinentals (UNP/NSC and a possible BNSF/CSX response).

Value pool location and trajectory. Value pool sits with the Class I operator. Shippers capture the productivity benefit (55% real cost decline since 1947 [1, latticework]) while railroads earn 14-20% ROIC. Trajectory: stable to slightly improving. Pricing in line with PPI plus a spread, volume tied to industrial production and global trade flows, capex roughly equal to depreciation plus growth. The merger could shift more value pool to UNP if approved.

Industry Verdict: Excellent

Inversion (Bear Case)

I am now playing the bear. The bull thesis above is wrong, and here is why.

The single event that kills this. The Surface Transportation Board denies the Norfolk Southern merger after 18 months of regulatory review, or approves it with conditions so onerous (open-access mandates, divestitures, rate caps on captive shippers) that the deal is value-destructive. Management has staked enormous credibility on this combination; a denial wastes 18-24 months of executive attention, breaks up the deal premium already in the stock, and signals to Washington that further rail consolidation is dead - which means UNP must compete with whatever defensive merger BNSF/CSX produces. Alternatively: a single major derailment with hazmat release in a populated area triggers Congressional re-regulation. The 2023 East Palestine disaster on Norfolk Southern's network produced bipartisan rail-safety legislation; a UNP-network disaster post-merger would be politically toxic. The combined regulatory-compact-renegotiation risk is asymmetric: low probability per year, but cumulative over a decade, and the downside is permanent multiple compression.

Why the moat is narrower than bulls think. Bulls treat railroads as bulletproof regulated monopolies; the reality is more fragile. (1) The captive bulk franchise is shrinking. Coal volumes have fallen ~50% from peak and will continue to fall as utilities retire plants - UNP's Powder River Basin book was once the highest-margin lane in American railroading and is becoming a stranded asset. (2) Crude-by-rail collapsed when pipeline capacity caught up post-2015; ethylene, plastics, and other chemicals can move by pipeline too. (3) Autonomous trucking - even a partial deployment - moves the rail/truck breakeven distance and steals intermodal share at the margin. Tesla, Aurora, Kodiak, and the OEMs are spending real money on this; if even half the marketed productivity gains arrive by 2030, intermodal economics get worse. (4) UNP's pricing power is bounded by the STB's rate review process; in any decade where political winds shift, captive shippers (utilities, agricultural cooperatives) will lobby for re-regulation, and the political cost of letting them lose is much smaller than the cost of letting railroads lose. The combination of declining captive volumes (coal), substitutable lanes (chemicals to pipelines, intermodal to autonomous trucks), and a politicized rate ceiling means the moat is shrinking, not widening.

Why management is worse than it appears. Recent operational metrics are excellent because UNP is in the easy part of a rail cycle: post-PSR cleanup, modest volume growth, fuel cost tailwind. Watch what happens in the next downturn. More importantly, the Norfolk Southern deal is the single largest capital allocation decision in UNP's modern history, and the announced terms are likely to be at a meaningful premium. Two things go wrong: (a) the price paid is too high because there is no competing bidder and synergy estimates always overshoot, especially on labor synergies that the Railway Labor Act effectively forbids in the early years; (b) integration risk is enormous - merging two unionized Class I networks across IT systems, dispatch protocols, terminal procedures, and labor agreements has historically been disastrous (Conrail breakup, CP/KCS still working through it). Three years of distraction means UNP's standalone operating performance regresses while BNSF gains share. The bull case credits management for the merger's optionality; the bear case charges it as the largest agency-cost decision in 30 years.

What bulls are extrapolating that won't hold. (1) ROIC of 14% has been sustained through a benign decade of low rates and strong intermodal. The next decade may include sustained higher rates, weaker globalization (which is the opposite of intermodal-positive), and tougher labor settlements - all margin headwinds. (2) Operating ratio of 59.8% is at or near the structural floor; further improvement is unlikely without volume growth, which depends on macro variables UNP does not control. (3) Reverse DCF implied growth of 4.3% looks easy, but on a base of low single-digit volume growth and bounded pricing, it requires share gains or successful M&A integration - neither guaranteed. (4) Bulls anchor on Buffett's BNSF endorsement [2, 4, 6], but Berkshire bought BNSF in 2010 at meaningfully lower multiples and with a private holding period; the public-market UNP holder does not have those luxuries.

Valuation trap. The IV base of $409 and high of $518 assume a continuation of 14% ROIC and reasonable reinvestment. If terminal ROIC compresses to 10% (regulated-utility level) because rate review tightens or labor takes a larger share of value, intrinsic value collapses by 30-40%. The stock at $266 looks cheap against the optimistic IV but fair against a regulated-utility multiple of 12-14x earnings. P/E of 23.8 is at the 10-year average (23.86) - so the stock is not even cheap on its own historical record, only on a forward IV calculation that assumes the moat doesn't erode. In a re-regulation scenario, the multiple goes to 14x and the earnings go down 15% - that is $266 -> ~$140, a 47% drawdown.

