SMS and Phone Verification Statistics 2026: Invalid Numbers, Landlines, Reassigned Numbers and Delivery Rates

Key takeaways

• Only 1.6% of US adults are landline-only, and 79.2% are wireless-only (July to December 2025).[1] Landlines on your list are a data-entry and decay problem, not a population problem.
• The FCC estimated that about 35 million numbers are disconnected and made available for reassignment each year. That figure dates from 2018.[2]
• The FCC's reassigned-numbers safe harbor costs $8 for 1,000 queries a month at the lowest tier.[3]
• No independent public benchmark exists for the US A2P delivery rate, the A2P rejection rate, or the show-rate lift from phone verification.
• In our experience, only about 40% to 45% of raw records on a list texted for years without validation end up as unique, SMS-capable numbers.
• Validation proves a number can receive texts. Verification proves the person controls it. Neither proves consent.

The short answer

About 97% of US adults have a wireless phone, and only 1.6% rely on a landline alone.[1] The FCC's last official estimate is that about 35 million numbers are disconnected and made available for reassignment each year.[2]

Your list still fills with landlines and dead numbers. People type their home or office line into forms, numbers churn, and nobody re-checks records collected years ago.

Here's what most "SMS statistics" pages won't tell you: no public, independent benchmark exists for the US A2P delivery rate, the A2P campaign rejection rate, or how much phone verification lifts show rate. A page that gives you one precise figure for any of those isn't citing data.

What does exist is a federal phone-status survey, the FCC's reassignment figure and safe harbor, carrier error codes, and list-cleaning results from operators. That's what this page compiles. Every figure carries its publisher, date and method.

Landlines: fewer people have them than your list suggests

The best source on phone ownership is the CDC's National Health Interview Survey, which asks households about their phones in person. Its latest release, published in September 2026, covers July to December 2025.[1]

1.6%

Share of US adults who are landline-only, July to December 2025. 79.2% are wireless-only.

Phone status	All adults	Adults 65+
Wireless-only	79.2%	63.4%
Wireless-mostly	12.8%
Dual use	3.2%
Landline-mostly	1.9%
Landline-only	1.6%	3.5%
No phone	1.0%

Source: CDC NCHS, NHIS Early Release, September 2026. The 18 to 24 landline-only estimate is suppressed as unreliable, not zero.

At household level, 80.0% of homes with adults are wireless-only, 18.6% have both a landline and a cell phone, and 1.0% are landline-only.[1]

Landline-only adults ticked up from 1.4% in the July to December 2024 release to 1.6%. With only 186 landline-only adults in the sample, that's within noise. Don't call it a trend.[1]

Two things follow for event and webinar lists:

• Older audiences carry more landlines. If your seminar audience skews 65+, expect more of them, though still only 3.5% are landline-only.
• Most landlines on your list belong to people who also own a cell. Nearly one household in five still has both.[1] When the form says "phone", some people type the home number.

The wider network is enormous. CTIA counts nearly 603 million US wireless connections in 2025, about 1.8 per American, but 46% of connected devices aren't phones.[4] Americans exchanged nearly 2.2 trillion SMS and MMS messages in 2024, person-to-person and business traffic combined.[5] Neither number tells you how many texts reach a real person.

Reassigned numbers: 35 million a year and an $8 safe harbor

A reassigned number is a consent problem, not just a delivery problem. The text goes through. It just reaches someone who never agreed to hear from you.

The FCC's 2018 order creating the Reassigned Numbers Database estimated that about 35 million numbers are disconnected and made available for reassignment each year. It also set a minimum aging period of 45 days before a disconnected number can be reassigned, and required providers to report disconnections monthly.[2] That estimate is about eight years old, and no newer official figure exists. You'll see it misquoted as "35 million reassigned". The FCC said made available.

The database comes with a safe harbor. Under 47 CFR 64.1200(m), a caller isn't liable for reaching a reassigned number if it had consent and the database wrongly reported the number as not disconnected.[6] The FCC declined to extend that protection to commercial databases.[2]

It's cheap. After a 20% price cut on April 28, 2025, Tier 1 costs $0.008 a query, or $8 for 1,000 queries over one month. Tier 2 is $0.006 a query, or $60 for 10,000.[3]

Usage is growing fast, according to a secondary summary of the database administrator's 2025 annual report. It cites 6.22 billion queries in 2025, up from 3.37 billion in 2024, and 314 million "yes" responses, meaning the number was disconnected after the consent date. That's about 5% of queries. It also reports about 28 million permanent disconnects a month.[7] We couldn't open the primary report, so treat these as indicative.

Carrier churn adds context, though churn isn't reassignment, because many customers port their number. AT&T's prepaid churn was 2.89% a month in Q4 2025, against 0.98% for postpaid phones.[8] Verizon's 2025 postpaid phone churn averaged 0.92% a month.[9] Prepaid-heavy audiences decay faster.

