LIN | Earnings Review — Q2 2026

The composition of this beat is the story. Linde delivered the largest revenue beat in at least twelve quarters (+3.03%, $9,289M vs $9,015.9M) and converted it into the thinnest EPS beat in at least twelve quarters (+0.22%, $4.50 vs $4.49). Two years ago LIN beat EPS by ~2% while missing revenue by ~1.5% — pure margin conversion. Today it beats revenue by 3% and EPS by a penny. The conversion engine has stalled.

Key metrics. Sales of $9,289M rose +9.3% YoY, the sixth consecutive quarterly acceleration off the 4Q24 trough of -0.2% and the fastest print in at least ten quarters. The driver is volume, not price: price has been a metronomic +2 pts for ten straight quarters, while volume went -1 → 0 → +1 → +2 pts, a two-year high. Every region is growing; APAC leads at +13.0% YoY. The sale-of-gas backlog broke open to a record $8.1B (+$1.0B QoQ) on a single advanced-node electronics win in the Western U.S. But adjusted operating margin fell to 29.5%, -55bps YoY (company-stated -60bps, -30bps ex cost pass-through), adjusted EBITDA margin -99bps — the worst in the six quarters of YoY data shown — and gross margin -164bps. All four core segments compressed simultaneously.

New guidance. FY26 adjusted EPS moved to $17.70–$17.90 (floor +$0.10, ceiling untouched for a third consecutive guide); 3Q26 to $4.45–$4.55 (mid $4.50 vs street $4.54). The headline reads as a guide-down, and it is not: consensus FY26 of $17.91 sits one cent ABOVE the top of the company's own range. The entire $0.05 FY raise is exactly the 2Q26 beat — implied 2H26 EPS is $8.97 both before and after the print, to the penny. Not one cent of back-half assumption changed, by management's own words and by the arithmetic.

Tone. Two vectors moved in opposite directions. Demand confidence is the best in eight quarters ("we have lapped the more difficult comps, and are starting to see green shoots of growth"; manufacturing "robust"; U.S. recovery "most prominent"; electronics +18% YoY) — and the economic assumption itself upgraded from contraction at the top end (2Q25) to no improvement at the midpoint. Self-assessment on execution is the most critical in the series: the CEO opened, unprompted, with "we are not satisfied with our margin performance for this quarter," and closed with "this is merely a guide. How we perform is what matters most. We know our owners expect more." Linde has not opened a call with a self-criticism in the seven transcripts on hand.

Contradictions (10 found, concentrated in margin promises). The FY26 margin-expansion algorithm was silently restated from "30 to 50 basis points" to "40 to 60 basis points" in May, immediately before a first half that delivered less than zero ex pass-through. APAC electronics-equipment dilution went from "a kind of a one-off" (1Q26) to "an integral part of our model, always have been and will continue to be" (2Q26). Headwinds are labelled "temporary… recover in the coming quarter" in the same call where both executives describe every driver as multi-year. And Lincare went from "robust and resilient" with "progress being made" (4Q24) to a whole-business divestiture review, "in part and as a whole" (2Q26).

Catalysts into 3Q26 (reports Oct 30, 2026). The Lincare strategic review is the largest discrete value event (~$0.21–$0.26 of annual EPS drag, with no divestiture in consensus); CFO White explicitly promised cost-action color on the October call; the backlog commitment to "finish the year with an 8 handle" is falsifiable arithmetic requiring at least $1.2B of net new 2H26 wins; and helium normalization — explicitly excluded from FY26 guidance — is un-modelled 2027 optionality. The uncomfortable comp: Air Liquide guided +100bps and Air Products delivered +110bps of margin expansion off the same electronics wave that cost Linde 121bps in APAC.

What the rows say together. Revenue: accelerating, six quarters running (+0.1% → +9.3%, a ~920bps acceleration off the trough). Volume: inflected, and it is real — the volume contribution turned positive in 4Q25 and doubled to +2 pts, while FX (+2 pts) is no longer the dominant swing it was in 1Q26 (+5 pts). Margin: decelerating, six quarters running — adjusted operating margin YoY went +115 → +79 → +5 → (45) → (10) → (55) bps; adjusted EBITDA margin +114 → +67 → +27 → (26) → (33) → (99) bps; gross margin's (164) bps is the sharpest compression in the window. EPS: accelerating, but on a thinner base — +5.3% → +10.0%, increasingly carried by revenue and a 2% lower share count (464.5M diluted, from 473.6M) rather than by margin.

