Halliburton Co HAL

Thesis

Halliburton is the world's #2 oilfield services firm, organized around two segments: Completion & Production (pressure pumping, frac, cementing, completions) and Drilling & Evaluation (directional drilling, fluids, wireline, software). Roughly 39% of LTM revenue is North America — almost entirely the US shale frac cycle — and 61% is international, where the company is currently riding a Middle East / Latin America capex wave.

The scorecard tells the honest story: composite 57, profitability 13/40, valuation 7/15. Ten-year average ROIC is -7.3%, distorted by the 2014-16 and 2020 down-cycles, when HAL took multi-billion writedowns. Five-year FCF conversion is 0.0% — every dollar of accounting earnings has been consumed by maintenance and growth capex. Net debt/EBITDA of 1.18x is fine in a $70 oil world and lethal in a $40 one, because revenue can fall 40% in twelve months (it did in 2015 and 2020).

Valuation: base IV $32.29, bull IV $47.52, current price $41.66 — P/IV = 1.29. The 5-yr ROIIC of 154% is real but reflects a cyclical recovery off a depressed base, not a sustainable reinvestment runway. Reverse-DCF implied growth is just 3.2%, which sounds modest until you remember the LT real growth rate of upstream service intensity is closer to 0-1%.

This is a pass. You buy HAL when WTI is $45 and everyone hates it, not at $41 in a Middle East-driven peak. Margin of safety becomes meaningful only below ~$28 (IV low).

Moat

Pricing power. HAL is a price-taker at the customer level. Operators (XOM, CVX, OXY, Aramco) award integrated project work via competitive bid. Frac pricing in the US Lower 48 is the canonical example: when rig count rolls over, pumping crews idle and HAL slashes per-stage pricing within a quarter. There is no Coca-Cola-style pricing. Verdict: NONE.

Switching costs. Modest and selective. In integrated international contracts (Aramco, ADNOC, Pemex), HAL embeds engineers, software (DecisionSpace), and proprietary chemistry into multi-year master service agreements. Rebidding is costly and risky for the operator, so incumbents get re-up. But the switching cost is calibrated against a small competitor set (SLB, BKR, regional NOCs' captives), and operators routinely split scope between two of the Big Three to preserve negotiating leverage. The Damodaran framework is instructive: switching costs only create value when the dominant supplier can extract rent. HAL cannot — Aramco audits its margins. [5] Verdict: NARROW in international integrated; NONE in NAM frac.

Network effects. None. Oilfield services is a one-to-one transaction; the value to operator A does not increase because operator B uses HAL.

Intangibles (brand, IP, regulatory). HAL owns thousands of patents in directional drilling, frac fluids, and reservoir software. The intangible that matters most is its 100-year operational reputation — the embedded knowledge of how to cement a wellbore in 14,000 ft of salt without losing the well. SLB has the same. Brand is a license to bid, not a license to extract excess returns. The 2010 Macondo episode (HAL was Deepwater Horizon's cement contractor) shows the regulatory intangible cuts both ways. Verdict: NARROW.

Cost advantages. This is where bulls argue HAL has a moat: scale in NAM (largest pressure-pumping fleet, vertically integrated sand and chemicals via the BJ legacy and the Multi-Chem business), and SAP/iCem manufacturing scale. The Damodaran scale-economies framework applies [5] — but the customer captures most of the surplus through bidding. The scale moat is real against a regional competitor (Pumpco, ProFrac) but irrelevant against SLB, which has equal or greater scale internationally. It is a cost parity with the #1 player, not a cost advantage. Verdict: NARROW vs Tier 2; NONE vs SLB.

Competitor stress test ($10B + 5 years). Could a hypothetical entrant with $10B and 5 years build a credible challenger? Mostly no — the labor, equipment, and operator certifications are too deep. But this misses the point: the threat is not entry, it is the structural overcapacity of the existing field. ProFrac, Liberty, NexTier (now Patterson-UTI), and a long tail of pressure-pumping pure-plays already have surplus horsepower. The frac market has been in a 'good enough capacity' state since 2015. The moat doesn't need to be breached because the rents aren't there.

Erosion risk. International is the bull case (60%+ of revenue, growing). But Middle East NOC contracts are increasingly being awarded to local-content-mandated competitors (Saudi-listed ADES, ADNOC Drilling). The cost-advantage moat erodes when the customer mandates that 60% of the work go to a local champion.

Buffett's framework would test: 'will this business reliably earn excess returns on incremental capital across a full cycle?' Ten-year ROIC of -7.3% is the answer. The moat exists only in the sense that HAL is one of three survivors of the 2014-2020 capacity rationalization. Survivorship is not a moat.

