Creating the Meditation Market

What we are trying to do

Put the world's contemplative traditions within reach of anyone who wants them.

The Sacred Act exists to make contemplative practice accessible enough that a person can actually have one, without needing to find the right teacher in the right city at the right time. We call that spiritual sovereignty: the practices of the world's traditions available to the person who wants them, on their own terms, in the moment they need them.

Interest is already near-universal. So is information. What is missing is guidance.

Fifty-one million people who want a practice. Eight thousand people trained to give one. That ratio is the problem, and it has never been solved, because a teacher's hours cannot be multiplied. Serving that intent pool one to one would take about two million teachers.

The second thing is happening right now. AI is being built into the devices people already own. Apple reported more than 2.5 billion active devices in January 2026, and rebuilt its assistant layer on frontier models in June. Whatever that layer can do becomes what a couple of billion people can do, without ever deciding to go and look for it.

Contemplative capability belongs in that layer. A person should be able to ask for a practice and receive the right one — a specific technique, from a real tradition, matched to the state they are actually in — with the same immediacy they now get directions or a translation. That is the product we are building, and it is why the timing matters: the distribution for this is being laid down across every device on earth over the next few years, and what rides on it is being decided now.

Making practice reachable this way creates a market. The people it reaches have wanted a practice for years and have had no way to get one that held them. Nobody is currently serving them.

What this report argues

Guidance is the constraint on this category. Removing it builds a new market.

The claim above is the conclusion. This report is the evidence for it.

It began as sixteen questions about our own market, and the answers pointed somewhere we did not expect. Almost every published figure on meditation measures the same small population: the people already practicing and already paying. One representative national survey measures the people standing just outside, and that turned out to be the largest number in the whole study — larger than the market itself.

Ten sections get from that number to the conclusion. One through three establish that meditation is a mainstream behaviour with a small economy, and that the money inside it goes to teachers' hours and objects. Four explains the price: the best-measured customer in the category is buying a sleep aid, and a remedy is a commodity. Five and six locate the failure — people leave in the first month, at the exact point a teacher would step in, and there are almost no teachers. Seven through nine show what has happened twice in other categories when a scarce expert became software, and how little of it has reached meditation. Ten counts the people looking. The final sections put a number on what closing the gap is worth, and state the case in six lines.

The number we arrive at: closing this gap creates $100M to $500M of annual consumer spending in the United States alone, against a global paid-app category of roughly $2 billion. That is created spending, and the model behind it is shown with its assumptions exposed.

A note on interest. We make one of the products in this report. Our own app appears once, in the audit in section nine, where its row is marked and describes the app as built. The app is not yet generally available. Everywhere else we report the numbers as we found them, including the ones that argue against us. The most important of those is in section three: willingness to pay for meditation content is low, and cost is the most-cited barrier in the category.

A note on sources. Every figure sits in one of four tiers, and the tier travels with the figure. Government surveys, peer-reviewed journals and direct company disclosures are cited freely. Market-report vendors are cited only to show that they disagree with each other. One widely repeated statistic is named at the bottom and refused. Our own model is labelled a model on its own face. The full apparatus is folded shut at the end of this page.

1. Meditation is now ordinary

Almost one in five American adults meditated last year. The rate has more than doubled in twenty years.

7.5%

4.1%

14.2%

18.3%

2002

2012 †

2017

2022

The 2012 reading is shown at reduced strength because it is widely treated as anomalous — it sits below the 2002 figure in a series that otherwise rises without interruption, and the questionnaire changed between cycles. A second team reading three of the same cycles put 2022 at 17.3%; the difference is population weighting. The paper pairs its 18.3% with a count of 60.53 million people; section one explains why that count uses the whole-population base.

The firmest numbers here come from the CDC's National Health Interview Survey, read by two independent teams in 2024. One, in Scientific Reports, weighted five NHIS cycles across 134,959 respondents and found that 18.3% of US adults used meditation in 2022, up from 7.5% in 2002. The other, in JAMA, read three cycles and put 2022 at 17.3%; the National Center for Complementary and Integrative Health adopted that figure. The difference is population weighting.