Absolute scale — quarterly sales ($M)

Growth trend — YoY, 10 quarters

The other side of the story — margin YoY change (bps)

Segment detail — all four core segments compressed simultaneously

APAC is the worst (-121bps) on lower-margin electronics equipment sales. Americas (-57bps) is the one management called out — U.S. homecare cost inflation plus a policy change, compounded by dilutive mix from double-digit U.S. hardgoods growth. CEO Sanjiv Lamba: "we are not satisfied with our margin performance for this quarter." The company is now evaluating the strategic fit of U.S. homecare, "in part and as a whole."

Heat key. Every EPS cell in the window is a beat — read the magnitude, not the flag. Darker green = larger beat; yellow = the 2Q26 penny. On revenue the sign flips permanently at 2Q25 and the beats widen monotonically since: +1.41% → +2.13% → +3.03%.

Beat rates and the deteriorating magnitude

EPS: deteriorating, monotonically. Four consecutive step-downs; the beat has been roughly halved twice over two years. 2Q26's +$0.01 / +0.22% is the smallest surprise in the 12-quarter window (next-smallest 4Q25 at +$0.02). Under the investing-principles framework this is an explicit red flag — deteriorating beat magnitude — and it is not offset by the fact that the beats keep landing. Twelve consecutive beats with zero misses means the beat itself carries almost no information content; the marginal signal is entirely in the magnitude and the guide.

The conversion gap — why $273M of revenue upside became $0.01

Stated fairly: not all of the $273M is margin-bearing. Management's bridge puts cost pass-through at +1% (roughly $85M, zero-margin by construction), FX at +2%, and acquisitions and engineering at +1% each. Underlying sales rose only +4%, split evenly volume and price. The street under-modelled the low-quality components of revenue and correctly modelled the earnings. That is not a beat you pay for.

Management's explanation, ranked

CFO Matthew White quantified the ex-Lincare picture: "The Americas business ex the US home care or Lincare business would be up 20 basis points on margin ex pass through... Excluding this, margins would have increased." So the underlying gas business did expand margin. The entire consolidated compression traces to one non-core asset plus mix — and in the RBC exchange White confirmed "Lincare is going to be the focus given that is the biggest driver."

Why the guide did not go up with the revenue beat. "While base volumes showed some recovery in the second quarter, we would like a few more quarters under our belt before incorporating this trend into future guides. Therefore, we are leaving the back half guidance assumption the same as before."

Structural point that shapes this entire section: Linde guides adjusted EPS only. There is no revenue guide, no gross/operating/EBITDA-margin guide, no FCF guide — never has been. Every "revenue" and "margin" line below is either consensus or a management-stated qualitative expectation, and is labelled as such. Anyone presenting a "revenue guide vs consensus" for LIN is fabricating one.

The new guidance, verbatim

3Q26: "Third quarter guidance range is $4.45 to $4.55. Or 6 percent to 8 percent growth. This assumes no currency impact from prior year but does assume a 1 percent FX headwind sequentially. Consistent with prior approach, range assumes no economic improvement at the midpoint." — CFO Matthew White

FY26: "The updated full year range is $17.70 to $17.90. Or 8% to 9 percent growth excluding a 1 percent FX tailwind assumption. This range raises the prior bottom end by $0.10 but leaves the top unchanged. While base volumes showed some recovery in the second quarter, we would like a few more quarters under our belt before incorporating this trend into future guides."

Closing line, new in tone: "Of course, this is merely a guide. How we perform is what matters most. We know our owners expect more, and the organization is committed to delivering on those expectations."

Waterfall — FY26 adjusted EPS guidance walk

Guidance vs expectations

Proof that the back half was left unchanged to the penny

The entire $0.05 FY26 raise is the 2Q26 beat versus the 2Q guide midpoint. Not one cent of back-half assumption changed. This is verifiable arithmetic, not a characterization. It also means the implied 4Q26 of $4.47 (+6.4% YoY) is $0.07 below the 4Q26 street of $4.54 (+8.1%) — the back half, not 3Q, is where the guide and the street actually disagree.

Reported trend vs guided trend — and why the deceleration is an illusion

Reported trend: sharply accelerating. Guided trend: decelerating. That divergence is the whole debate on this print — and it is mostly mechanical.

1. FX lapping (the dominant reason). FY26 guidance assumed a 3% FX tailwind in 1Q26, 2Q26 carried +2%, and 3Q26 explicitly assumes zero YoY FX plus a 1% sequential headwind. Roughly 200–300 bps of the apparent deceleration is FX rolling off, not demand.
2. Comp normalization. 1H26 lapped the weakest stretch of the cycle (1Q25 revenue +0.1%). Lamba: "we have lapped the more difficult comps."
3. Deliberate non-recognition of the volume recovery. "We would like a few more quarters under our belt."

