What the 2025 NSDUH Shows About Addiction in America

The 2025 Survey Results Are Out

The National Survey on Drug Use and Health is the federal government’s annual count of substance use, substance use disorders, mental health, and treatment. SAMHSA interviews tens of thousands of Americans aged 12 and older for it every year, and it’s the source behind most addiction statistics you’ll see cited anywhere. The 2025 results were published in July 2026^[1].

I spent a full day in this report, and the new numbers resist a simple headline.

Four findings organize everything else:

• Alcohol use disorder, binge drinking, cigarette smoking, and underage drinking all declined since 2021.
• Marijuana use and nicotine vaping increased, largely among adults.
• Most people who needed treatment did not receive it, and the survey points to perceived need as a major reason.
• Recovery remains common, and it gets far less attention than the problem does.

Alcohol Still Accounts for the Most Substance Use

The report shows that in 2025, 162.8 million people, 56% of everyone aged 12 and older, used tobacco, vaped nicotine, drank alcohol, or used an illicit drug in the past month^[2].

One substance dominates that count^[2]:

• Alcohol: 129.1 million past-month drinkers
• Illicit drugs: 47.5 million
• Tobacco: 46.9 million
• Nicotine vaping: 29.3 million

America’s substance story is still, first and by a wide margin, an alcohol story. That’s why alcohol produces the largest single disorder count every year, and why the alcohol trend below matters more than its lack of drama suggests.

What Improved Since 2021

These are the declines SAMHSA itself flags as statistically significant against its 2021 baseline^[2]:

Alcohol Use Disorder Fell From 10.6% to 8.9%

That drop equals four million fewer Americans meeting criteria for alcohol use disorder in 2025 than in 2021^[2]. Binge drinking declined in every age group over the same window, and cigarette smoking fell to 12.4% of the population^[2].

The survey does not tell us why. Possible contributors include generational drinking patterns and shifting habits after the pandemic years, but NSDUH measures prevalence, not causes. What it establishes is the direction: on the substances that account for most of America’s addiction burden, the numbers are moving the right way.

Substance Use Declined Among Adolescents

The turn is sharpest among the youngest^[2]:

• Underage drinking fell from 15.6% to 13.7%.
• Underage nicotine use declined.
• Past-year major depressive episodes among adolescents dropped from 20.8% in 2021 to 15.1% in 2025.
• Serious mental illness among adults declined from 7.9% to 6.9%.

These findings deserve more attention than they get. On measure after measure, teenagers are using less than they were four years ago.

What Increased

Four measures moved the other way, each flagged as a significant change. The baseline years differ because the survey tracks its newer measures from 2022^[2]:

Adults Drove the Increases

The report is specific about who is behind the rises. The marijuana increase came entirely from adults 26 and older; use among adolescents and young adults declined^[2]. Vaping among young adults fell even as the national number climbed, and hallucinogen growth also concentrated in the older group. More on the vaping pattern is at nicotine addiction.

One possible explanation is that these substances are becoming more normalized among adults, through legalization, marketing, and changing social norms. The survey does not measure attitudes, so it cannot confirm why use rose. What it does establish is the shape: for decades the worry ran young, and in this data the increases belong to adults while teenagers pull back.

Suicide Plans and Attempts Rose Among Adults

Serious thoughts of suicide held flat among adults, but plans rose from 1.4% to 1.8% and attempts from 0.7% to 0.9%^[2]. A stable rate of suicidal thinking alongside more plans and attempts is the most concerning combination in the report, and the survey offers no explanation for the gap.

If you’re having thoughts of suicide, call or text 988 (the Suicide & Crisis Lifeline). It’s free, confidential, and available 24/7.

Why the Fentanyl Estimate Is Likely Too Low

The report shows 768,000 people (0.3%) used illegally made fentanyl in the past year of 2025^[2]. Next to fentanyl’s role in overdose deaths, that number looks impossibly small.

Important limitation: SAMHSA attaches the caution itself. Illegally made fentanyl is often present in counterfeit pills and heroin without the user’s knowledge, so self-reported use understates real exposure^[2]. A survey can only count the people who know what they took.

The overdose data counts outcomes rather than awareness, and that’s where fentanyl’s footprint actually shows. The fentanyl statistics cover that side.

Most Substance Use Disorders Were Classified as Mild

The majority of people with a past-year substance use disorder in 2025 had a mild one, and about 1 in 5 had a severe one^[2].

Severity Was Highest for Stimulant and Opioid Disorders

The exceptions run exactly where you’d expect. Among people with a disorder involving stimulants (cocaine, meth, or misused prescription stimulants), 47.9% were classified severe. For opioid use disorder (heroin or misused prescription opioids), 38.4% were severe^[2].