Moat verdict: NARROW (and shallow enough to be effectively NONE through-cycle).

Management & Capital Allocation

Jeff Miller has been CEO since 2017 and Chairman since 2019. He is a 30-year HAL lifer who came up through the cementing business — operationally credible, technically literate. The capital-allocation record across his tenure is mixed and must be evaluated through a cycle lens.

1) Reinvestment. HAL spent $1.5-1.6B/yr in capex through 2024-2025, roughly 6-7% of revenue. The company has consciously held capex below the 2014 peak ($3.3B) and refused to chase incremental NAM frac fleets in 2022-2023 when smaller competitors over-built. This is correct discipline, and it is the single most important capital-allocation decision in this industry. Grade on reinvestment: B+.

2) Acquisitions. The Baker Hughes deal that died in 2016 cost shareholders $3.5B in break fees — a self-inflicted wound from chasing scale rather than returns. Since then, M&A has been small and bolt-on (Athlon Solutions, Optime, Resoptima for AI reservoir software). Reasonable. Grade: C+ (legacy drag, current discipline OK).

3) Debt. Net debt/EBITDA 1.18x is conservative for this industry, down from 3x+ at the 2020 trough. Long-dated maturity ladder, mostly fixed-rate. Interest coverage isn't reported in the scorecard (null) but is comfortably above 8x at current EBITDA. Grade: A-.

4) Buybacks. This is where the story turns ugly. HAL bought back roughly $1B/yr of stock in 2023-2024 at average prices in the $32-40 range. The base IV is $32.29. So the company has been retiring stock at or modestly above intrinsic value — fine, not great. But share count change over 10 years is +0.28% — meaning a decade of buybacks has barely offset SBC dilution. The $3.5B in break fees plus the writedowns from the 2014 cycle mean shareholders have not enjoyed a meaningful per-share count reduction. Grade: C.

5) Dividends. Cut from $0.72 to $0.18 in 2020. Restored to $0.68 annualized currently. The cut was correct but tells you what 'commitment to the dividend' means in a commodity-services business: it doesn't. Grade: B.

Communication quality. Earnings calls are direct and operationally specific. Miller does not over-promise on NAM frac recoveries — he has consistently telegraphed the 2024-2025 NAM softness. International commentary is clear-eyed about Mexico/Pemex risk. The 10-K and proxy are standard quality, no obvious accounting tricks. Pay structure is roughly 80% performance-linked, with ROCE and FCF as primary metrics — appropriate.

Munger lens. The Buffett canon emphasizes managers who 'concentrate capital in a few high conviction ideas' and 'maintain discipline and let compounding unfold.' [2] Miller's discipline on NAM capacity is genuinely Buffett-flavored. But the underlying business cannot compound — the soil is too poor. As Buffett would say, 'when a manager with a reputation for brilliance tackles a business with a reputation for bad economics, it is the reputation of the business that remains intact.' Miller is good. The business is the problem.

Capital allocator: B. Above average for the industry, mediocre for an investable compounder. The grade is dragged up by current capacity discipline and dragged down by the legacy BHI break fee and a buyback program that has merely held share count flat.

Industry Structure

Oilfield services sits at the bottom of the energy value chain, and Porter's Five Forces explains why ROICs are structurally poor.

1) Bargaining power of buyers — VERY HIGH. The customer base is concentrated: the top 20 E&Ps and NOCs (Aramco, ADNOC, Petrobras, Pemex, ExxonMobil, Chevron, ConocoPhillips, Occidental, EOG, Devon) account for the majority of global upstream capex. These customers employ procurement organizations whose explicit mandate is to drive service costs lower year-over-year. Aramco's IKTVA program forces local-content procurement; US E&Ps run formal Dutch auctions. There is no Coca-Cola-style customer fragmentation that would protect supplier margins.

2) Bargaining power of suppliers — MODERATE. HAL buys steel, sand, guar gum, electronics, and labor. Sand prices spiked in 2017-2018 then collapsed; HAL responded by vertically integrating into in-basin sand mines. Labor (especially frac crews and offshore engineers) was a binding constraint in 2022 but has eased. Suppliers are not the problem.

3) Threat of new entrants — MODERATE in NAM, LOW internationally. US frac entry is cheap by industrial standards — a single fleet costs $50-80M and you can be operational in 9 months. This is why the 2017-2019 frac glut happened. Internationally, the entry barriers (operator certifications, Aramco prequalification, decades of safety record, $1B+ in working capital to fund 90-day receivables) are real. Net: the segment that matters most to HAL profitability (international) has good entry barriers; the segment that matters most to investor sentiment (NAM frac) has terrible ones.