Ex-FX, the guided trend accelerates. Management stated 1Q26 EPS grew +10% reported but +5% ex-FX. The 3Q26 guide of +6% to +8% carries no FX at all — a fully organic +6–8% versus +5% organic in 1Q26. On the only apples-to-apples basis available, guided underlying EPS growth is accelerating by roughly 100–300 bps, the exact opposite of the headline. Present the ex-FX line or the chart tells the wrong story.

Guidance trajectory vs the same point last year

The guidance trajectory is unambiguously better than a year ago on every axis. A year ago Linde could not raise at all and embedded outright contraction at the top of the range. This year it raised the floor twice, removed the contraction assumption, and guides ~270 bps faster FY growth against a record backlog rather than a shrinking one.

Guidance accuracy — the reason to discount the "miss"

Four consecutive quarters at or above the top of the range, plus a full year that finished at the top of the final range. Against that record, an FY26 midpoint of $17.80 with the top pinned at $17.90 for a third straight guide is best read as the same conservatism that produced five straight top-end outcomes.

Other KPI guidance issued on the call

"Accel" below is the quarter-over-quarter change in the YoY growth rate, in basis points — a positive number means growth got faster than it was the quarter before.

The four inflection points

Net of the two trend lines, EPS growth is roughly where it was two years ago — but the composition has completely reversed. In 3Q24 Linde grew EPS 8.5% on 2.5% revenue. In 2Q26 it grew EPS 10.0% on 9.3% revenue. The company is working roughly four times as hard on the top line for the same earnings growth.

Quality of the acceleration — only ~4 of the 9.3 points is underlying

The 2Q26 bridge per CFO White: FX +2%, acquisitions +1%, engineering +1%, cost pass-through +1%, underlying +4%. Price has been a metronomic +2% in all eight quarters — the entire variable is volume, which has moved -1% → 0% → +1% → +1% → +2%. That volume line is the real signal, and it is genuinely and steadily improving.

Verdict: good and improving on volume, bad and deteriorating on margin, flattening on both. Under a trajectory-over-absolutes lens this is a hold-quality inflection, not a buy-quality one — you want revenue acceleration to re-accelerate with margin, and right now the two are trading off against each other. The next two quarters resolve it: if 3Q26 delivers the promised sequential margin recovery while volume holds +2%, the spread re-widens and the compounding story is intact. If homecare drags again, Linde is a mid-single-digit-underlying business paying for its EPS growth almost entirely with FX, buybacks and pass-through.

Ten dated catalysts that move the LIN earnings algorithm over the next four quarters, ranked by EPS sensitivity. Next print: 3Q26 on 2026-10-30.

Catalyst 1 — Lincare, the swing factor management just put in play

The catalyst is asymmetric: an exit removes a ~$0.21–$0.26 annual drag and re-rates Americas mix, while "we're fixing it" is the outcome already in numbers. External policy signal: CMS's 2026 DMEPOS final rule (CMS-1828-F, published 2025-11-28) reinstated competitive bidding but excluded legacy oxygen and CPAP — that removes a tail risk but does nothing for the actual squeeze, since CY2026 carries only a 2.0% update factor on non-CBP items against the labor-cost inflation Lamba named as the primary headwind. The structural math does not fix itself, which is precisely why the strategic review exists.

Catalyst 4 — the backlog commitment is checkable arithmetic

This is a falsifiable, self-imposed, dated commitment — exactly the kind of promise-tracking that separates good management teams from bad. Lamba made the same call one quarter earlier ("could potentially have an 8 handle on it") and delivered it two quarters early; precedent at 3Q25 was also good ("I had said 3 months ago my expectation is we will end the year with a 7 handle... despite starting up $1 billion in projects" — delivered). Project economics: post-tax double-digit unlevered IRR hurdle, 2–3 year build, then a ramp; revenue-to-capex converts at 20–50% depending on energy pass-through and tolling.

Catalyst 5 — the competitor contrast is the most uncomfortable fact in this quarter

Both peers expanded margin by triple-digit basis points off the same electronics wave that cost Linde 121 bps in APAC. LIN's defence — that low-margin equipment sales buy future gas pull-through — is coherent and consistent with a decade of on-site practice, but it is an assertion that only the backlog conversion in 2028–2029 can prove.