One correction, since this figure is quoted constantly. The Scientific Reports paper pairs its 18.3% with a count of 60.53 million people. It reached that count by applying an adult prevalence to the 2020 census estimate of the whole population, around 331 million. Applied to the US adult population of 267 million, the same 18.3% is about 49 million adults. This report uses percentages, and states the base wherever it gives an absolute number.

Women meditate more than men, 16.3% against 11.8% in the 2017 cycle. College graduates meditate at roughly twice the rate of adults without a high-school diploma. Practice is most common in the western United States. In the 2022 data the most represented practitioners were between 25 and 44, female, more educated, and western.

No federal survey measures frequency. For that, the evidence comes from elsewhere.

2. The biggest group in meditation is the one nobody serves

Fifty-one million American adults intend to meditate and have no practice. Thirty million have one.

Pre-Intention

69.7% · 186.1M

Not meditating and not planning to

Preparation

19.1% · 51.0M

Intend to meditate. Are not meditating.

Action

2.3% · 6.1M

Practicing, under six months in

Maintenance

9.0% · 24.0M

Practicing, established habit

Absolute figures apply each share to a US adult population of 267 million (Census Bureau, 2024). The second row is the finding: roughly 51 million American adults hold an intention to meditate and do not have a practice — more than one and a half times the number who do. Stages sum to 100.1% through rounding.

The figure comes from a representative Ipsos KnowledgePanel survey of 2,000 US adults, mapped onto a stage model of habit formation: Pre-Intention for people not meditating and not planning to, Preparation for people who intend to and are not, Action for those under six months into a practice, Maintenance for those with an established one.

The people who intend to practice outnumber the people practicing by better than one and a half to one. Every meditation app on earth competes for the 11.3%. No product serves the 19.1%.

The behavioral funnel underneath shows the same thing.

Half the country has tried this. A seventh is doing it now. Around 100 million American adults have tried meditation and have no practice today. They are lapsed. People arrive willing, and something in the experience fails to hold them.

Cost

42.4%

Doubt that it works

35.5%

Time it takes

21.2%

Not interested enough

20.5%

Respondents could name more than one, so these do not sum to 100%. Read against the field evidence, the ranking is more interesting than it looks: in a randomized experiment across 1.33 million Spotify Premium users, messaging that removed practical friction lifted activation most (odds ratio 1.55 against control), messaging about benefits lifted it less (1.38), and messaging that addressed not knowing how to meditate underperformed the control condition outright (0.66). Messaging that removed practical work lifted activation. Messaging that reassured people performed worse than saying nothing.

The field evidence on which barrier moves is worth sitting with. In a randomized experiment across 1.33 million Spotify Premium users, messages that removed practical friction — the quiet room, the time, the logistics — lifted activation by half. Messages reassuring people that they did not need to know how to meditate correctly performed worse than saying nothing at all.

Knowledge is still a barrier. Reassurance does not remove it. Telling someone they cannot get it wrong leaves them without instructions. What moved people was removing the work.

3. The money goes to teachers' hours and objects

Seven dollars in ten never reach software, and the ones that do buy a library.

Sort the category's spending by what it purchases. Almost none of it buys instruction that arrives when a person sits down.

Global mental wellness market

$181B

of which meditation and mindfulness$7.1B · 3.9%

Meditation and mindfulness spending

$7.1B

of which paid apps~$2B · 28%

Each row rescales to the shaded part of the row above it: the sliver in the first bar is the whole of the second. Read downward, paid meditation apps are about one percent of what the world spends on mental wellness. The $2B app figure is the softest number in the chart — no one publishes it directly, and it is assembled from subscriber counts and list prices.

Start with the size. Meditation and mindfulness accounts for under four percent of what the world spends on mental wellness, and paid apps take under a third of that. A category with hundreds of millions of participants monetizes at about the scale of a mid-sized software company.

The composition matters more, and nobody publishes it.

