Oneok Inc OKE

Thesis

ONEOK is a midstream toll-bridge: it gathers, processes, fractionates and transports natural gas, NGLs and (post-Magellan) refined products across the Bakken, Permian, Mid-Continent and Gulf Coast. Customers pay fees per barrel/cubic foot moved; commodity exposure is mostly indirect, through volume rather than price. The compounding case rests on three legs: (1) Bakken and Permian volume growth feeding hard-to-replicate fractionation at Mont Belvieu; (2) Magellan, EnLink and Medallion acquisitions adding refined-products and crude scope at synergy-friendly multiples; (3) a price-regulated FERC framework that protects long-haul tariffs.

The scorecard tells a more cautious story. ROIC 10y avg of 6.62% and ROIIC 5y of 7.9% are below most pipeline peers' cost of capital — the deals have grown EBITDA but not per-share intrinsic value at a cracking pace. FCF conversion is 43.1%, share count is up 11.95% over a decade (largely Magellan stock), and net debt/EBITDA sits at 4.97x — investment-grade but with no slack. The TTM P/E of 18.25x is a third of its 10-year average of 32.72x, reflecting market skepticism about the deal-financed growth model.

The IV math: base IV $109.70, low $69.84, high $143.71, current price $90.36 — px/IV = 0.82, an 18% discount to base. Reverse-DCF implies 11.47% growth, which is achievable but not conservative for a 4.97x-levered roll-up. With composite score of 60/100, this is a respectable but not exceptional name. Buy below the low IV of $70 (margin of safety becomes meaningful); trim above the high IV of $144. At $90 it's a Hold — interesting yield, but not Buffett's fat pitch.

Moat

ONEOK's moat is overwhelmingly a cost advantage rooted in irreplaceable physical infrastructure, lightly seasoned with switching costs. The other three classic moat types (pricing power above inflation, network effects in the Metcalfe sense, intangibles like brand) are either weak or absent.

1. Cost advantages — narrow-to-wide depending on basin. Buffett's writeups of MidAmerican's pipelines [1][4][6] are the right frame: 'Kern River and Northern Natural pipelines, which carry about 9% of the natural gas consumed in the U.S.' [4]. Long-haul pipes are economic monopolies on routes — once steel is in the ground and rights-of-way are negotiated, a competitor faces a $10B+ greenfield bill plus 5-10 years of FERC and environmental permitting to duplicate the route. ONEOK's NGL pipeline system from the Mid-Continent and Bakken to Mont Belvieu is the same kind of asset. The Mont Belvieu fractionation complex specifically — where ethane, propane, butane, isobutane and natural gasoline are separated from a Y-grade NGL stream — is one of only a handful of such facilities in the world, and ONEOK is one of three large players. Replicating it would cost billions and would face enormous community opposition. The Magellan acquisition added the largest refined-products pipeline system in the US (~9,800 miles) — another network Buffett would recognize as 'far-sighted' infrastructure [3].

Competitor stress test ($10B + 5 years): A new entrant with $10B and 5 years could not build a competing Bakken-to-Conway-to-Mont Belvieu NGL system. They could perhaps build one parallel segment in the Permian (where pipeline build-out is ongoing — see Plains, Energy Transfer, Enterprise) and that's where ONEOK's moat is narrowest. The Bakken NGL system, by contrast, is genuinely fortress-like; ONEOK is the dominant gatherer there because of legacy Williams asset roots.

2. Switching costs — narrow. Producers sign long-term gathering and processing contracts (often 10-20 years, sometimes life-of-lease dedications). Once a wellhead is connected to ONEOK's gathering system, switching means re-laying gathering pipe — uneconomic for individual wells. But producers can route new drilling to competitor systems, and shippers on long-haul lines can renegotiate at contract end. The switching cost protects existing volumes, not new ones.

3. Pricing power — narrow. Most long-haul tariffs are FERC-regulated under cost-of-service (which caps return) or indexed to PPI plus a small adder. Buffett's note about MidAmerican's regulators allowing 'an appropriate return on the huge amounts of capital we must deploy' [3] applies — but the same regime caps upside. Gathering and processing fees are not FERC-regulated and have more pricing flexibility, but they're competitive within a basin.

4. Network effects — minimal. Pipelines are point-to-point assets. Adding one shipper does not increase value for the next shipper the way an additional Visa cardholder does. There is a mild positive: a denser gathering grid reduces marginal-mile capex per new well, which is a scale economy, not a network effect.