4) Threat of substitutes — LOW now, MODERATE-RISING long-term. No one is going to substitute hydraulic fracturing in the next 10 years — it is the only economical way to produce US shale. But the broader substitute is electrification of end-use demand, which compresses oil demand growth toward zero post-2030 in IEA NZE scenarios. HAL's whole revenue pool is a derivative of upstream capex, which is a derivative of expected oil demand.

5) Competitive rivalry — VERY HIGH. SLB, HAL, and BKR form a stable Big Three globally, but each has a different segment mix and they compete fiercely for marquee international integrated contracts. In NAM frac, the rivalry includes ProFrac, Liberty, NexTier/PTEN, and dozens of regional players. The 2014-2020 cycle saw 60%+ price decline in pressure pumping. Capacity discipline is fragile — every up-cycle invites new fleets back into the market.

Value pool location. The lion's share of upstream value accrues to the resource owner (NOCs, mineral rights holders) and the integrated operator. Service companies harvest a thin, volatile slice — historically 6-10% operating margins through-cycle for HAL, vs 20%+ for the supermajors and 40%+ for low-cost NOCs. The value pool is migrating internationally (good for HAL) but also being squeezed by NOC local-content mandates (bad for HAL).

Trajectory. Cycle-up through ~2027 is plausible if Middle East capex sustains and US recovery occurs. But the secular trajectory is challenged: peak oil demand within 15 years is the consensus base case for IEA, BP, and Equinor.

Industry Verdict: Poor. A through-cycle ROIC of -7.3% is not an aberration — it is the structural output of an industry where buyers have full pricing power, capacity is easily added, and the underlying commodity is in long-term secular decline.

Inversion (Bear Case)

I am short HAL at $41.66. Here is why.

1) The single event that kills this. OPEC+ defends market share rather than price. The catalyst is plausibly a 2026-2027 Saudi-UAE coordination breakdown over quotas, or a Trump-administration push to crash oil to fight inflation, or a sustained Chinese demand disappointment as EV penetration accelerates past 50% of new-car sales. Any of these takes WTI from $70 to $50, and at $50 sustained for 18 months, the US frac market re-collapses. NAM revenue (37% of HAL) drops 30-40%, EBITDA margins compress 400bps, FCF turns negative, and the dividend is cut for the second time in seven years. Internationally, NOCs slow-walk capex acceleration. Stock falls to ~$22 (IV-low minus a recession discount).

2) Why the moat is narrower than bulls think. Bulls argue HAL has scale in pressure pumping and entrenched international integrated-services positions. Reality: in NAM, ProFrac (PFHC), Liberty Energy (LBRT), and Patterson-UTI/NexTier have collectively built fleet capacity that exceeds rational demand at any oil price below $80. Pricing discipline is a prisoner's dilemma that lasts exactly until one player blinks. Internationally, Aramco's IKTVA program and ADNOC's local-content mandate are systematically transferring share to ADES, ADNOC Drilling, and other state-influenced champions. SLB has spent ten years building digital and AI offerings (Delfi platform) that HAL is still trying to match (Resoptima acquisition was tiny). The moat that bulls cite is being eroded simultaneously by capacity discipline failure (NAM) and political localization (international).

3) Why management is worse than it appears. Jeff Miller is operationally credible and has avoided fresh capex mistakes — but the 2016 BHI deal that cost $3.5B in break fees happened on this management team's watch (Miller was COO). The 10-year share count is flat (+0.28%) despite roughly $8B of buyback spending across the decade — meaning shareholders have funded SBC dilution and approximately nothing else from buybacks. The 2020 dividend cut from $0.72 to $0.18 vaporized the income-investor base. Most damning: a -7.3% ten-year ROIC means this management team, across a full cycle, destroyed capital. Bulls focus on 2023-2024 operational execution and ignore the through-cycle math.

4) What bulls are extrapolating that won't hold. Three extrapolations: (a) Middle East capex super-cycle continues for a decade — implausible because Aramco capex is already at multi-decade highs and the marginal barrel from Saudi capacity additions does not need pressure pumping; (b) AI/data-center natural gas demand creates a new secular driver for North American gas drilling — true at the margin but small relative to the oil-driven NAM rig count; (c) the 5-year ROIIC of 154% is sustainable — false, this is recovery off a depressed 2020 base and will mean-revert toward the 10-year ROIC of -7.3% within three years. The reverse-DCF implied growth of 3.2% is barely below GDP, but in a peak-oil-demand world for the underlying commodity, even that may be optimistic.

