Delta Air Lines Inc DAL

Thesis

Delta Air Lines (DAL) is the highest-quality network carrier in North America. It runs a global hub-and-spoke system with structurally premium cabins (Delta One, Premium Select, Comfort+) that now generate roughly 60% of passenger revenue, plus a co-brand partnership with American Express that delivered ~$7B in 2024 and management targets ~$10B by 2027. This 'two businesses in one' model — a cyclical airline bolted to a steady-state royalty stream from Amex — is what differentiates DAL from peers like AAL and UAL.

The scorecard tells a sober story. Composite is 63/100 with profitability only 13/30 and valuation 14/30. The 10-year average ROIC is -3.88% — the GFC and COVID write-downs dominate the average — but the trailing 5-year ROIIC is 99% on a small denominator, suggesting incremental capital deployed since 2020 has earned well. Owner earnings TTM are $3.34B. Net debt / EBITDA is a manageable 0.97x and share count has shrunk 1.7% over a decade, with no meaningful dilution post-COVID.

The IV math is unforgiving. IV-base $67.38, IV-low $62.82, IV-high $98.72; current price $68.98 implies px/IV of 1.024 — fully valued vs base case. Reverse-DCF implied growth is -1.04%, meaning the market is pricing modest decline; that is not heroic, but neither is there margin of safety. To get a Buffett-grade 30% margin, I want DAL nearer $47 (mid between IV-low and a 25% MoS off base). At $69, this is a 'wait for the cycle' stock, not a 'back up the truck' compounder.

Moat

Delta's moat is real but narrow, concentrated in three of the five classic moat sources.

1. Brand / Premium positioning. Delta has spent fifteen years repositioning from a commodity carrier to the 'premium' US airline. Operationally this shows up in industry-leading on-time performance and completion factor — Delta routinely runs >99.5% completion vs ~98.5% for peers — and in customer NPS that is meaningfully above United and American. Pricing-wise, premium cabin revenue now exceeds main-cabin revenue and is growing 2-3x faster. This is the single most important moat element because main-cabin economy is and always will be a Bertrand-competitive commodity. Delta has effectively segmented its own seat map into a premium product (where it earns moat-like returns) and a commodity product (where it does not). [1]

2. Loyalty / Switching Costs (SkyMiles + Amex). SkyMiles paired with the American Express co-brand is the crown jewel. Amex pays Delta to acquire miles, which Amex then sells to cardholders. In 2024 Amex remuneration was ~$7B; Delta has guided to ~$10B by 2027. Importantly, this revenue stream has airline-like top-line but software-like economics — incremental margin on the marginal mile sold is very high, and it is largely insulated from oil prices, recessions, and labor costs. Network effect: status-laden frequent flyers are reluctant to start over at a competitor, so SkyMiles produces genuine switching costs for the ~30M most valuable customers. [2]

3. Network / Slot Economics. Delta's hubs at ATL, DTW, MSP, SLC, and JFK are fortress hubs — ATL is the world's busiest airport and Delta has ~80% share. JFK and LGA slots are scarce, government-allocated assets. Internationally, Delta's joint ventures with Air France-KLM, Virgin Atlantic, Korean Air, LATAM, and Aeromexico produce metal-neutral revenue sharing on premium long-haul routes. These assets cannot be replicated; you cannot build another ATL. [3]

4. Cost / Scale. Delta is NOT the low-cost producer. CASM (cost per available seat mile) is structurally higher than ULCCs like Spirit/Frontier and slightly above Southwest. Pilot contract (2023) put Delta's pilot costs ~30% above pre-pandemic levels and locked them in until 2027. Delta competes by earning a price premium, not by cutting cost. This is a defensible posture only as long as the premium gap holds. [4]

5. Regulatory / Intangibles. Slot controls at JFK/LGA/DCA, FAA pilot certification (1,500 hour rule limits supply), and bilateral air service agreements all act as soft barriers to entry. These have helped pricing recover post-COVID. However, regulators have blocked some JV expansions (JetBlue-American NEA) so the regulatory shield is two-edged. [5]

Why returns have still been mediocre. The 10-year average ROIC of -3.88% reflects two facts: (i) airlines have brutal operating leverage to fuel and demand shocks; one bad year wipes out three good ones, and 2020 was historically bad; (ii) capex is enormous and continuous — fleet modernization runs $5-6B/year and never stops. Even a wide moat in pricing struggles to overcome these structural drags.

Durability test. Will Delta's premium-and-loyalty moat be wider in 2035? Probably similar. Amex will continue paying for miles, ATL will remain a fortress, and demographic tailwinds (aging affluent travelers) favor premium. But the moat is narrow in the sense that one cycle of fuel + recession can erase years of FCF, and the business cannot meaningfully widen the moat through reinvestment because incremental seats are commodity.