Why the Mild Majority Matters

The popular image of addiction describes the severe minority, not the typical case. Most people who meet criteria for a substance use disorder in this data have a mild disorder.

What this does not mean: mild is not the same as harmless. A mild disorder is a diagnosis, it can progress, and the substances whose disorders skew severe, stimulants and opioids, are the ones where waiting carries the highest cost.

Only 1 in 6 People Who Needed Treatment Received It

Among everyone classified as needing substance use treatment in 2025, 16.0%, or 7.6 million people, received it^[2].

Most Untreated Adults Did Not Perceive a Need

Of the 38.1 million adults with a substance use disorder who got no treatment, 94.5% did not seek treatment or believe they needed it. Widening the lens, 86.4% of everyone with a substance use disorder, 38.5 million people, received neither treatment nor any other kind of help that year^[2].

What this means: the survey suggests that recognition and perceived need are a major part of the treatment gap, alongside whatever happens after someone decides to seek care.

What this does not mean: it does not establish that cost, insurance, availability, stigma, or distance are unimportant. The survey measures whether people perceived a need, not every barrier they would have hit once they looked. Someone who has already concluded treatment is unaffordable or out of reach may also report not seeking it.

Co-Occurring Conditions Often Went Untreated

Of the 18.5 million adults with both a mental illness and a substance use disorder in 2025, about 2 in 5 received treatment for neither condition^[2]. Clinicians call that combination a dual diagnosis, and the standard of care is to treat both together.

The Survey Added an Anxiety Measure

Beginning with this report, NSDUH includes the GAD-7 screening scale. In 2025 it found 17.5 million adults (6.6%) with moderate or severe symptoms of generalized anxiety disorder in the past two weeks, with the highest rate among young adults 18 to 25 (12.2%)^[2].

There is no trend line yet; the measure is too new. It matters anyway, because anxiety is a condition many people report managing with alcohol or other substances, and until now the survey that tracked the substances did not measure it. The survey itself does not connect the two; it now simply watches both. The broader picture is in the mental health statistics.

Most Adults With a Past Problem Say They Are Recovering

In 2025, 30.6 million adults (11.6%) said they had ever had a problem with drugs or alcohol. Of that group, 73%, or 22.3 million people, consider themselves to be in recovery or to have recovered^[2].

The same pattern holds for mental health: 69.4% of the 63.8 million adults who ever perceived a mental health problem consider themselves recovering or recovered^[2].

Important limitation: these are self-assessments, not clinical verifications. The survey asks people where they stand, and this is what they report. Even read conservatively, recovery is a common outcome in this data, and it is significantly underrepresented in how addiction gets discussed. The fuller picture, including what’s known about how people get there, is in the addiction recovery statistics.

Why These Numbers Don’t Compare to Older Ones

NSDUH changed how it collects data in 2021 and moved to updated diagnostic criteria. SAMHSA is explicit that 2025 estimates compare only with 2021 through 2024, never with 2020 or earlier^[1]. A chart drawing a line from 2019 to today is comparing two different rulers, whatever its source.

The treatment questions changed again in 2024, so the 1-in-6 figure stands alone and cannot be trended against any earlier year^[2].

Statistics on AddictionHelp.com carry their own year and are never stitched across those breaks, which is why some figures here differ from older articles elsewhere. Different rulers, honestly labeled.

Where to Find the Full Breakdowns

Every figure above comes from the federal source documents and is year-stamped.

The substance-by-substance detail lives on the statistics pages:

• Addiction statistics covers the national picture in one place
• Alcohol statistics tracks the biggest single disorder and the biggest decline
• Marijuana statistics follows the fastest-growing use numbers
• Nicotine addiction statistics shows cigarettes down and vaping up
• Fentanyl statistics covers the exposure the survey can’t self-report
• Mental health statistics spans mental illness, suicidality, and the new anxiety measure
• Addiction recovery statistics tells the recovery story in full
• The most commonly used drugs ranks substances by real prevalence

If Any of This Applies to You

The data section ends here. This part is practical.

If the treatment-gap numbers made you wonder about your own situation, that wondering is worth taking seriously. The survey’s clearest finding is that most people with a diagnosable problem never conclude they have one.

A quick, private self-assessment is a reasonable first step. Treatment options explains what care actually looks like, level by level, and free support groups exist in nearly every community for you or your family.

The numbers in this report will change again next year. What doesn’t change is the person reading them at 1 a.m., wondering whether their situation counts. It counts.