Paid apps

~$2.0B · 28%

Everything that is not an app

~$5.1B · 72%

Human access that does not scale

Classes, teachers, retreats, courses, teacher training

Objects

Books, cushions, beads, chimes, journals, coloring books, trackers and biofeedback gadgets

Listed as GWI defines the segment. Unsized: a bar chart here would be numbers we invented.

GWI does not publish a public breakdown of the $7.1 billion by category. The split above is the app estimate against a residual, so both figures are estimates. What GWI counts inside the segment is documented, and it is listed beneath. Roughly seven dollars in ten go to a human being whose hours cannot scale, or to an object that sits in a room. Very little buys instruction that arrives when a person needs it. GWI’s own summary of the segment: “while millions of people worldwide practice meditation, only a small fraction spend money on it.”

GWI counts the whole segment: classes and teachers, retreats, courses and teacher training, books, cushions, beads, chimes, journals, coloring books, and biofeedback gadgets. That list falls into two piles. One is human access that cannot scale — the teacher's hour, the retreat's bed, the course's cohort. The other is objects — things you buy once and put in a room. Roughly seven dollars in ten go to one pile or the other.

The category has monetized scarcity and objects. It has never sold guidance at scale. A cushion supplies no instruction. A retreat supplies four days of it a year, to the few hundred people who can attend. The rest of the year a person is alone with a library.

Few people pay for the software.

Meditated at least once in the past year · US

48.9M

18.3% of US adults, NHIS 2022

Have an established or new practice · US

30.2M

11.3% of US adults — the Action and Maintenance stages

Pay for Calm or Headspace · worldwide

6–7M

The two largest paid meditation apps, combined

The first two bars are United States only. The third is worldwide and still a fifth of the bar above it. The paying population of the two largest meditation apps on earth is a fraction of the Americans who already have a practice, and a smaller fraction of the fifty-one million who want one. Both US figures apply survey percentages to a US adult population of 267 million.

Calm has an estimated 3.5 to 4.5 million paying subscribers, down from a peak near five million in 2022, at $14.99 a month or $69.99 a year. Headspace has around 2.8 million. Those two are the only meditation apps clearly at that scale; every other paid subscriber count in the category is undisclosed or small. Download counts are much larger — Calm and Headspace have each passed a hundred million cumulative installs — and they are a vanity number. Free-to-paid conversion in this category runs around four to six percent.

What subscribers say the software is worth is the number that argues hardest against our own thesis.

Worth it to a once-a-week meditator

$4.90 / mo

Worth it to a near-daily meditator

$5.90 / mo

Calm, billed annually

$5.83 / mo

$69.99 a year

Calm, billed monthly

$14.99 / mo

Waking Up, billed annually

$10.83 / mo

$129.99 a year — 1.9× Calm's annual rate, and the strongest 90-day retention in the category

One axis, dollars per month. The lighter bars are what users said a meditation subscription is worth to them; the darker bars are what the category leader charges. The annual plan lands almost exactly on the daily practitioner’s stated ceiling, and the monthly plan asks two and a half times it. Stated willingness to pay is a soft instrument — people under-report in surveys — and the shape of the gap survives the caveat. The last bar complicates the story. Waking Up charges nearly twice Calm’s annual rate, sells depth, and holds its users better than anyone in the category. The $5.90 ceiling was measured on Calm and Headspace subscribers. It measures what a content library is worth.

When 5,391 Calm and Headspace users were asked what a meditation subscription is worth, weekly meditators said $4.90 a month and near-daily meditators said $5.90. Calm's annual plan works out to $5.83. The category leader is priced at the ceiling of what its most committed users say the content is worth, and its monthly plan asks two and a half times that. Cost is also the most-cited barrier in the figure two sections up, named by 42% of meditators.

That does not settle the price question. Those numbers measure willingness to pay for a library of recordings, which is the only thing the respondents had ever been sold. The same population pays $420 to $980 for four sessions of personal instruction, as sections four and six show. Price sensitivity here reflects what is on offer. It remains a real risk: any product asking a premium has to earn it inside the first session, or it will be priced like the shelf it sits next to.