Moat verdict: NARROW.

Management & Capital Allocation

Ed Bastian became CEO in 2016 after years as CFO/President under Richard Anderson. Glen Hauenstein (President) runs the revenue side; Dan Janki is CFO. The team is regarded as the most disciplined in the industry, and the capital-allocation record reflects that — within the constraints of being an airline CEO.

Capital allocation scorecard. The composite capital allocation sub-score is 20/30 — the strongest of the four sub-scores. Track record:

• Buybacks. Pre-COVID Delta retired roughly 25% of its share count between 2013 and 2019 at an average price in the high $40s — a brilliantly-timed program that compounded per-share owner earnings. Buybacks were correctly suspended in 2020 (Delta took CARES Act payroll support, which forbade them) and have only been cautiously resumed; share count is essentially flat post-COVID. 10-year share count change is -1.68%, modest because COVID interrupted the program.
• Dividends. Delta paid a dividend pre-COVID, suspended it, and reinstated it at $0.10/quarter in 2023, raised to $0.15. Token rather than meaningful — appropriate for a debt-laden cyclical.
• Debt management. This is where management deserves real credit. Net debt peaked at ~$22B in 2021; by 2025 it is ~$15B and net-debt/EBITDA has fallen to 0.97x — investment-grade territory. Delta is targeting BBB+ by 2027. Refinancing has been opportunistic; the SkyMiles bond (2020, 9% coupon) was an expensive but necessary survival move and management has been repaying it.
• Capex. $5-6B/year on fleet, primarily Airbus A321neo, A330neo, A350. Discipline shown by retiring older 757s and 767s rather than chasing growth. Notable refusal to participate in the 737 MAX boom — a contrarian but correct call given MAX issues.
• M&A. Equity stakes in foreign carriers (Aeromexico, Virgin Atlantic, China Eastern, LATAM) — mixed record. China Eastern was written down, LATAM still in workout. The strategy of buying influence rather than control is sensible for an industry with foreign-ownership caps but execution has been bumpy.
• The Boeing 717 fleet purchase from AirTran/Southwest at distressed prices (2013) is a textbook case of opportunistic asset acquisition.

Compensation. Bastian's pay is mostly equity-linked and tied to ROIC, on-time performance, and customer satisfaction. Pay was correctly cut during COVID. No egregious 'pay-for-pulse' signs.

Communication. Investor days are substantive. Management has been honest about what the business is — they describe themselves as 'high-quality industrial' rather than tech-style compounders. The $10B Amex target is specific and measurable; they have been transparent about the SkyMiles co-brand economics.

Mistakes acknowledged. The Monroe refinery (acquired 2012) has been a curiosity — sometimes a fuel-cost hedge, sometimes a drag — but management has stopped pretending it is strategic and runs it as a side asset.

Insider ownership. Low — Bastian's stake is meaningful in dollars (~$30M+) but tiny as a percentage. This is typical for a 60-year-old industrial; not a red flag but not the founder-level skin in the game I prefer.

Verdict. A capable, disciplined operator with a genuinely good buyback record pre-COVID and a credible debt-paydown program post-COVID. Constrained by the structural reality that airline CEOs cannot reinvest at high rates of return — the best they can do is not destroy capital, return cash, and stay solvent through cycles. By that bar, Bastian's team executes well.

Capital allocator: B.

Industry Structure

Porter's Five Forces — US legacy airline industry.

1. Rivalry among existing competitors — HIGH. The US has consolidated to four majors (Delta, United, American, Southwest) plus a handful of LCC/ULCCs (JetBlue, Alaska/Hawaiian, Spirit, Frontier, Allegiant). On most domestic city-pairs there are 2-4 competitors. Capacity decisions are made on long lead times (aircraft ordered 3-7 years ahead) and demand is volatile, which historically has produced destructive price wars. Post-2010 consolidation reduced rivalry meaningfully — the industry has been profitable in 9 of the last 14 years, a record without precedent in airline history. But 2020 reminded everyone that one demand shock erases years of profits. Rivalry is currently rational but the structure is fragile.

2. Threat of new entrants — MEDIUM-LOW. Capital intensity, slot scarcity at key airports, pilot supply constraints (1,500-hour rule), and the operational complexity of running a hub network are real barriers. However, ULCCs prove that point-to-point entry is feasible with leased aircraft and outsourced ground handling. Spirit's bankruptcy and Frontier's struggles show even the entrants have a hard time, which paradoxically protects incumbents. Bottom line: nobody is starting another Delta, but existing low-cost capacity can flow into and out of routes.

