ANCHOR Check - Linde (LIN)

Linde is one of the cleanest ANCHOR businesses on the board.

It sells molecules that industrial customers need to keep plants, fabs, hospitals, food systems, refineries, metals operations, and manufacturing lines running. The moat is not an app. It is installed infrastructure, long-term contracts, route density, engineering depth, purity reliability, and customer switching pain.

ANCHOR Score: 52 / 60

Badge: ABIP ANCHOR Certified

Gates:

Linde is durable because its product is not cognition. It is industrial gas supply embedded inside real customer operations. AI helps demand through semiconductors and helps operations at the margin. It does not replace the need for molecules delivered reliably at scale.

The market is focused on electronics, backlog, margins, and whether AI-driven semiconductor capex can turn into durable earnings growth.

That is fair. Linde reported Q2 2026 sales of $9.3 billion, up 9%, with underlying sales up 4%. Adjusted EPS grew 10%, adjusted operating margin was 29.5%, and project backlog reached $11 billion. Management also raised the lower end of full-year 2026 adjusted EPS guidance to $17.70–$17.90.

The AI angle is real. But it is not “Linde uses AI.” The angle is that advanced fabs need ultra-high-purity gases, and Linde just signed a long-term agreement tied to a major semiconductor expansion in Phoenix, including a planned $1 billion investment in on-site gas infrastructure.

Linde’s business is boring in the best possible way.

Industrial gases are not optional inputs for many customers. They are part of the process. Oxygen. Nitrogen. Argon. Hydrogen. Helium. CO₂. Specialty gases. Electronic gases. Medical oxygen. Food gases. Welding gases. Refining gases. Fab gases.

AI can optimize routing.

It cannot make a semiconductor fab run without ultra-high-purity gases.

AI can improve maintenance scheduling.

It cannot build an air separation unit overnight.

AI can forecast demand.

It cannot remove the cost, permitting, reliability requirements, safety burden, or customer switching risk attached to industrial gas supply.

This is the key: Linde is not selling an information product. It is selling operational continuity.

Customers do not want “a cheaper UI.” They want the gas to show up, at spec, without interruption.

A — Asset-Embedded: 9/10

Linde is deeply embedded in physical infrastructure. The company manufactures and distributes industrial gases through production plants, pipeline complexes, distribution centers, delivery vehicles, on-site plants, tanks, and customer-adjacent systems. Its 10-K says most industrial gas products are uneconomical to transport more than a few hundred miles, which makes local production and density matter.

This is not a loose vendor relationship. In on-site supply, Linde builds plants on or next to customer sites and supplies product directly by pipeline. That is embedded.

N — Non-Discretionary: 8/10

The demand is not perfectly recession-proof. Manufacturing, metals, chemicals, energy, and electronics cycles still matter.

But the use cases are mission-critical. Healthcare oxygen. Semiconductor gases. Food and beverage gases. Refining hydrogen. Steel oxygen. Manufacturing nitrogen. These are not nice-to-have workflow tools. Linde serves healthcare, chemicals and energy, manufacturing, metals and mining, food and beverage, and electronics.

The customer may run fewer tons in a downturn. But if the plant is operating, the gas supply has to work.

C — Capital-Intensive: 9/10

This is a capex business.

Linde spent $5.261 billion on capital expenditures in 2025, primarily for new plant and production equipment tied to backlog growth. For 2026, management expects capex of $5.5 billion to $6.0 billion.

That matters. AI-native competitors can spin up software. They cannot cheaply replicate Linde’s installed plant base, route density, safety systems, customer site integration, and engineering backlog.

H — Hard to Replace: 9/10

Linde’s on-site contracts typically run 10–20 years, include minimum purchase requirements, and contain price escalation provisions. Some long-term supply contracts run up to 30 years.

That is the opposite of low switching cost.

The customer is not swapping a dashboard. They are replacing a physical supply system attached to uptime, safety, quality, engineering, and plant operations.

O — Obsolescence-Resistant: 9/10

AI does not obsolete industrial gases.

It may change which customers grow fastest. More electronics. More advanced packaging. More data-center-related industrial activity. More specialty gas demand. Less of something else.

But the underlying requirement survives: physical production needs physical inputs.

Even semiconductor progress helps Linde more than it threatens it. The Phoenix semiconductor agreement is a clean example: AI chip demand may be digital, but fabs are industrial monsters. They consume power, water, chemicals, gases, land, permits, construction labor, and uptime guarantees.

R — Real-World Demand: 8/10

Linde has real-world demand across multiple end markets, not one fragile vertical.

The 2025 revenue base was about $34 billion, with industrial gases sold through merchant, on-site, and packaged gas channels. In Q2 2026, volumes increased primarily from electronics, manufacturing, and chemicals and energy.

The weakness: some demand is cyclical. Europe can be soft. Manufacturing can slow. Chemicals can retrench. Helium and rare gases can have pricing cycles.

But the core demand is tied to production, healthcare, food, energy, fabs, metals, and industrial uptime. That is real-world demand.

Energy costs matter. Electricity and natural gas are major inputs. Linde has pass-through mechanics in many contracts, but not every cost spike is painless.

Project execution matters. Large plants can run late, over budget, or underperform. That hurts customer trust and returns.

Electronics concentration can cut both ways. AI chip capex is a tailwind today. If fab buildouts slow, the growth narrative cools.

Regulation and safety risk are real. This business involves plants, pipelines, storage, trucks, hazardous materials, environmental rules, and operational failure risk. Linde’s own filing flags fire, toxic release, explosions, vehicle accidents, environmental damage, litigation, and regulatory exposure.

Pricing discipline matters. The business looks great when price, productivity, and capital discipline offset inflation. It looks less great if competition forces bad contracts or returns compress.

If I’m right:

Linde keeps compounding because customers keep needing molecules, fabs keep needing purity, and industrial gas supply remains a physical bottleneck. AI becomes demand support, not disruption.

If I’m wrong:

The analysis is too bullish if electronics capex rolls over hard, new projects earn weaker returns, energy cost pass-through breaks down, or customers push pricing harder than expected.

What would change my mind:

Evidence that long-term contract quality is weakening. Lower minimum commitments. Worse price escalation. Poor project returns. Backlog growth that requires bad capital. Or margin pressure that productivity cannot offset.

AI Tailwind

AI helps Linde through semiconductor demand, advanced manufacturing, forecasting, routing, maintenance, and operating efficiency. But the real point is simpler: AI needs chips, chips need fabs, and fabs need ultra-high-purity gases.

AI can write the model.

It still needs someone to deliver the molecules.

— Connor
Alpha Before It Prints

🔗 Click to Connect with me

© 2026 Alpha Before It Prints
Unsubscribe