4. The category sells a remedy, and remedies are commodities

The category's best-measured customer is a 48-year-old woman treating insomnia. That fact sets the price of everything.

There is one good dataset on who pays for meditation software: a cross-sectional survey of 12,151 paying Calm subscribers, published in JMIR.

	The Calm subscriberMeasured, n=12,151	The practitionerHypothesis — never surveyed
What they are buying	Relief from a symptom	A practice they can develop
Top reason for opening it	Sleep — 63%, ahead of stress at 62%	To build a practice
Median age	Mean 48.6 · 45–64 is the largest band at 44.5%	Younger — SBNR skews under 50
Health context	77% report sleep difficulty; 41% carry a mental health diagnosis	Not primarily clinical
What it competes with on price	Melatonin, a white-noise machine, a sleep podcast	A teacher, a course, a retreat
What they pay	$69.99 a year, and they say it is worth $5.90 a month	$129.99–$199 a year; $420–$980 for four TM sessions
What makes it work	They slept	The practice fit them, and it progressed

The two columns carry different weight. The left column is measured — a peer-reviewed survey of twelve thousand people who were paying for the product at the time they answered. The right column is assembled from adjacent populations who have demonstrated the behaviour separately: Waking Up’s subscribers at nearly double Calm’s price, Transcendental Meditation’s course buyers, coaching clients, and the spiritual-but-not-religious population in section ten. Nobody has surveyed a premium personalized meditation buyer, because the product they would be buying does not exist yet. Treat the right column as the hypothesis this report is testing.

The subscriber is older than the meditating population at large — a mean age of 48.6, with the 45-to-64 band the largest single group at 44.5% and only 2.5% under 25. She is overwhelmingly female, at 80%. She is affluent: 39% report household income over $100,000, and separate polling finds a comparable skew among Headspace users. She is often unwell. Seventy-seven percent report difficulty sleeping. Forty-one percent carry a mental health diagnosis — a third with anxiety, a quarter with depression. Fifty-seven percent have some chronic diagnosis.

What she says she is there for: sleep, at 63%, is the top reason, ahead of stress at 62%, depression and anxiety at 54%, and general health at 40%. Fifty-six percent use Sleep Stories.

This is a remedy purchase, and remedies are priced against other remedies. The alternatives to a sleep app are melatonin, a white-noise machine, a podcast, an audiobook: all cheap, all one-off, none asking for a relationship. A product competing in that set will be priced at five or six dollars a month however good it is, because that is what the neighbouring shelf costs.

The $4.90-to-$5.90 ceiling from the previous section is a fact about sleep aids. It was measured on Calm and Headspace subscribers, who are buying one.

One product in the category takes a different position. Waking Up sells consciousness, philosophy and instruction. It charges $129.99 a year, close to double Calm's rate, and third-party analysis puts its 90-day retention ahead of everyone else in the category. It is a smaller business than Calm by an order of magnitude. It is also the only evidence inside software that a different buyer exists and pays more.

The rest of the evidence for that buyer sits outside software, and this report has already counted it: the people paying $420 to $980 for four Transcendental Meditation sessions, the ones paying $234 an hour for a coach, the 22% of American adults who describe themselves as spiritual but not religious. All of them are buying access to a person who knows something.

The limit on this section. Everything in the right-hand column of that table is inference. There is no survey of the premium personalized meditation buyer, because there is no premium personalized meditation product to have surveyed them with. What can be said is narrower and still useful: the population that pays a premium for contemplative instruction exists, it has been paying for decades, and it has never been offered software.

5. The drop-off happens exactly where a teacher would be

Nineteen in twenty people who install a meditation app stop opening it within a month.

Day 1

69.4%

of people who installed opened it

Day 15

3.9%

still opening it

Day 30

3.3%

still opening it

Median retention across 93 mental health apps measured from an independent usage panel rather than from anything a developer chose to publish. Mindfulness and meditation apps do marginally better than the field at day 30 — a median of 4.7% against 3.3% — and hold their users longer per session, at a median 21.5 minutes a day against 13.0 across all categories. The product people download does not keep them.