5. Intangibles — none meaningful. No brand, no patents, no regulatory licenses that competitors lack (FERC certificates are obtainable; environmental permits are the actual barrier).

Erosion risks: (a) Permian dominance shifts — Energy Transfer, Enterprise and Targa each have competing NGL takeaway projects. (b) Energy transition & demand — long-dated NGL/refined-products demand is the single biggest 10-year question; if US LPG export demand softens or ethane crackers move offshore, fractionation economics compress. (c) Acquisition integration risk — the Magellan refined-products business is structurally lower-growth than NGLs, and the company is still proving the synergy thesis. (d) Re-contracting risk — many high-margin legacy contracts roll off and are renegotiated at lower rates as basins mature.

The scorecard's 6.62% ten-year ROIC and 7.9% ROIIC are the moat's most damning evidence: a true wide moat would compound capital at 12%+. ONEOK earns barely above its cost of capital, and the marginal dollar reinvested earns less than the average dollar already in the ground. That is the signature of a narrow, not wide, moat — albeit a durable one.

Moat verdict: NARROW

Management & Capital Allocation

Pierce Norton (CEO since 2021) inherited a Williams-spinoff legacy and has executed three of the largest midstream M&A deals of the decade: Magellan Midstream ($18.8B, closed Sept 2023), EnLink Midstream public-unit roll-up plus Medallion (combined ~$5.9B, closed in 2024-2025), and a Delaware Basin JV. The 12% increase in share count over the past decade is almost entirely Magellan stock issuance — the EnLink/Medallion deals have been mostly cash. Let's grade against Buffett's five capital-allocation choices.

Reinvest in operations. ONEOK reinvests heavily — annual capex has run $2-4B, with about half maintenance and half growth. ROIIC of 7.9% over five years is the verdict: the marginal dollar earns roughly its cost of capital. Compare with Buffett's MidAmerican framing — 'Last year MidAmerican retained more dollars of earnings — by far — than any other American electric utility' [2] — there, regulators allow a fair return; here, the issue is that growth capex chases volume, and volume's marginal economics in a maturing basin are mediocre. Grade: C.

Acquisitions. Three large ones in three years. Magellan was bought for ~10x EBITDA, before synergies — a fair-not-cheap price for an irreplaceable refined-products network. EnLink/Medallion buy-in unified the GP/LP structure, eliminating IDR drag, which is a textbook value-additive move. The pace, however, is concerning: roll-ups that compound by acquisition rarely earn ROIC above peer averages, because they eventually run out of cheap targets. Public statements emphasize 'synergies' and '$1B run-rate by year three' — these have a track record of partial delivery in midstream. Grade: B-.

Debt. Net debt/EBITDA of 4.97x is at the upper end of investment-grade midstream — Magellan was famously low-leverage (~3x) before the deal, so OKE's approach has been to lever Magellan's clean balance sheet. The good news: management has publicly targeted 3.5x and is paying down. The bad news: at 4.97x with rates higher than during the deal, refinancing risk and reduced dry powder for the next downturn are real. Grade: C.

Buybacks. Modest. OKE issued shares for Magellan (the share count is up 11.95% over a decade — most of that in 2023). Buybacks have been opportunistic but small. There is no indication management has been disciplined about average P/IV when buying — the historical pattern is dividend-prioritized return of capital, not buybacks. With shares trading at 0.82x base IV, this is the moment a great capital allocator would buy back stock aggressively; we are not seeing that. Grade: C.

Dividends. OKE pays a high (~5%) dividend yield with a multi-decade record of either holding or increasing. The dividend has been the cornerstone of the equity story, and the Magellan deal was specifically pitched as accretive to 'distributable cash flow per share' from year one. This is good capital discipline — paying out cash you cannot redeploy at high returns is correct. Grade: A-.

Communication quality. Investor materials are clear; the segment EBITDA bridge is well-disclosed; goodwill and acquisition accounting (visible in 10-K XBRL tags for EnLinkAcquisitionMember, MagellanAcquisitionMember, MedallionAcquisitionMember, DelawareBasinJVAcquisitionMember) is broken out properly. The framing is promotional in the standard MLP/midstream way (focus on adjusted EBITDA, not GAAP, and on DCF/share, not FCF/share) — Buffett would prefer they show owner earnings. The 43.1% FCF conversion is notably lower than the 'distributable cash flow' figures management emphasizes — read both numbers. Grade: B.