Catalyst 9 — the contrarian read: management sees a recovery it refuses to guide

"While base volumes showed some recovery in the second quarter, we would like a few more quarters under our belt before incorporating this trend into future guides. Therefore, we are leaving the back-half guidance assumption the same as before." — CFO Matthew White

Against that conservatism, the signals management itself described: U.S. package gases mid-to-high-single-digit and hardgoods double-digit (management's own stated leading indicator); manufacturing the fastest-growing industrial end market with aerospace over a third of it; APAC volume +6% for a second consecutive quarter versus roughly flat through 2025; "we have lapped the more difficult comps"; and metals customers sounding a "slightly higher degree of optimism." Guidance assumes no economic improvement at the midpoint — so the FY26 range is a floor built on a deliberately stale back-half assumption, and 3Q26 is the first of the "few more quarters" that would let it into the FY27 guide in February.

What to watch, dated

Topic-intensity shift. Keyword mentions across the full transcripts: Lincare / homecare went 3 (3Q25) → 0 (4Q25) → 1 (1Q26) → 11 (2Q26). Backlog went 14 → 19 → 13 → 28. The call polarized into the two things that moved the print — a record $8.1B sale-of-gas backlog (management's story) and U.S. homecare (the street's story). Homecare went from a throwaway mention to the densest topic on the call in one quarter. The street has repriced Lincare from noise to thesis risk.

Notably un-asked: the FY26 guide midpoint ($17.80) sitting below street ($17.91), and the +0.2% EPS beat — the thinnest in twelve quarters.

The seven deflections, ranked by how much the missing answer matters

The answers that did land

The pattern, and what it means

The seven deflections are not randomly distributed. Six of seven cluster on the two questions that determine whether the margin story is a temporary dislocation or a structural break. Anything requiring a number on margins or homecare was resisted; the one hard figure that emerged came only because an analyst showed his own math and asked management to correct it — and management corrected it accurately, which suggests the reticence is about volunteering, not about knowing. Anything about the durable quality of the top line was answered qualitatively. Anything forward-looking on helium was answered with a definition instead of a direction.

Conversely, every question about the growth story was answered generously and specifically. This is not an evasive management team — Linde volunteered several answers that cost them something, and the CEO opened with an unprompted "we are not satisfied with our margin performance." It is a management team that is specific where it has conviction and vague where it does not yet have a plan. The asymmetry itself is the signal: the backlog and electronics answers are underwritten; the margin-recovery and Lincare answers are not yet.

October 30 is the date the deferred answers come due. Four things to check: (1) the cost-action disclosure White explicitly promised — if it does not appear, credibility takes a real hit because it was volunteered, not extracted; (2) whether the homecare review gets a date, a perimeter ("in part" vs "as a whole"), or an outcome; (3) whether APAC volume growth holds once the electronics equipment sales anniversary; (4) whether the promised sequential operating-margin improvement actually lands against the 29.5% 2Q base.

Scope: all seven earnings-call transcripts on hand (4Q24 through 2Q26) plus the repo's local fact inventory. This is not a "no contradictions found" quarter. The cluster is unusually dense and it is concentrated in one place: margin promises. Management set an explicit, repeatedly-escalated FY26 margin-expansion commitment in February, re-affirmed and raised it in May, then delivered flat-to-down margins in 1H26 and re-labelled the drivers "temporary" while simultaneously describing those same drivers as multi-year and structural.

Tested and dismissed — not contradictions

What this means

Nine of the ten contradictions run in one direction: an optimistic forward commitment, followed by a quiet reframing when the number arrives. The FY26 margin algorithm was raised (30–50 bps becoming 40–60 bps) immediately before two quarters that expanded margins by less than zero; the APAC equipment drag was "a one-off" and is now "an integral part of our model"; helium improvement was pre-labelled pure upside and delivered none; a structural restructuring whose benefits were due in 2H26 now requires a second round of cost actions with the back-half guide untouched; and Lincare went from "robust and resilient" with "progress being made" to a whole-business divestiture review.

Against the investing-principles management test — guidance accuracy, promise tracking, absence of conflicting statements across transcripts — LIN passes on guidance accuracy at the EPS line (twelve consecutive beats, though 2Q26's +0.2% is the thinnest of them) and fails on promise tracking and internal consistency at the margin line. The EPS algorithm is being defended by the buyback and by cost pass-through optics while the operating margin commitment quietly slips. That is the specific thing to interrogate on the October call, where management has already promised to quantify the new cost actions.