This is measured from an independent panel of real usage across 93 mental health apps. More than eighty percent of the decline happens in the first ten days. Whatever the download numbers say, the product fails its users immediately and at scale.

The same pattern shows up away from software, in the only population-level reading of what happens to people after they learn a practice.

• Still practicing regularly25%
• Practicing occasionally32%
• Stopped entirely43%

The best population-level attrition reading available anywhere. It is British; no US survey isolates continued practice. Two in five people who learned a practice no longer do it. The quarter who continue are the population any subscription business is selling to.

Two in five stop entirely. A quarter keep going. The failure happens immediately after arrival, in the stretch where a beginner needs someone to tell them what to do next and there is no one there.

6. People pay ten times more for a teacher, and teachers cannot scale

Eight thousand certified teachers. Fifty-one million people who want one.

Meditation app, one year

$69.99

a year, unlimited sessions

Private meditation teacher

$75–150

an hour, one to one

Professional coach, global average fee

$234–244

an hour, one to one

Transcendental Meditation course

$420–980

four sessions, then follow-up for life

One axis, dollars per transaction. The transactions are not alike. The darker part of each bar is the bottom of the published range and the lighter part is the top. The smallest number here buys twelve months of a library; the largest buys four sessions with a human being and a teacher for life. Consumers pay roughly ten times more for the second thing.

Transcendental Meditation is the cleanest proof in the category. TM sells four sessions of one-to-one, personalized instruction on an income-based scale from $420 to $980, including follow-up for life; the fee was once as high as $2,500. It has taught over five million people since the 1960s. A TM course carries less content than a meditation app. It sells a person who assesses you, gives you a specific practice, and stays reachable.

The lapsed hundred million never got that product. It does not exist in anything like the required quantity.

Teachers the US intent pool would need

2.04M

51 million people, one guided session a week

Every credentialed coach on earth, of any kind

122,974

ICF Global Coaching Study 2025 — most of them do not teach meditation

Teachers certified by the largest meditation program

8,000

Mindfulness Meditation Teacher Certification Program, 70+ countries

The arithmetic, in full: 51 million intenders at one guided session a week for 48 weeks is 2.4 billion sessions a year. A teacher working 25 one-to-one hours a week for 48 weeks delivers 1,200. That is 2.04 million teachers, for one country. The two lower bars are drawn at a minimum width because at true scale they would be invisible — the entire global coaching profession is 6% of the bar above it, and the certified meditation teachers are four tenths of one percent. Guidance is expensive because there is almost none of it.

There is no Bureau of Labor Statistics occupation code for "meditation teacher." The closest official proxy, exercise trainers and group fitness instructors, counted 370,100 jobs in 2024 at a median of $46,180. The largest certification program in the field has produced just over 8,000 teachers across seventy countries. The entire credentialed coaching profession worldwide — of every kind, most of it nothing to do with contemplative practice — is 122,974 people generating $5.34 billion, and 53% of them earn under $30,000 a year from it.

The supply side fails in both directions. There are nowhere near enough teachers for the people who want guiding, and the ones who exist mostly cannot make a living at it. Guidance costs what it costs because there is almost none of it. Software is the only mechanism that has ever changed that ratio for anything.

7. Spending is growing far faster than the number of people practicing

Spending grew 18.9% a year. Participation grew about five.

The most useful growth number here is the gap between two measured rates.

Consumer spending on meditation and mindfulness has grown 18.9% a year, to $7.1 billion, according to the Global Wellness Institute, the one body here that publishes its methodology. Over a comparable window, the share of US adults who meditate went from 14.2% in 2017 to 18.3% in 2022: about 5% a year.

Spending is growing close to four times faster than the number of people practicing. The category is deepening while the population barely moves. That is what a market looks like when it serves its existing customers adequately and has no mechanism for bringing in new ones.