3. Bargaining power of suppliers — HIGH. This is the worst force for airlines. (a) Aircraft: duopoly Boeing/Airbus with 5-7 year backlogs; airlines take what they can get on whatever terms are offered. (b) Engines: GE/Pratt/Rolls-Royce — highly concentrated. (c) Fuel: commodity but volatile and uncontrollable; refining margins squeeze airlines in dislocations like 2022. (d) Labor: pilot unions extracted 30%+ pay raises in 2022-2023. ALPA, AFA, and IAM hold real pricing power. (e) Airports/ATC: monopoly suppliers of slots, gates, and air traffic services. Airlines are price-takers across virtually every input.

4. Bargaining power of buyers — MEDIUM-HIGH. For commodity main-cabin economy, buyers shop on price via Google Flights, Kayak, Expedia and switching cost is zero. For premium and corporate buyers, switching is harder — corporate travel managers negotiate annual contracts; high-status flyers don't want to start over at miles 0 elsewhere. Delta's strategy of premiumization is precisely an attempt to shift mix from high-buyer-power economy to low-buyer-power business/first.

5. Threat of substitutes — MEDIUM. Domestically, driving substitutes <500 miles, Amtrak in NEC, video conferencing for business meetings (a permanent post-COVID demand drag of perhaps 10-15% on business travel). Internationally, no real substitute exists. Long-term, high-speed rail in the US is unlikely to materialize meaningfully. Zoom is the most important substitute, and it has already mostly happened.

Industry economics. Long-term industry ROIC is roughly cost-of-capital — Buffett's 'death-trap' framing has empirical support. Periods of consolidation and discipline (2014-2019) produce attractive returns; periods of dislocation (2001, 2008, 2020) wipe them out. Cumulative industry profits since deregulation in 1978 are essentially zero in real terms. Delta's premium positioning is the least-bad place to be, but the gravitational pull of the industry is downward.

Industry Verdict: Poor.

Inversion (Bear Case)

The bear case for DAL — written without softening.

1. The 10-year ROIC tells the truth that one good cycle is hiding. The scorecard's most important number is ROIC 10y avg = -3.88%. This is not a typo, not a footnote — it is the actual return on invested capital that Delta has earned over a full cycle that includes both the post-GFC golden age of consolidated airline profitability AND COVID. If you cannot earn your cost of capital across a full cycle, you are not a compounder, you are a capital-cycle play. Buying at any price requires you to believe the next 10 years will look more like 2014-2019 than 2009-2024. That is a forecast, not an investment thesis. The composite score is 63 with profitability at 13/30 — the deterministic scorer is telling you exactly this.

2. The premium-cabin and Amex narratives are partially circular. Management and bulls cite premium revenue growth and Amex contribution as evidence of business-model transformation. But: (a) premium cabin growth has been industry-wide (UAL, AAL also up sharply), so this is a sector tailwind not Delta-specific moat widening; (b) the Amex contract was renegotiated in 2019 and is locked in until 2029 — the visible $10B target is the contractual ramp, not new business creation; (c) credit-card economics depend on Amex's premium-card franchise, which itself is sensitive to consumer discretionary spending and to interchange-fee regulation. The Durbin amendment did not touch Amex but a future Congress could, and the CFPB has signaled interest. If Amex remuneration plateaus or compresses, a meaningful slice of Delta's narrative valuation evaporates.

3. Pilot costs are a permanent step-change, not a cycle. The 2023 pilot contract gave 34% raises over four years and locked in industry-wide rates that United, American, and Southwest then matched. This added several billion to industry-wide annual costs permanently. CASM-ex-fuel is structurally higher than pre-COVID, and the next contract (2027) will not give it back. Combined with the FAA's 1,500-hour rule constraining supply, labor leverage is permanent.

4. Fleet capex is a treadmill, not an investment. The fleet is among the oldest among US legacies — average age ~15 years. Delta has avoided the 737 MAX (correctly) but that means the next decade's fleet renewal cycle is largely ahead, not behind. $5-6B/year of capex roughly equals owner earnings of $3.3B plus depreciation; in a normal year FCF after maintenance capex is modest. The reverse-DCF implied growth of -1.04% is the market's polite way of saying it doesn't believe the next decade produces meaningful growth in owner earnings.

5. The debt overhang is improved but not gone. Net debt is ~$15B with net debt/EBITDA at 0.97x — fine in a good year, terrifying in a bad one. EBITDA can fall 70% in a recession; that ratio could be 3-4x within twelve months of a real downturn. Investment grade is not yet locked in. A 2026-2027 recession would force a return to hoarding cash, halting buybacks, and possibly equity issuance at distressed prices — the 2020 SkyMiles bond at 9% is the warning of how that goes.

6. Substitution is only partly priced in. Business travel is structurally 10-15% lower than 2019 and is unlikely to fully recover. The most profitable passenger — the corporate road warrior on a flexible fare — has been partially replaced by Zoom forever. Delta has filled the seats with premium leisure, but premium leisure is more recession-sensitive than corporate.