CTIA's guidance adds a carrier-side expectation: process deactivation files regularly, for example daily, and remove deactivated numbers.[10]

Invalid and fake numbers: what we know and what nobody measures

People don't always want to give a phone number. In Baymard's research on ecommerce checkout, 14% of 1,026 US shoppers said they were reluctant to provide one, and test users admitted typing fake numbers. Baymard also found 39% of sites required a phone number without explaining why.[11] That's checkout, not event registration. It explains how fake numbers get into forms. It isn't an invalid-rate benchmark.

Consumers are broadly open to business texts. In SimpleTexting's January 2026 survey of 1,000 US consumers, 85.6% said they were opted in to texts from at least one business.[12] Opted in to some business isn't consent to yours.

What nobody has measured neutrally is the share of web-form numbers that are invalid. Precise invalid-rate figures circulate on vendor pages without a study behind them. One vendor's invalid-rate figure wasn't on its page when we opened it.

Delivery rates: the honest table

Question	What the evidence says
Average US A2P 10DLC delivery rate	No independent public benchmark. Carriers and The Campaign Registry don't publish one, and the precise percentages on listicles cite no study.
A2P campaign rejection or approval rate	None published by The Campaign Registry, carriers or providers.
Share of web-form numbers that are invalid or landlines	No neutral study. Vendor claims only, none with a method.
Show-rate lift from passcode or line-type checks at opt-in	No public data.
Current annual reassigned numbers	Only the FCC's 2018 estimate of ~35 million made available a year.[2]
Filtering of texts with shortened links	One vendor figure: 1.96% filtered with third-party shorteners vs 0.07% with Twilio's own, across 63 million+ messages, period undisclosed.[13]
Daily send caps	Published. T-Mobile caps: 1,000 segments a day for Sole Proprietor brands, up to 2,000 for Low-Volume Standard, 2,000 to unlimited for Standard by Trust Score.[14]
"98% open rate"	Untraced. CTIA's 2022 FCC comments call it an estimate, citing an FCC advisory committee report.[15] SMS has no open tracking.

'No benchmark' means we searched for a method-disclosed, independent source and found none.

On drip sends from a dedicated number on a registered campaign, with links on the brand's own domain, the target we hold is 95% or better delivery. That's our operating target, not an industry benchmark, because there isn't one to compare it with.

Why carriers filter as hard as they do: Americans reported $470 million in losses to text scams in 2024, more than five times the 2020 figure.[16] Text was the contact method in 7% of 2025 fraud reports with a loss, behind websites and apps at 31% and social media at 28%.[17] The FCC logged about 15,300 unwanted-text complaints in 2021, 18,900 in 2022, 12,000 in 2023 and 15,600 in 2024 through August 6.[18]

Reading the failure log: error codes that tell you which problem you have

Your message log already sorts failures into list problems and sender problems. Twilio's error codes are the clearest public taxonomy:

After we fixed the filtering on one event send (the carrier-filtering post-mortem has the details), the failures left over were list problems: inactive or out-of-service numbers, landlines, countries the sender wasn't set up for, and app-only numbers. Those map straight onto 30003, 30005 and 30006.

Lookups can catch most of these before you send. Twilio's Line Type Intelligence returns 12 line types, including mobile, landline, fixed VoIP, non-fixed VoIP, toll-free and pager.[24] Twilio has also launched a line-status check for whether a number is active, and claims 99.1% accuracy for US numbers without publishing a method.[25]

Our list-validation funnel

When a client has texted the same list for years without validating it, we run it through three stages before anything is sent:

1. Raw records. Every contact row with a phone field, including duplicates across forms, imports and events.
2. Unique numbers. After normalizing formats and deduplicating.
3. SMS-ready numbers. After a line-type lookup removes landlines, invalid and unreachable numbers, and after opt-outs are suppressed.

In our experience, only about 40% to 45% of raw records survive to the last stage. No public industry figure exists to compare that with, so treat it as our rule of thumb from list cleans.

Put that next to the CDC's 1.6% landline-only figure and the gap is obvious. The population hasn't gone back to landlines. Lists decay, people type the wrong number, and duplicates pile up across every import.

Validation vs verification vs consent

These three get blurred constantly, and they answer different questions:

• Validation (a line-type or line-status lookup) tells you a number can receive texts.
• Verification (a one-time passcode at opt-in) tells you the person filling in the form controls that phone.
• Consent (the disclosure and agreement on your form) tells you that person agreed to hear from you.

A verified number with no consent is still a number you shouldn't text. Our guide to SMS consent, registration and verification covers how to capture all three on one form.

Numbers to stop repeating

• "SMS has a 98% open rate." An estimate with no traced measurement.[15]
• "35 million numbers are reassigned every year." The FCC said disconnected and made available, in 2018.[2]
• Any single "average SMS delivery rate". No independent source publishes one.
• "Phone verification proves consent." It proves control of the phone, nothing more.