For the wider frame: the global wellness economy reached $6.8 trillion in 2024 and is forecast to $9.8 trillion by 2029. Mental wellness is its second fastest-growing sector at 12.4% a year, reaching $181 billion in 2022 with a projection to $330 billion by 2027; the US alone accounts for $125 billion. Hold that against the spend stack in section three: people direct under four percent of their mental wellness money at meditation, the practice with the deepest evidence base and the longest lineage in the category.

The other growth figures in this category should be ignored.

Persistence Market Research

9.9%

Grand View Research

14.67%

Global Wellness Institute

18.9%

Consumer spending on meditation and mindfulness — a different quantity to the four around it, and the only one from a body that publishes its method

Research and Markets

29.6%

Technavio

55%

0%20%40%60%

Four vendors, sizing the same market, disagree by a factor of 5.6. That spread is itself the finding — the category is poorly measured, and any single CAGR quoted as fact is quoting one vendor’s model. We cite none of them as authority. The GWI figure is set apart because it measures consumer spending. It does not substitute for the others.

We have read the meditation-app market-size reports. They are low quality, they conflict wildly, and they are the source of nearly every growth statistic quoted in this category's pitch decks. We cite none of them as authority, and anyone reading a meditation-app CAGR should ask which of these five numbers it came from.

8. When the expert becomes software, the market gets built

Four in five of the people who showed up had never been in that market at all.

McKinsey's research across two studies is the most-cited work on personalization. It most often drives a 10 to 15% revenue lift, with company-specific results spanning 5 to 25%; it can cut acquisition costs by up to half and lift marketing return by 10 to 30%. The finding that matters most is comparative: companies that grow faster take 40% more of their revenue from personalization than their slower-growing peers. Seventy-one percent of consumers expect personalized interactions and 76% are frustrated without them.

This has now happened twice in categories built like this one.

Language learning

Language learning had the same structure meditation has now: a real practice, a large latent population, and a delivery model that only worked if you could afford a human teacher. As of the third quarter of 2025 Duolingo reports 50.5 million daily active users and 11.5 million paying subscribers, with bookings past a billion dollars for the year. The category's previous leader, Rosetta Stone, never passed about $270 million in annual revenue in its best years before free mobile apps took the ground out from under it.

The users came from outside the market. Duolingo has reported, from its own survey at the time of its 2021 IPO, that roughly four in five of its US users were not learning a language before they arrived. It built a market out of people who wanted the thing all along and had no affordable way to get it.

We flag that figure. It is quoted consistently across secondary coverage of the S-1 and attributed to Duolingo's own survey, but the filing would not serve to us and we have not read the sentence in the original. Treat it as strong and unverified.

Software itself

The second case is happening now, and it is larger. Writing software was the most expert-gated skill in the economy: it required years of training, or the money to hire someone who had them. Since the beginning of 2025 that gate has been dismantled by tools that let a person describe what they want and receive working software — Cursor, Claude Code, Replit, Lovable, Vercel's v0, GitHub Copilot.

What happened to the population is the point.

Most of the people arriving were never going to be engineers.

Duolingo †

~80%

were not learning a language before they arrived

Lovable

~80%

building with it have no technical background

Vercel v0

63%

of its 4M+ users are not developers

These are the only three figures we could find that measure the thing directly, and all three are company-published — read them with that in mind. The Duolingo bar is drawn at reduced strength because we could not retrieve the primary filing; it is consistently reported and attributed, and we have not read it in the original. What survives the caveats is the shape: when the expert a practice required becomes software, most of the people who show up were never in the market.

Lovable reports that about 6% of the people building on it are engineers, and that its largest single group is founders, at 45.7%. Replit's chief executive has said publicly that the company is no longer building for professional coders; its revenue went from $10 million at the end of 2024 to $150 million by September 2025, at which point Bloomberg reported a $3 billion valuation. Cursor passed $500 million in annual recurring revenue at a $9.9 billion valuation inside three years.

The four in five is the number that matters. It is the same fraction as Duolingo's, in a different category, for the same reason. The practice required a scarce expert, and most people could not get one.

