ANCHOR Check - United Parcel Service (UPS)

UPS is not an AI-native software story.

Good.

It is a real-world logistics network with hard assets, route density, labor complexity, regulatory exposure, and customer concentration risk. Under ANCHOR, UPS clears the bar. Not because it is pretty. Because replacing it is miserable.

Badge: ABIP ANCHOR Certified

Gates:

UPS owns a scaled physical delivery network that AI can optimize but not replace. The moat is density, infrastructure, trust, customs capability, and operational muscle. The weakness is labor cost, cyclicality, Amazon exposure, and margin pressure in a market where volume growth is not guaranteed.

The market is focused on the Amazon glide-down.

Fair.

Amazon represented 10.6% of UPS revenue in 2025, and UPS is deliberately reducing that volume because it was lower-margin and strategically messy. Management is trying to swap cheap volume for better mix: healthcare, SMB, international, returns, and higher-value packages. In Q1 2026, UPS revenue was $21.2 billion, down slightly, while U.S. domestic revenue fell 2.3% on expected volume declines. Revenue per piece rose 6.5% domestically. That is the story: fewer packages, better packages.

The investor question is simple.

Can UPS shrink the wrong volume without breaking the network economics?

UPS is an atoms business.

The choke point is not “can software predict demand.”

The choke point is whether a package can be picked up, sorted, flown, cleared, routed, loaded, delivered, returned, billed, and recovered across a giant network without the unit economics collapsing.

AI helps here.

It can improve routing.
It can forecast volume.
It can automate customer support.
It can optimize pricing.
It can reduce waste inside hubs.

But AI does not replace the hub in Louisville.
It does not create ground density in every ZIP code.
It does not negotiate the Teamsters contract.
It does not clear customs.
It does not absorb peak season chaos.
It does not make an unprofitable Amazon package profitable by magic.

UPS’s real durability is not brand fluff.

It is physical network density plus operational complexity plus trust at scale.

That is hard to copy.

It is also expensive to maintain.

A — Asset-Embedded: 9/10

UPS is deeply embedded in physical infrastructure.

Air hubs. Ground routes. Vehicles. aircraft. sorting facilities. customs brokerage. billing systems. shipper integrations. The company says its ground fleet serves substantially all business and residential ZIP codes in the contiguous U.S., and its air network is anchored by Louisville with additional U.S. and international hubs.

This is not a thin software layer.

This is a network of steel, labor, route density, customer systems, and delivery habits.

N — Non-Discretionary: 7/10

Shipping is not fully non-discretionary.

Some e-commerce volume is cyclical. Some consumer parcels are discretionary. Some retail demand can disappear fast.

But businesses still need to move goods. Healthcare still needs logistics. Cross-border commerce still needs customs. Returns still need handling. SMBs still need delivery infrastructure.

UPS is not selling luxury attention.

It is selling movement.

That earns a strong score, but not a perfect one.

C — Capital-Intensive: 9/10

This is a capital-heavy business.

UPS guided for about $3.0 billion of capital expenditures in 2026 after generating $88.7 billion of revenue in 2025. It also disclosed nearly $3.0 billion of purchase commitments as of year-end 2025.

A new competitor cannot spin this up with a pitch deck and an API.

The capex wall is real.

H — Hard to Replace: 8/10

UPS is replaceable at the margin.

A shipper can move some volume to FedEx, USPS, regional carriers, Amazon Logistics, DHL, freight brokers, or gig-style same-day players.

But replacing UPS at national scale is different.

The company served 1.6 million shipping customers and more than 10.7 million delivery recipients daily in 2025. That kind of density matters.

A delivery network gets stronger when it already has stops on the route.

That is the moat.

Not invincible.

Very real.

O — Obsolescence-Resistant: 8/10

AI does not obsolete parcel logistics.

It improves it.

The more software compresses commerce, the more physical fulfillment still has to happen somewhere. Digital demand eventually becomes a box, a label, a return, a customs form, a cold-chain shipment, or a truck roll.

The risk is not that AI kills UPS.

The risk is that AI improves everyone’s logistics stack and makes pricing more transparent.

That pressures margin.

It does not erase the need.

R — Real-World Demand: 8/10

UPS has obvious real-world demand.

Packages move because goods move.

Healthcare moves.
Parts move.
Returns move.
International shipments move.
SMB commerce moves.

The issue is mix.

Low-margin residential e-commerce can fill the network and dilute economics. UPS is trying to move away from the wrong packages and toward higher-value ones. That makes strategic sense, but it is execution-heavy.

The demand is real.

The profitability of each package is the fight.

Labor is the first risk.

UPS had approximately 460,000 employees at year-end 2025, with nearly 80% of U.S. employees represented by unions, primarily package-handling and transportation workers. The national Teamsters master agreement runs through July 31, 2028.

That protects operating discipline in one sense.

It also limits flexibility.

Second risk: Amazon.

UPS is reducing Amazon volume on purpose, but shrinking a large customer relationship while reconfiguring the network is not clean. In Q1 2026, Amazon was down to 8.8% of revenue, according to CEO commentary reported by Supply Chain Dive, but the transition still has to land.

Third risk: network deleveraging.

Package networks hate empty space. If volume falls faster than cost comes out, margins get hit.

Fourth risk: regulation and compliance.

UPS runs through aviation, customs, transportation security, environmental rules, labor rules, data protection, and safety regulation. The company is not just delivering boxes. It is operating inside a dense regulatory machine.

Fifth risk: AI-enabled competition.

AI can help shippers compare rates, optimize carrier selection, split volume, and reduce dependence on any single carrier. That does not kill UPS. It can weaken pricing power.

If I’m right:

UPS exits the Amazon glide-down with a cleaner mix, better revenue per piece, and a network sized for profitability instead of vanity volume. AI becomes an operating lever, not a strategic threat.

If I’m wrong:

The volume loss creates too much network deleveraging, labor costs stay too rigid, and customers use AI-enabled logistics tools to squeeze carrier pricing harder than UPS can offset with mix.

What would change my mind:

Evidence that SMB, healthcare, international, and returns volume cannot replace the lost Amazon density at acceptable margins.

Or evidence that regional carriers and Amazon Logistics are taking enough profitable volume to make UPS’s national network less special.

AI should help UPS more than it hurts.

Routing, forecasting, pricing, customer support, facility automation, claims, fraud, customs workflows, and capacity planning all benefit from better intelligence. But AI does not remove the need for physical delivery density.

This is not software being disrupted by software.

This is software trying to make the machine less expensive to run.

AI can optimize the route.

It still can’t put the package on the porch.

— Connor
Alpha Before It Prints

🔗 Click to Connect with me

© 2026 Alpha Before It Prints
Unsubscribe