One genuine mitigant, and it should not be lost: the contradictions are concentrated in Lincare and in mix, not in the core industrial gas franchise. Lamba's disclosure that Americas margin ex-Lincare would have been up 20 bps ex pass-through is consistent with the Daloopa-sourced segment history and indicates the leadership economics of the core business are intact. The contradictions are about how honestly the deterioration in one non-core asset — and the mix cost of the electronics growth engine — has been communicated over the last six quarters.

Linde is the single best real-time macro instrument in the industrial complex: it sells atmospheric gases to essentially every physical-goods producer on earth, prices to local inflation, and its U.S. packaged-gas/hardgoods line is management's own stated leading indicator. That makes this call more valuable for what it says about other people's businesses than about Linde's own quarter.

Macro scorecard

Named counterparties and what the call means for them

The two read-throughs that actually matter

1. U.S. advanced-node semiconductor capex is being underwritten with hard cash by a supplier with a double-digit-unlevered-IRR hurdle. A $1B backlog add for Western-U.S. advanced-node fabs, plus ~$800M in the Taiwan JV, plus construction already started under reimbursable LOIs, plus "a few more large opportunities we are currently pursuing," plus an expectation the backlog ends the year "with an 8 handle" after $1.3B of start-ups roll off. Linde does not build ASUs on hope. This is among the cleanest third-party confirmations available that the AI fab buildout extends multiple phases beyond what is already announced.

2. The reimbursement-capped U.S. homecare model is breaking under labor inflation, and the best operator is heading for the exit. ~$40M/quarter of drag, no pricing power, a new management team already installed, aggressive pruning already done — and it still is not enough. For LIN this is a sentiment-inversion setup (removing the drag would have made Americas margin +20 bps ex pass-through instead of negative). For the DME/HME sector, it is a warning that the marginal operator is in worse shape than the tape suggests.

The sharpest single macro comment

"We will likely look to take some cost actions this quarter... I mean, it is clear you are seeing more inflation around the world, and that is something that we have to manage through our productivity and our actions. And in some regions, you are seeing growth, which supports it. In other regions, you are seeing inflation without the growth. And that is an area we are gonna focus on specifically for this quarter." — CFO Matthew White

"Inflation without the growth" is a stagflation description, from the CFO of a company operating in roughly 80 countries. Note the asymmetry versus pricing: Linde priced +2%, "generally tracked with local inflation" — pricing is matching inflation, not out-running it. That is why margins compressed despite a record top line. For companies with weaker pricing power than Linde — which has among the best in industrials — this is a margin warning for 2H26 estimates across labor-heavy, price-taking industrial and services businesses, especially those with reimbursement-capped revenue.

What was conspicuously not discussed

Accelerating on revenue, decelerating on margin — and for the first time this cycle the margin line is winning the argument. Sales growth has climbed for six straight quarters to +9.3% with volume at a two-year high and a record $8.1B backlog behind it. But adjusted operating margin has fallen YoY for three consecutive quarters (worst at -55 bps), gross margin compressed 164 bps, and the EPS beat narrowed to +0.2%, the thinnest in twelve quarters. LIN is converting a genuinely better demand environment into progressively less incremental profit.

Three things are true at once, and the market conflated them on print day:

1. This was a raise, not a cut. The FY26 floor went up $0.10; consensus of $17.91 sits one cent above the company's own ceiling. Against a five-for-five top-end track record, the guide is the same conservatism that has produced five straight top-end outcomes.
2. The headline guided deceleration is FX, not demand. Ex-FX, guided EPS growth accelerates from +5% organic in 1Q26 to a fully organic +6–8% in 3Q26. Present the ex-FX line or the chart tells the wrong story.
3. The margin problem is real, is concentrated, and is not yet fixed. It is one non-core asset (Lincare, ~$0.21–$0.26 of annual EPS) plus growth-linked mix. The core gas franchise expanded margin — Americas ex-Lincare was +20 bps ex pass-through. But management raised its own margin algorithm right before missing it, called permanent drivers temporary, and deferred every number that would pin down the fix.

Under a trajectory-over-absolutes lens this is a hold-quality inflection, not a buy-quality one. You want revenue acceleration to re-accelerate with margin; right now the two are trading off. October 30 resolves it. If 3Q26 delivers the promised sequential margin recovery, sizes the cost actions White volunteered, and dates the Lincare review while volume holds +2%, the spread re-widens and the compounding story is intact — with helium normalization and a homecare exit as un-modelled 2027 optionality on top. If the EPS surprise stays at a penny while revenue beats again, the margin algorithm is broken and a Lincare exit becomes necessary rather than optional.