7. The valuation is not cheap. At $68.98 vs IV-base $67.38, px/IV = 1.024. P/E TTM 12.29 looks low but this is a peak-cycle E. Adjusted for through-cycle earnings, the multiple is 18-20x — well above what an industry with -3.88% 10-year ROIC deserves. P/E 10y average is 22.4 only because it includes COVID's negative E (which math-explodes the multiple). A normal-EPS Delta deserves ~10x, putting fair value closer to $50-55.

Bear-case kill scenario. A 2026 recession + spike in fuel + capacity discipline cracking + Amex normalization. Owner earnings drop to $1.5B, multiple compresses to 9x, equity below $40.

If I am right, the stock could be worth $42 within 2-3 years.

Lollapalooza Bias Check

Biases the analyst (me) is exposed to right now.

1. Recency bias. Delta has just printed two strong years (2023, 2024). Premium cabins are full, Amex revenue is at record levels, debt is being paid down, and management is articulate. It is very easy to extrapolate the recent vibe and forget that this is the same industry where DAL went bankrupt in 2005. The base rate of airline ROIC over 50 years is approximately the cost of capital. Recency bias pushes me to weight the last 3 years more than the prior 47.

2. Narrative seduction (the 'two businesses in one' story). The story that Delta is 'really' a high-quality loyalty/Amex business with an airline attached is genuinely partly true and genuinely partly a marketing line. The bull narrative is exciting and clean; the bear narrative ('it's still an airline') is boring. Munger warned that if a story is too good, recheck the math. Premium + Amex is real, but it does not change that 70% of revenue is selling commodity seats with negative incremental ROIC across cycle.

3. Authority / social proof. Buffett bought airlines in 2016 and sold them in 2020. The fact that he sold is more informative than the fact that he bought, and yet bulls cite the 2016 buy more often. Munger called airlines a 'death trap.' If I am about to deviate from Munger on his home turf, the bar should be higher than 'the recent cycle has been good.'

4. Anchoring on the IV. The base IV of $67.38 makes the stock look fairly valued, which subtly invites a 'Hold' verdict — comfortable, defensible, lazy. The honest answer is that for cyclical commodity-adjacent businesses, IV is more uncertain than for true compounders, and 'fair value' on a deep cyclical at peak-cycle earnings is a place to be cautious, not neutral.

5. Liking / familiarity bias. Most analysts (and I) fly Delta and find it genuinely better than American. That positive customer experience leaks into the security analysis. The product is good; that does not mean the equity is.

6. Deprival super-reaction (FOMO). Travel is hot, the consumer is spending, and the stock has run from the lows. Missing further upside has emotional weight. The antidote is the reverse-DCF: at $69, the market already prices a return to mid-cycle, so the upside requires the cycle to extend further than consensus, AND for IV to expand. That's two bets, not one.

Net effect. Lollapalooza pulls toward 'Hold' or even 'Buy.' Inversion pulls toward 'Trim.' The honest middle is Hold with discipline — only get interested closer to IV-low.

10-Year Outlook

Will I be glad I owned DAL in 2036?

The honest answer requires separating two questions: (a) will the company still exist and be a leader, and (b) will its equity have compounded.

On (a): Yes, with high probability. Delta will be one of three or four global network carriers in 2036, ATL will still be the world's busiest airport, the SkyMiles-Amex partnership will still be paying nine or ten figures annually, and the brand will still be the premium pick among US legacies. The 10-year survival question is essentially settled.

On (b): Much less certain. The compounding math requires owner earnings to roughly double — from ~$3.3B today to ~$6-7B — at a multiple that is at least maintained. Doubling owner earnings requires either (i) the premium-cabin mix continuing to rise at recent pace for another decade (plausible but rate-of-change must slow as mix saturates), (ii) the Amex contract renegotiating favorably in 2029 (possible but counter-party Amex will negotiate hard), and (iii) no severe macro shock — no recession, no fuel crisis, no pandemic, no war disrupting key routes. Demanding all three is a lot.

The more likely 10-year outcome: owner earnings grow ~3-4% nominally, cyclically, with a couple of bad years; multiple stays around 10-12x normalized E; total return roughly tracks earnings growth + dividend, call it 6-8% annualized. That's a fine bond-substitute return, not a compounder return.

The deeper concern is that Delta cannot reinvest its earnings at high rates of return. Every retained dollar that goes into another A350 earns the cost of capital, not 20%. The only way DAL becomes a 'great' long-term compounder is if Amex revenue and ancillaries grow into a much larger share of the business — possible, but a structural shift, not a continuation of trend.

CONFIDENCE: medium