The structure they share

A latent population that already wants the thing. Near-total awareness with no entry path. A scarce, expensive human expert as the only reliable way in. And a tool that supplies the expert's availability.

Meditation has that structure exactly, and section six is the proof: 51 million people who intend to practice, against 8,000 certified teachers.

The disanalogy. Code has an objective pass condition — it runs or it does not — and that condition is what let AI coding tools improve as fast as they did. A meditation has no such signal. Whether a practice landed is subjective, delayed, and easily confused with mood. Any product claiming to do for contemplative practice what Cursor did for software has to manufacture the feedback loop that software got for free, out of completion, self-report, and physiological measurement. That part of the analogy does not transfer.

9. Personalization in this category stops at the signup questionnaire

Several apps read your health data. None uses it to decide what you practice.

Product	Questionnaire at signup	Ongoing state check	Biometric input	Generated audio	Prescribes a named technique from a source text
Calm
Headspace
Insight Timer
Balance
Happier
Muse
Endel
Vital
The Sacred Act

present partial absent

Five columns; on a narrow screen the table scrolls sideways, and the rightmost column is the one that matters. Filled is present, half is partial, an outline is absent. Partial means the capability exists but does not drive what the practitioner is given — Calm and Insight Timer read Apple Health’s mindful minutes, for example, but do not prescribe from them. We build one of the products in this table. Its row describes the app as built, and the app is not yet generally available. This audits what is visible from outside.

The landscape sorts into three tiers and a gap.

Questionnaire recommendation. Calm, Headspace, Insight Timer and Balance all ask a short set of questions at signup and use the answers to sort an existing library. Balance markets itself most aggressively on this. Happier, rebranded from Ten Percent Happier in 2024, goes furthest by checking in monthly about mood, need and available time. This is recommendation. It decides which shelf to show you.

Biometric input. Very limited. Several apps read Apple Health's mindful minutes, a record of what you already did. Muse, which sells hardware, provides real-time biofeedback and has added AI-driven personalized feedback on top of it.

Generated audio. The newest tier and the fastest moving. Vital generates spoken sessions on demand from what a user types. MeditAItion, Cerebral AI and Wondercraft generate custom meditations. Endel produces adaptive soundscapes that respond to activity and biometric data. Insight Timer has added AI-generated soundscapes. These systems generate audio; none diagnoses a state and matches it to a structured body of technique.

The gap. We could not find a product that combines all three: an assessment of state, temperament and context; biometric input; and prescription from a structured corpus of techniques drawn from primary sources, rendered as generated audio for that person. The traditions worked this way. A teacher assessed a student and assigned a practice, and the practice was a specific named method with its own mechanism and its own contraindications. Apps supply a catalogue instead. The retention curve in section five measures the result.

10. A large population is looking for a practice without an institution

Seven in ten Americans call themselves spiritual. Three in ten belong to no religion, and they skew young.

Pew's 2023 study of 11,201 American adults found that 70% describe themselves as spiritual in some way, including 22% who are spiritual but not religious. The religiously unaffiliated are now 28 to 29% of US adults, larger than Catholics or evangelicals, and they skew sharply young.

The counterweight is in Pew's own reading of its data. The spiritual-but-not-religious share rose from 19% in 2012 to 27% in 2017, then settled back to 22% in 2023, and Pew notes evidence that Americans are becoming somewhat less spiritual as well as less religious. This population is large and durable. Its growth is flat.

The 51 million intenders fit this picture. Seven in ten Americans call themselves spiritual and three in ten belong to no religion: a large number of people looking for a practice, with no institution to hand them one.

Creating this market is worth $100M to $500M a year

Three percent of the intenders is $304 million a year that does not exist today.

The question is how much bigger this market becomes if someone solves the guidance problem. We built the answer two ways, so the two can be checked against each other.

