HAL: Better Adjusted Margins, A Tougher Cash Test

The 60-second read

Five things to know

1. Two profit stories: adjusted operating income rose 7.0% sequentially to $731M; GAAP operating income fell to $624M.
2. International carries the mix: $3.543B international revenue offsets a $2.160B North American business that declined sequentially.
3. Cash requires a second definition: $368M operating cash less $266M capex is $102M; PP&E proceeds lift company free cash flow to $156M.
4. Recovery remains an outlook: management’s early 2027 view is mid-single-digit growth and about 15% operating margins, assuming flat North America.
5. Price disciplines the call: $34 base value versus the $32.98 pre-release close leaves little cushion; $27.20 would provide a 20% margin of safety only if the thesis holds.

Three scenarios, one probability-weighted number

Illustrative 12-month values; analyst assumptions, not forecasts or confidence intervals.

Signposts: what would change our mind

These are Charged Alpha decision thresholds, not company guidance. Check actual reported data at each date. The October 20 call was scheduled for 08:00 CDT and had not occurred at this research snapshot; no transcript or post-call guidance is implied.

Price and decision frame

Halliburton supplies the services, equipment and technology that customers use to develop and maintain oil and gas production. The investment question is not simply whether energy demand grows. Customers must authorize spending, Halliburton must deliver the work at acceptable margins, and invoices must become cash. Each stage can move at a different pace. This quarter illustrates that separation: adjusted profitability improved sequentially, while working capital absorbed cash.

Our reference is the $32.98 October 19 regular-session close, which predates the release. It is a valuation anchor, not a claim about how the market closed after these results. An earlier eligibility screen also captured extended-hours activity, but this packet uses one dated price consistently throughout its models. The earnings release arrived before the market opened on October 20; SEC filing acceptance does not establish the issuer’s exact publication minute. The scheduled conference call is a separate event.

The practical decision is HOLD with moderate conviction and high uncertainty. We can see a route to better earnings, but our central value already requires improvement. A useful business at a fair price does not automatically become an attractive purchase. We would want either stronger evidence of repeatable cash generation or a wider discount before upgrading the call.

What the quarter actually delivered

Revenue of $5.703B increased 1.8% from the year-ago quarter and was essentially flat sequentially. The composition was more important than the headline: Completion and Production contracted while Drilling and Evaluation expanded. A near-flat consolidated number can conceal a useful shift toward international drilling work, but it also makes it harder to claim a broad recovery across the whole portfolio.

Adjusted EPS of $0.60 compares with the saved FMP estimate of $0.584; revenue compares with $5.581B of provider consensus. Those are modest beats, computed at approximately 2.7% and 2.2%. The estimate basis is provider supplied, and adjusted EPS must not be compared with GAAP EPS as though the two were identical. The release is the authority for actual reported figures. FMP’s newest financial-statement row contains incomplete cost fields, so those zeros are not treated as zero operating costs.

Year-over-year GAAP EPS rose from $0.02 to $0.50, but the old period carried substantial charges and a tax valuation allowance. That spectacular percentage change is a poor shorthand for current business momentum. Adjusted operating income was actually below the prior-year level. The stronger claim is narrower: profitability improved sequentially on broadly unchanged sales. Whether that is the beginning of a durable trend depends on subsequent margins and collections.

Where the improvement came from

Completion and Production produced $3.089B of revenue and $457M of operating income. Both were down sequentially, primarily because of lower stimulation activity in the Western Hemisphere, partly offset by the Middle East. Its operating margin remained around 14.8%. This business still provides the largest share of revenue, so weaker pressure pumping can dilute progress elsewhere even when international operations are improving.

Drilling and Evaluation generated $2.614B of revenue and $391M of operating income. Higher activity in the Middle East and Latin America and more drilling-related work in the Western Hemisphere helped offset weaker international testing. Its operating profit rose more quickly than revenue, lifting its margin to roughly 15.0%. That is the clearest operational positive in the quarter, although a single quarter does not establish the sustainable margin on the next wave of contracts.