Synthesis. This is competent, deal-driven management running a mature business. They're not destroying capital, they're not hiding things, but they are not the disciplined per-share value compounders Buffett seeks ('Berkshire couldn't have better partners' is decidedly not how I'd describe Norton). The dividend and the deal pace are the two things they do best; ROIIC and the share-count math are the two things to watch.

Capital allocator: B-

Industry Structure

Porter's Five Forces on US midstream NGL and natural gas pipelines:

1. Threat of new entrants — LOW. Pipelines are the textbook example. Buffett: 'Kern River, extends from Southwest Wyoming to Southern California… is undergoing a $1.2 billion expansion that will double throughput by this fall' [1] — once the line is in the ground, no one builds a competing one. FERC certification, environmental review, eminent domain coordination, and capital intensity (often $5-10M/mile for large-diameter pipe) push payback periods to 10+ years and raise the bar for new entrants. The Keystone XL cancellation is the cautionary tale: even with capital, political risk can kill a project. Threat is low and getting lower as anti-pipeline activism intensifies — which protects incumbents.

2. Bargaining power of suppliers — MODERATE. ONEOK's 'suppliers' are E&P producers in the Bakken, Permian and Mid-Continent. In a single basin, a few large producers (ExxonMobil/Pioneer, ConocoPhillips, Chevron, Continental, Hess) can negotiate. They typically sign long-term dedications in exchange for capacity guarantees. When commodity prices crash, producers slow drilling and ONEOK's volumes drop — that's the cyclical bite, not pricing pressure per se. In a strong basin growth phase, suppliers compete to get takeaway capacity, flipping the leverage.

3. Bargaining power of buyers — MODERATE. Downstream buyers are refiners, petrochemical crackers, LPG export terminals, and end-use utilities. Mont Belvieu fractionation customers (Dow, Shell, ExxonMobil) are large but have limited alternatives — there are only a few US fractionation hubs. Refined-products shippers on the Magellan system are Marathon, Phillips 66, Valero — large, sophisticated, but committed to the network because trucking refined fuels is uneconomic. Pricing for FERC-regulated long-haul is set by tariff, limiting buyer leverage.

4. Threat of substitutes — MEDIUM, RISING. This is the most important force. Substitutes for natural gas: renewables + storage for power generation, heat pumps for residential heating. Substitutes for NGLs: bio-based feedstocks, electrification, recycled plastics policy. Substitutes for refined products: EVs (the long horizon), biofuels (near term). The 10-year demand outlook for NGLs is more robust than for refined gasoline (US LPG exports to Asia are still growing; ethylene demand for plastics still grows globally), but the 25-year outlook is less clear. This force is the asymmetric risk in the OKE thesis.

5. Industry rivalry — MODERATE. Within a basin, midstream is roughly oligopolistic: ONEOK, Energy Transfer, Enterprise Products Partners, Targa, Williams, Plains. Returns above cost of capital are competed away when too many parallel pipes get built (the 2014-2016 Permian over-build is the historical example). Rivalry intensifies during basin-investment booms; it eases during droughts. Currently, with capital discipline and consolidation (EnLink, Magellan, Crestwood, etc.), rivalry is moderate-and-easing.

Value pool location and trajectory. The midstream value pool sits between the producer (E&P) and the end consumer (refiner, petrochemical, exporter). It captures economic rent from the geographic separation. The pool grew from ~2010-2022 with shale, has roughly stabilized, and faces secular pressure from energy transition over a 20-year horizon. The smart money in midstream now is in basins with the longest reserve life (Permian) and in molecules with the most resilient demand (NGLs/petrochemicals, LNG export, refined products in EM). ONEOK is well-positioned in the Bakken (legacy fortress), middling in Permian (catch-up), and well-positioned in NGLs and refined products.

Industry Verdict: Good — high barriers to entry, regulated returns floor, but secular substitution risk and a maturing basin profile cap upside. Not 'Excellent' (which I'd reserve for railroads, where substitutes don't exist).

Inversion (Bear Case)

I am now the short-seller. The bull thesis on ONEOK rests on the premise that midstream is a regulated toll-road that compounds at high single digits in perpetuity. Here is why that's wrong.