If I am right, the stock could be worth $140-$180 within 3-5 years.

Lollapalooza Bias Check

Active biases I detect in myself analyzing UNP right now:

Authority bias (strong). Buffett owns BNSF outright. Buffett wrote glowingly about railroads in 2010, 2012, 2013, 2014, 2015, and 2016 letters [1-6]. The canon section of this brief is dominated by Buffett quotes endorsing the rail business model. I am pattern-matching to 'Buffett likes railroads, therefore UNP must be a buy.' But Buffett bought BNSF in 2010 at significantly lower multiples, in a private transaction with no exit pressure, and as part of a larger Berkshire conglomerate that smooths the volatility. None of those advantages apply to a public-market UNP holder. The endorsement is real but not directly transferable. I have to actively discount it.

Anchoring (strong). The brief gives me an IV base of $409 and a current price of $266. The 0.65 P/IV ratio is anchoring me on 'undervalued.' But the IV calculation embeds assumptions - sustainable ROIC, growth, discount rate - and any of those could be 10-20% wrong. If true IV is $320 instead of $409, the stock is fairly valued, not cheap. I should treat the IV range as a probability distribution, not a target.

Recency / confirmation bias (medium). UNP just reported best-ever 2025 operating metrics: best-ever safety, 100% manifest SPI, 99% intermodal SPI. I am extrapolating these into the future. But operational excellence cycles - PSR-era darlings became service disasters within four years. The current peak performance is the most likely point in the cycle to mean-revert.

Commitment / consistency (medium). Once I write 'WIDE moat' in step 5, I am psychologically committed to that conclusion through steps 6-12. The bear case has to fight my own narrative momentum. The mandatory inversion is designed to counteract this - I should weight it more heavily than I instinctively want to.

Social proof (medium). Railroads are a perennial value-investor favorite. Whitman, Klarman, Munger, Buffett, and various boutique value shops have owned them for decades. The concentration of smart money in this name pushes me toward consensus. But consensus value can be wrong: rails were value traps in the 1970s and 1980s, when productivity hadn't yet started compounding.

Deprival super-reaction (mild). The Norfolk Southern merger creates fear-of-missing-out on the transcontinental optionality. I am tempted to treat this as a free call option, but if the deal fails the stock falls and management credibility is damaged.

Incentive bias (mild). The pipeline produces results in JSON; the analyst is rewarded for a confident recommendation. There is institutional pressure toward 'Buy' or 'Hold' rather than 'Too Hard,' even when the honest answer is uncertainty about the merger outcome.

Balancing these: I should require a wider margin of safety than the IV-base / IV-low spread alone would suggest, and treat my WIDE-moat verdict as conditional on regulatory and merger outcomes that I cannot predict.

10-Year Outlook

Same fundamental business in 2036? Yes, with very high confidence. Trains will move bulk freight on UNP's track in 2036 with the same physics, the same right-of-way, and largely the same regulatory compact. The Class I structure will likely be more consolidated (UNP+NSC, BNSF+CSX), but the asset will be recognizable.

Customer base larger? Approximately the same to modestly larger. U.S. industrial production grows ~1-2% annually long-term; international trade flows are more volatile but trend up over decades. Coal will be a smaller piece, intermodal a larger piece, ag a stable piece. Net customer base in dollar terms grows with nominal GDP.

Profit per customer higher? Yes, modestly. Productivity flywheel - 55% real ton-mile cost decline since 1947 [1, latticework] - shows no sign of stopping. Train lengths still growing, locomotive productivity still improving, AI/automation in dispatch and PSR-style network design still has runway. Pricing should keep pace with PPI plus a small spread. Owner earnings should compound mid-single-digits before buybacks, high-single-digits per share after buybacks.

Moat wider? Probably modestly wider if NSC merger closes (true transcontinental scale advantage), modestly narrower if it fails (Stranded duopoly, growing autonomous-trucking competition). Net: directionally similar, with a fat-tailed upside on a successful merger.

Single biggest threat? Regulatory regime change. Either (a) STB blocks the merger and tightens rate review on captive shippers, or (b) a major hazmat incident triggers Congressional re-regulation that compresses ROIC structurally. Both would produce 30-50% intrinsic-value damage that no operational excellence can offset. Secondary threat: autonomous long-haul trucking achieving real intermodal share by 2032-2035.

The ten-year picture rests on the regulatory compact holding [3, 6]. That compact has held for a century and the alternative is unthinkable, so I rate the base case as solid. But the merger introduces a 3-5-year window of elevated regulatory and integration risk that I cannot handicap with confidence above medium.

CONFIDENCE: medium