Geographically, North America revenue fell to $2.160B from $2.276B, while international revenue rose to $3.543B from $3.438B. Middle East/Asia was the principal sequential regional growth contributor. The release also discusses offshore contract awards, automation and digital work. These are useful indicators of competitive positioning, but contract announcements without disclosed revenue, timing and margin should not be converted into invented backlog or earnings forecasts. We value their potential through explicit scenarios rather than adding unsupported contract dollars.

Follow the adjustment bridge

The sequential decline in GAAP operating income does not contradict the increase in adjusted operating income. In Q2, $95M of net credits lifted reported profit above the adjusted measure. In Q3, $107M of net charges reduced it. That $202M swing more than offset the $48M improvement in adjusted operating profit, producing a $154M decline in GAAP operating income. Both presentations matter: the adjusted series helps compare operations, while the GAAP series retains the economic consequences of restructuring and investment decisions.

The Q3 reconciliation contains $107M of severance, $25M of inventory costs and write-downs, and $14M of fixed and other asset write-offs, partly offset by $36M of investment gains and $3M of other credits. It is a net basket, not simply one restructuring expense. Some cash effects may arrive on a different schedule from recognition. Repeated exclusions deserve scrutiny even when a particular item is unusual. Our valuation does not assume that every charge disappears permanently.

Below operating income, net interest and other expense reduce pretax income to $513M. Tax takes $94M, leaving $419M of consolidated income and $417M attributable to Halliburton after noncontrolling interests. The $42M SAP migration cost remains inside adjusted operating income. Adding it back again would create a more flattering metric than the company reported and would weaken comparability.

The same nine quality checks

Cash conversion was the main weakness. Q3 operating cash flow was below consolidated net income, and the gap becomes more consequential after capital expenditure. Working capital used $330M in the quarter. Receivables rose to $5.515B and inventory to $3.265B. Some of this can accompany international growth and normal seasonal timing, but that explanation needs confirmation through subsequent collections. Revenue growth does not pay for buybacks until customers pay.

The receivables chart uses quarter-end receivables divided by quarterly revenue, multiplied by ninety. The inventory series uses the same revenue denominator. These are transparent monitoring proxies, not formal average-balance DSO or cost-based inventory days. The current release does not disclose enough cost detail to compute a comparable Q3 inventory-days measure. Showing a clearly labeled proxy is more informative than treating an incomplete provider zero as a valid denominator.

Stock-based compensation is also not zero merely because the provider’s cash-flow row says zero. The audited annual report discloses $213M of stock-based compensation cost for 2025, after $223M in 2024 and $219M in 2023. We use that annual context and mark current-quarter SBC unavailable in the release. The scorecard is a disciplined set of questions, not an accounting-fraud verdict. Neither a small minority interest nor a favorable tax rate cancels weak cash conversion.

Cash, debt and shareholder returns

The company’s $156M quarterly free cash flow is a valid disclosed non-GAAP measure, but it includes $54M of proceeds from selling property, plant and equipment. Operating cash flow of $368M less $266M of capex produces $102M before those proceeds. We show both figures because asset disposals may be a recurring feature of fleet management, yet they are not identical to cash generated by the quarter’s operations after reinvestment. Neither definition should be substituted silently for the other.

Cash returns of $391M consisted of $250M of repurchases and $141M of dividends. They exceeded both quarterly free-cash-flow measures. The nine-month comparison is less extreme but still worth watching: OCF less capex was $772M, versus $984M of cash returns; company free cash flow was $947M. This does not by itself imply a financing crisis. It does mean the sustainability of returns should be judged across a full cycle, including working-capital reversals and reinvestment needs.

Cash ended at $1.843B against $7.162B of debt, including $90M classified as current. Our $5.319B net-debt measure subtracts cash only. We do not assume every investment, deferred tax asset or receivable is immediately distributable. Operating leases are economic obligations discussed separately, not debt included in this particular valuation convention. The same convention is held constant across scenarios. Cash falling while returns exceed internally generated cash would eventually require a more conservative capital-allocation stance.