1. The single event that kills this — credit market regime shift. ONEOK runs at 4.97x net debt/EBITDA. Investment grade, but barely. The midstream sector carries roughly $400B of debt collectively. If long rates re-set 200bps higher and stay there — a regime, not a cycle — the cost of refinancing $30B+ of OKE debt over the next decade compounds into a meaningful EBITDA haircut. Worse: if a 2008-style credit event hits and investment-grade spreads blow out, OKE could see covenant pressure exactly when EBITDA is also under pressure (recession + wider spreads + commodity crash is the trifecta). The dividend would be the first thing to go. A dividend cut on a yield-anchored stock is a 30-50% drawdown event, fast. This isn't a tail risk — it's exactly what happened to Kinder Morgan in 2015. KMI cut the dividend 75% and the stock halved.

2. Why the moat is narrower than bulls think. The bull narrative says 'irreplaceable infrastructure'. The reality: only the long-haul backbone is irreplaceable. Gathering systems are very replaceable — competitors lay parallel pipe routinely in Permian and Eagle Ford. Processing capacity is fungible — you can build a new fractionator in 2-3 years. ONEOK's specific Bakken position is fortress-like, but the Bakken is a maturing basin with declining rigs; that fortress protects a shrinking volume base. Permian, where the volumes are growing, is where ONEOK is weakest, in fierce competition with Energy Transfer, Enterprise, Targa, and Plains. The 6.62% ROIC over a decade is the actual signal: this is not a wide-moat business. It's a mediocre-moat business that includes one fortress (Bakken) and a lot of competitive sand.

3. Why management is worse than it appears. Pierce Norton has done three large deals in three years. The historical base rate for transformative midstream M&A delivering promised synergies is poor — Energy Transfer's history is the cautionary tale, and even Williams's deals have a mixed record. Magellan synergies were pitched at $400M run-rate in year three; the actual delivery is opaque because the segment reporting was reshuffled, which is exactly what happens when the synergy story is missing the target. EnLink/Medallion add complexity at a time when the company hasn't fully digested Magellan. The 11.95% share-count growth is the tell: management has been a seller of stock at low multiples to fund deals, which is the opposite of what disciplined capital allocators do. They have not bought back meaningful stock at 0.82x IV — exactly when they should. This is empire-building under a yield-discipline veneer.

4. What bulls are extrapolating that won't hold. (a) NGL volume growth in the Bakken — basin is past peak rig count; ROW pipeline volumes are flat-to-declining absent a large oil-price spike. (b) Mont Belvieu fractionation premium — new capacity coming online from Enterprise and Targa will compress fractionation spreads. (c) Magellan refined-products demand — refined-products demand in the US is structurally flat-to-declining; EVs at 8% of new sales (and growing) will eventually bend the gasoline-mile curve. (d) Permian catch-up — bulls assume OKE wins meaningful Permian share via EnLink/Medallion. EnLink's Permian assets are subscale relative to the dominant trio (ET, EPD, TRGP). (e) Long-rate regime — bulls assume rates revert. They may not.

5. Valuation trap. The TTM P/E of 18.25x looks cheap vs. the 10y average of 32.72x. But the 32.72x average includes the ZIRP yield-bubble period (2015-2021) when midstream MLPs traded at absurd cap rates. The normalized fair multiple for a 4.97x-levered, 6.6% ROIC business is closer to 12-14x earnings, not 18x. So OKE could be a value trap: cheap vs. a bubble multiple, fair vs. a normal multiple, and expensive vs. a stressed multiple. EV/FCF of 28.5x is the more honest metric — that is expensive for any business not compounding at 15%+. Reverse-DCF implied growth of 11.47% looks plausible but bakes in successful Magellan/EnLink synergy delivery; if synergies disappoint, implied growth drops to 6-7% and the stock is fairly valued at ~$70.

Multiple compression scenario: 12x earnings × $5/share owner earnings = $60. Dividend cut scenario: 10x earnings × $4 = $40, plus the market's re-rating of the perceived dividend safety across the sector — call it $35-45.

If I am right, the stock could be worth $50 within 3 years.