5) Valuation trap (multiple compression / regime change). Current P/E TTM is 17.2x — well above the 10-year average of 13.9x. EV/owner-earnings is approximately 18-19x. Energy services has historically traded at 10-13x mid-cycle EBITDA and 12-15x earnings. The current multiple is pricing a continued international up-cycle. If the market accepts (a) peak oil demand by 2032, (b) NOC localization, and (c) reflexive cycle dynamics, the multiple compresses to 10-12x EPS, which on $2.40 mid-cycle EPS gives a $24-29 stock. The IV-low of $27.43 corroborates this. Importantly, the scorecard's note that 'maintenance capex is uncertain (>50% spread)' means the IV could be even lower if economic depreciation is closer to the high end.

Cumulative catalyst path. WTI rolls below $60 in 2026. NAM frac pricing drops 15%. International growth decelerates from 12% to 4%. Q3 2026 guidance miss. Multiple compresses from 17x to 11x. Dividend held but no growth. Buyback paused.

If I am right, the stock could be worth $22 within 18 months.

Lollapalooza Bias Check

Authority bias — active. HAL is a household name, the second-largest player in a Buffett-watched sector, and routinely covered by sell-side analysts with 'Buy' ratings citing 'best-in-class international platform.' The instinct is to defer to the consensus that this is a high-quality cyclical. The scorecard's 57 composite is consistent with that consensus — but the underlying components (profitability 13/40, valuation 7/15) tell a different story. I have to override the consensus framing.

Anchoring — active. I am anchored to the current $41.66 price and to the 2024 narrative that HAL is in the early innings of an international super-cycle. The decade ROIC of -7.3% should anchor me instead. When the through-cycle return on capital is negative, no reinvestment story creates value.

Recency bias — active. The most vivid recent data points are 2023-2024: strong international bookings, Aramco contract awards, AI/data-center natural-gas commentary, and a 5-year ROIIC of 154%. These are 30-month observations being implicitly extrapolated against an 80-year industry history of cyclical capital destruction. The historical base rate dominates the recent observation.

Confirmation bias — active in the opposite direction. Because I started this analysis already skeptical of oilfield services as a Buffett asset class, I am at risk of dismissing genuinely improved capital discipline at HAL (lower capex/revenue, real maintenance capex restraint, conservative leverage). I have tried to acknowledge in the management section that Miller's capacity discipline is the right instinct.

Social proof — partially active. Energy is back in vogue with value investors who cite Buffett's Occidental position as validation of the sector. But Buffett's OXY position is an upstream resource-owner stake (oil reserves), not an oilfield-service stake. The two are economically opposite: OXY benefits from low service costs; HAL benefits from high service costs. Conflating them is a category error.

Incentive bias. Sell-side analysts have an incentive to maintain coverage on the largest names in their sector, which biases ratings toward Buy/Hold. Management has an incentive to emphasize the international story over the NAM weakness because the international story sounds like a secular thesis rather than a cyclical one.

Deprival super-reaction — not active here. I do not own the stock, so there is no loss-aversion bias.

Net effect. The Lollapalooza of authority + anchoring + recency + social proof all push me toward a more positive conclusion than the math warrants. The discipline of writing a literal inversion section was the correction.

10-Year Outlook

Same fundamental business model in 2036? Mostly yes — operators will still need cementing, completion, drilling, and reservoir characterization services. The form factor (mix of NAM frac vs international integrated) will shift, but the basic service is durable.

Customer base larger? Probably smaller in absolute count. The trend is toward concentration: NOCs growing, public US E&Ps consolidating (Pioneer/XOM, Hess/CVX, Endeavor/Diamondback, Marathon/COP), and a long tail of private E&Ps being rolled up. Concentration shifts bargaining power further to the customer.

Profit per customer higher? Unlikely. Operator procurement organizations are getting more sophisticated, and AI-driven procurement (which HAL ironically helps customers build) further commoditizes service pricing.

Moat wider? No. NOC localization is eroding the international moat; US capacity discipline remains a prisoner's dilemma. The only moat that may widen is in software/AI reservoir characterization — but here SLB's Delfi platform is years ahead, and tech-native players (BPX, Earth AI, Hartree-Fock startups) are entering.

Single biggest threat. Peak oil demand. The IEA's 2026 World Energy Outlook places global oil demand peak in the early 2030s in the Stated Policies scenario and earlier in the Announced Pledges scenario. EV penetration in China is already past 50% of new-vehicle sales as of 2025. When global demand begins to decline, upstream capex compresses structurally — not cyclically — and the OFS sector enters secular contraction. HAL would survive (cash-generative through wind-down) but would not compound.

Could a smart 12-year-old explain why this earns excess returns in 2036? No — I cannot construct a credible narrative in which HAL's per-share intrinsic value in 2036 is meaningfully higher than today's $32 base IV after accounting for cycle position, dividends paid, and the secular demand trajectory.

CONFIDENCE: low.