Three routes, one disciplined call

The operating-profit route is the primary model. Our base assumes $23.5B of forward annual revenue and a 15% operating margin, generating $3.525B of operating income. A 9.5 times multiple gives $33.488B of enterprise value. Subtract $5.319B of net debt and divide by 832M diluted shares to obtain $33.86, rounded to $34. These are our assumptions. The margin is informed by management’s early outlook, but our revenue level, multiple, debt convention and probability weights are independent analytical choices.

The earnings route starts from the same $3.525B operating profit, deducts $336M of annualized net interest and $108M of other expense, then applies a normalized 22% tax rate. At eleven times resulting earnings per share, value is $31.77. The normalized tax rate is an assumption rather than the quarter’s reported rate. This route prevents a temporarily low tax rate from becoming a permanent valuation benefit. We retain other expense instead of assuming it vanishes.

The equity-cash route assumes $2.2B of normalized annual cash available after capex and values it at an 8% cash yield. Dividing $27.5B by 832M shares gives $33.05. Because operating cash flow already includes interest, this is an equity route and we do not subtract net debt a second time. The cash assumption exceeds the $1.600B trailing OCF-less-capex figure, so it requires recovery; it is not a disguised current yield.

All three routes land near the low-to-mid thirties. At the reference price, our model enterprise value of $32.758B implies about $3.448B of operating profit at the selected 9.5 times multiple, or a 14.7% margin on $23.5B of sales. The price therefore already anticipates much of the improvement. High uncertainty warrants a 20% margin of safety, producing a conditional $27.20 entry threshold. That is a review point, not an automatic purchase instruction if the business deteriorates.

Compare with the Street carefully

The saved provider target consensus is $43, well above our $34 base. That difference is useful because it identifies where optimism may already reside, but it is not a clean post-earnings disagreement with freshly revised analyst models. The newest saved target story is dated October 19, before the current release, and records Piper Sandler lowering its target to $42 from $43. Another saved entry records UBS at $42 in September. We preserve those dates instead of presenting old targets as reactions to today’s print.

The ratings endpoint contains sixty-four records, mostly Buy or Strong Buy. Treat that as the provider’s ratings snapshot rather than sixty-four unique analysts who have just reviewed this quarter. Target averages, ratings counts and annual-estimate analyst counts come from different datasets and need not describe the same population. No same-day revision wave is claimed.

FMP’s annual path shows approximately $22.350B revenue and $2.856B EBIT for 2026, then $23.563B revenue and $3.347B EBIT for 2027. The latter implies about a 14.2% EBIT margin. Management’s early operating-margin ambition is around 15%, but provider EBIT and company adjusted operating income may not share every adjustment. The comparison points to the debate; it does not prove a precise earnings upgrade. A better test will be a dated, internally consistent revision after analysts incorporate the release and call.

Management promises to grade

Management’s early 2027 outlook combines mid-single-digit revenue growth with operating margins of about 15%, an improvement described as roughly two hundred basis points. The company says that outlook assumes a flat North American market and does not require a hydraulic-fracturing recovery there. That framing matters: a North American rebound is upside to management’s thesis, not a prerequisite it explicitly built into this early view. It also places more responsibility on international execution, mix and internal cost discipline.

The release’s confidence about multi-year service demand reflects management’s assessment of energy disruption, inventory rebuilding and energy security. We treat that as an outlook, not a verified future order book. The timing of customer budgets, permits, supply chains and project execution can still delay conversion. International revenue can grow while collections lag, which is why margin and cash signposts must be evaluated together.

The $107M severance charge shows that the organization is taking action on its cost base. A charge is not evidence that all savings have already arrived. We need subsequent operating leverage and cash follow-through before assigning full credit. Similarly, repurchasing shares supports per-share economics only if the price paid is sensible and funding does not weaken the balance sheet. Our moderate conviction reflects an identifiable improvement path with several unfinished tests.

1. What portion of the expected margin improvement comes from cost savings versus price, utilization and mix?
2. How much restructuring cash remains to be paid, and on what timetable?
3. What collection profile should investors expect as international work expands?
4. How would buybacks change if full-year OCF less capex does not cover cash returns?
5. Which measurable operating conditions would make the early 2027 margin outlook unattainable?