Lollapalooza Bias Check

Biases active in me as the analyst right now:

Authority bias. I am leaning on Buffett's MidAmerican writeups [1][3][4][6] to validate a 'pipelines are good businesses' frame. But Buffett owns regulated electric utilities with NaturalNorthernGas as one piece of a diversified portfolio inside Berkshire's tax-advantaged structure. That is a different beast from a pure-play midstream C-corp with 4.97x leverage funding deals with stock. The authority transfer is misleading. Pipelines that Berkshire owns enjoy MidAmerican's retained-earnings discipline ('MidAmerican has not paid a dividend since Berkshire bought into the company in early 2000' [5]); ONEOK has the opposite policy. I should not assume the moat-quality maps cleanly.

Anchoring bias. I am anchoring on the IV base of $109.70 vs. the price of $90.36 and concluding 'cheap'. But the 32.72x ten-year average P/E that helps justify base IV includes the 2015-2021 yield-bubble. If I anchor on a more conservative P/E of 14x normalized owner earnings, IV drops below the current price. The $109.70 is a Goldilocks number; the range $69.84-$143.71 is honest and very wide for a reason — the maintenance capex spread is >50% per the scorer notes.

Recency bias. Magellan synergies, EnLink integration, and a strong 2024-2025 NGL volume year color my read of management. Two years from now we'll see the integration's actual outcome; midstream M&A regularly looks great in years 1-2 and disappoints in years 3-5. Williams's history of Williams Partners and the eventual $11B impairment is the cautionary record.

Incentive-caused bias (in management, that I might miss). Management is paid on adjusted EBITDA growth and DCF/share, not on FCF/share or per-share IV growth. They are compensated to do dilutive deals that grow the EBITDA headline. I might be giving the deal pace too much benefit of the doubt because the press releases are well-written.

Confirmation bias. I started with 'midstream toll-road, regulated, durable' and kept finding evidence for it. But the 6.62% ROIC and 7.9% ROIIC are screaming 'mediocre business' and I had to actively restrain the urge to explain them away. Munger: 'the best businesses by far for owners continue to be those that have high returns on capital and that require little incremental investment to grow' [3]. ONEOK fails both halves of that test. The honest read is: this is a fair business at a fair price, not a great business at a fair price. I should not let the interesting moat stories override what the capital-return numbers are telling me.

Social proof. Berkshire owns OXY (oil), Chevron (oil), and pipelines through MidAmerican. There's a 'Buffett likes energy infrastructure' halo. He likes his energy infrastructure. He has not, to my knowledge, been a public buyer of OKE.

Net effect: my pre-corrected output would have been a Buy. Post-correction, it's a Hold leaning to 'Buy below $70'.

10-Year Outlook

Same fundamental business model in 2036? Mostly yes. ONEOK will still be moving NGLs, natural gas, and refined products through pipes. Volumes might be lower (energy transition), might be higher (LNG and LPG export tailwinds, AI-driven gas-fired power demand). The toll-road economics will remain.

Customer base larger? Probably similar in count, possibly smaller in volume. Bakken producers will be fewer (consolidation continues); Permian producers will be more concentrated (Pioneer-into-XOM is the pattern). Petrochemical buyers and LPG exporters are likely larger and more international. Refined-products customers (refiners) are likely fewer.

Profit per customer higher? Uncertain. FERC tariffs index roughly to inflation. Gathering and processing margins compress as basins mature. Synergies from Magellan/EnLink could improve unit economics if executed. Net: probably flat-to-slightly-up in real terms.

Moat wider? Probably narrower. Permian competitive intensity will continue. Mont Belvieu sees new fractionation supply. Gathering systems get parallel competition. The Bakken fortress holds, but it's a smaller share of a maturing basin. Magellan refined-products is structurally more stable, less moat erosion there.

Single biggest threat over 10 years. A 2008-style credit event combined with a commodity price crash. ONEOK at 4.97x net debt/EBITDA cannot absorb both at once without a dividend cut. Secondary threat: faster-than-expected EV adoption denting refined-products demand to the point where Magellan tariffs flatline. Tertiary threat: a serious pipeline accident that triggers a regulatory clamp-down on midstream (Mariner East, Colonial post-2021 cyberattack patterns).

Confidence assessment. The business will look broadly similar in 10 years. The price/per-share IV math is less clear. I have medium confidence that this remains a profitable, dividend-paying business; medium confidence that it will compound per-share IV at 7-9% over 10 years; lower confidence about whether the current management's deal pace continues to add or subtract value at the margin. The wide IV range ($69.84-$143.71) and the 'maintenance capex uncertain (>50% spread)' scorer note are honest acknowledgments that I cannot tightly bound the future.

CONFIDENCE: medium