The telecom industry has spent decades getting very good at moving tiny pieces of plastic around.

SIM cards were ordered, packaged, shipped to stores, distributed to dealers, assigned to customers and, eventually, replaced when somebody lost a phone or needed a different format. Operators built forecasting models around that physical flow. Too much stock meant wasted money. Too little meant an angry customer standing at a retail counter.

eSIM does not eliminate that supply-chain problem. It changes what “stock” means.

That distinction is becoming harder to ignore. Apple’s current iPhone 17 generation is eSIM-only in a much wider group of markets than the original U.S.-only shift, including Canada, Mexico, Japan and several Gulf countries. Google has gone in the same direction with U.S. Pixel 10 models. Samsung, meanwhile, continues to expand eSIM support across its Galaxy portfolio, even where physical SIM remains available.

For operators, the signal is clear: the device market is moving faster than many provisioning operations.

The warehouse is becoming a server

Physical SIM stock is easy to picture. Digital SIM inventory is not.

With eSIM, an operator still needs credentials, profiles, activation logic, backend capacity, and a reliable route from the customer request to the SM-DP+ platform and, finally, into the device’s eUICC. Depending on how the operator has implemented provisioning, profiles may be prepared in advance or generated and assigned dynamically. GSMA’s consumer Remote SIM Provisioning architecture and Android’s implementation make clear just how many digital components now sit behind what looks to the customer like a simple activation.

So the old question — “How many SIM cards do we have in the warehouse?” — becomes several questions.

How many profiles are available for activation? How quickly can new ones be created or assigned? Are activation codes being consumed correctly? Which channels are drawing from which pools? Can the platform handle a sudden device launch? What happens if a provisioning partner, entitlement server or operator backend slows down?

READ MORE: Provisioning Is Becoming the Boring Part of eSIM

This is inventory management, just without boxes.

And digital shortages can be nastier than physical ones. A retail store can sometimes borrow SIM cards from another branch. A failed profile-allocation workflow at 8 p.m. on launch day can block thousands of customers at once.

Apple changed the planning horizon

Apple removed the physical SIM tray from U.S. iPhones beginning with the iPhone 14. At first, many operators outside the United States could reasonably treat that as somebody else’s operational problem.

That excuse is disappearing.

Apple now lists iPhone 17-series models as eSIM-only not only in the United States but also in Canada, Mexico, Japan, the UAE, Saudi Arabia, Bahrain, Kuwait, Qatar and Oman, among other listed markets. That matters because operators can no longer assume they will have several comfortable device generations to prepare.

The pressure is not only Apple’s. Google says U.S. Pixel 10 models, except the Pixel 10 Pro Fold, are eSIM-only. Samsung’s 2026 support pages show eSIM across flagship Galaxy S26 devices and an increasingly broad set of Galaxy A, Z and tablet models.

Android is not uniformly abandoning the SIM tray. That is actually what makes planning harder. Operators must support a mixed estate: physical SIM, eSIM, dual-SIM combinations, transfers and conversions, often across prepaid, postpaid, enterprise and MVNO brands at the same time.

Forecasting needs new inputs

Traditional SIM forecasting leans heavily on subscriber growth, retail demand, replacement rates and distribution stock.

An eSIM-first forecast needs device intelligence too.

Operators should know what percentage of their active base is eSIM-capable, which handset launches are likely to drive migrations, how many customers typically replace devices in each quarter, and how much transfer traffic follows major Apple, Google or Samsung releases.

More importantly, they need visibility into the provisioning funnel itself.

A headline number such as “500,000 profiles available” is not enough if 40,000 are stuck in an unusable state, one sales channel is burning through allocations unexpectedly, or a transfer process requires manual intervention.

Android’s current transfer architecture shows why. Google’s Android documentation describes a flow involving the device transfer client, carrier entitlement server, activation code, Local Profile Assistant and SM-DP+. Google also recommends that carrier backends process most SIM-swap requests in under five seconds.

That is not warehouse logistics. It is real-time service orchestration.

Smaller operators feel the squeeze first

Large mobile groups can throw engineering teams, multiple vendors and redundancy at the problem. Smaller MVNOs and regional brands often cannot.

Many depend on host operators, aggregators or third-party eSIM platforms for parts of the provisioning chain. That can make eSIM fast to launch commercially but harder to observe operationally. If something fails, the brand facing the customer may not control the system causing the failure.

This does not mean every small MVNO should build its own SM-DP+ or entitlement stack. For many, that would be expensive overkill.

READ MORE: Google Pixel 10 Could Go eSIM-Only in the US

A better alternative is demanding stronger visibility from suppliers: profile status, allocation rates, failure reasons, transfer performance, channel consumption and clear capacity thresholds. The strategic asset is not owning every component. It is knowing what is happening across them.

That is where many current eSIM offers still need work. Activation is marketed heavily; profile lifecycle operations are discussed far less.

The next stockroom is invisible

The physical SIM will not disappear everywhere at once. Samsung still supports physical-plus-eSIM configurations in many markets, and millions of customers will remain on devices with SIM trays for years.

But operators that use that fact as a reason to postpone change are reading the transition backwards.

Apple has already shown that a device maker can remove the tray market by market. Google is now reinforcing the pattern on Android. GSMA’s consumer Remote SIM Provisioning specifications continue to evolve — SGP.22 v2.7 was published in April 2026 — while Android is building transfer deeper into device setup itself.

The competitive difference will therefore move away from who can distribute plastic most efficiently.

It will move toward who can forecast digital demand, expose profile availability, automate transfer and recovery, and see provisioning problems before customers do.

The best operators will treat eSIM profiles less like downloadable accessories and more like live network inventory. MVNOs should demand the same visibility from their platform partners.

Because once the SIM tray is gone, “out of stock” does not disappear.

It just becomes an API response.