The strongest cases on both sides

The bull case is coherent. International activity is improving, Drilling and Evaluation is showing better operating leverage, and a flat North American market could be enough for a larger earnings base if management executes. Automation and integrated contracts may improve efficiency and strengthen customer relationships. A North American recovery could then provide incremental upside. Our bull scenario gives this possibility a meaningful weight rather than dismissing it because current cash flow is weak.

The bear case is also grounded in the numbers. North American activity remains under pressure, adjusted operating profit is lower than a year earlier, and international expansion is consuming working capital. Repeated restructuring and investment adjustments complicate a simple normalized-profit story. If customers delay spending or payment, operating leverage can work in reverse while cash returns compete with reinvestment. A cyclical business deserves scrutiny when valuation assumes a recovery before cash confirms it.

Where we land is between those arguments. Halliburton does not need to be broken for its shares to offer only a modest expected return from this price. Conversely, one weak cash quarter does not erase its franchise. The key is to keep scenario uncertainty visible and size the price concession accordingly. Our risk rankings are qualitative judgments, not statistically estimated probabilities, and they can change when new evidence arrives.

What changes the decision next

The next useful catalyst is evidence, not simply a date on a calendar. The October 20 call can clarify the early 2027 margin bridge, the geographic demand outlook and the cash effects of restructuring. This packet was written from the release before that call, so answers cannot be presumed. A later update should distinguish genuinely new guidance from a repetition of the release’s broad ambition.

The next quarterly results will give a more complete full-year cash view and a chance to see whether receivables convert. Seasonal cash generation can make a single quarter look unusually weak or strong. That is why our capital-return test uses both the quarter and the full year. We will also review whether improved margins are accompanied by growing net debt or reduced reinvestment, since either could make the apparent recovery less attractive.

We would upgrade the investment case if profitable international growth converts into repeatable cash while the price leaves adequate room for uncertainty. We would reduce value if the margin path slips, collection risk rises, or sustaining capex proves higher than our cash model assumes. Price alone can improve prospective returns, but it cannot repair a broken operating thesis. The decision thresholds are designed to force a fresh review rather than encourage passive adherence to a stale target.

Data, methods and source limits

All dollar amounts in the financial tables are USD millions unless a per-share or billion-dollar label says otherwise. Quarterly income, segment, geographic, balance-sheet and cash-flow figures come from the saved SEC earnings exhibits. Three audited annual reports provide the longer operating and cash history, stock compensation and corporate-structure context. The raw source manifest stores exact URLs, retrieval times and hashes; the companion data document retains the primary table bodies as well as normalized values and computed fields.

The security is Halliburton common stock, not an ADR or a special voting-class proxy. The annual-report cover lists the common shares, and KPMG is the identified auditor. Ordinary director appointments were reviewed in the current eligibility screen. The consolidated statements can include controlled subsidiaries and variable-interest entities under ordinary consolidation rules; we do not infer a separate offshore VIE investment structure from that accounting wording. Material country and customer credit exposure remains a business risk even without such a structure.

Ratios use the denominator explicitly stated beside them. Quarterly receivable and inventory proxies use ending balances and ninety days. The model uses 832M diluted weighted-average shares for consistency, while the provider’s current outstanding-share figure is approximately 833.13M. This small difference explains why our modeled market capitalization need not equal the provider quote’s market capitalization exactly. Net debt is debt less cash; leases and less-liquid assets are discussed without silently changing the definition between scenarios.

The historical reaction table associates a report date with the first available regular close and the preceding close. It does not isolate earnings from market moves. The current report has no completed regular-session reaction at this snapshot. Annual consensus and targets are saved provider observations, not independently audited forecasts. Current-quarter cost detail and SBC are not fully available in the release; incomplete provider zeros are excluded rather than treated as actual results. A current comparable cost-based inventory-days measure, Beneish score and Altman score are therefore not presented.

This is a new exact-period packet. Known public HAL Q1 and Q2 videos are prior-period records and were preserved. We have not invented a prior packet signpost grade or claimed that an inaccessible historical model made today’s call. Future research should compare the signposts above with subsequent reported values and preserve this dated valuation alongside the outcome.