Conservative · 1% of intenders

$36M–$101M

0.51 million new paying practitioners

Base · 3% of intenders

$107M–$304M

1.53 million new paying practitioners

Upside · 5% of intenders

$178M–$507M

2.55 million new paying practitioners

This is a model, not a forecast. Each bar is one assumption made explicit: what share of the 51 million American adults who intend to meditate could be converted into a sustained paying practice by a product that removes the barriers in the figures above. The dark part of each bar prices them at today’s category rate of $69.99 a year; the lighter part extends to $199, the rate a personalized product can ask. The base case creates $304 million of annual spending that does not exist today — from one country’s intenders — against a global paid-app market of roughly $2 billion. The upside case is a quarter of that entire global market, and all of it is new.

The bottom-up model, in full. 51 million US adults hold an intention to meditate and have no practice. Convert 1%, 3% or 5% of them into a sustained paying practice and you get 510,000, 1.53 million or 2.55 million new paying practitioners. Priced at today's category rate of $69.99 a year, that is $36M, $107M or $178M in annual consumer spending. Priced at $199 — what a personalized product can credibly ask, and roughly a fifth of what TM charges once for four sessions — it is $101M, $304M or $507M. All of it is new. None of it is taken from Calm.

For scale: the entire global paid meditation-app market is around $2 billion. The base case adds 15% to it, from one country's intenders alone. The upside case adds a quarter.

The top-down check. Meditation and mindfulness is $7.1 billion of a $181 billion global mental wellness market, or 3.9%. If solving activation moved that share to 6%, the category gains $3.8 billion worldwide. At 10%, it gains $11 billion. Those are share-shift thought experiments and we hold them loosely. They bracket the bottom-up model from above: our US base case is roughly a twentieth of the low end of the global top-down figure, which is what one country and one product category should look like. The bottom-up model is the conservative one, and it is the one we would defend.

What the model assumes, so it can be argued with

That the 19.1% Preparation figure holds. It comes from one representative survey of 2,000 adults, and it is the single load-bearing input. If it is half that, halve everything above.

That intention converts at all. Nobody has demonstrated a 3% activation rate on this population, because nobody has built the product that would test it. The number is an assumption.

That the price holds. Stated willingness to pay for meditation content is $4.90 to $5.90 a month. The $199 case rests on the claim that personalized guidance is a different good from a content library, which the TM and coaching figures support and no software product has yet proven.

That the created demand is additive. Some fraction of these people would have eventually paid Calm. We assume most would not, on the evidence that they have already had a decade of the opportunity and have not.

The opportunity is the fifty-one million

The market is small because the product is thin.

The case, in six lines

Interest is high. 92% of Americans know what meditation is, half have tried it, and 51 million intend to practice right now without doing it.

Guidance is the constraint. Two in five people who learn a practice stop, nineteen in twenty app installs are gone within a month, and the moment of failure is right after arrival, where a teacher would be.

Teachers work, and there are almost none. Serving the US intent pool one-to-one would take about two million teachers. The largest certification program has produced eight thousand.

The category's low price is a fact about its customer. The best-measured meditation buyer is a 48-year-old woman treating insomnia, and a remedy is a commodity, priced against melatonin. The same market pays $420 to $980 for four TM sessions and $234 an hour for a coach.

Removing the expert is the mechanism that has closed a gap of this shape twice: four in five of Duolingo's US users were not learning a language before it existed, and four in five of the people now building software with AI have no technical background.

Doing it here creates, on conservative assumptions, $100M to $500M of annual consumer spending in the United States alone, against a global paid-app category of $2 billion.

Competing for the 11.3% is the wrong game. That population is well served, price-anchored at six dollars a month, and defended by two companies with a hundred million downloads each. The 19.1% is unserved, unpriced, and larger. A product that takes a beginner from intention to a practice builds the market Calm has spent a decade failing to activate.

The risk, restated: if a personalized product is received as another meditation library, it will be priced like one, and the model above collapses to its low column. The wedge has to be a depth of personalization a practitioner can feel in the first session.

Claims we do not make

Seven figures that would have flattered this report, and why they are not in it.

Any single meditation-app CAGR, stated as fact. The published estimates run from 9.9% to 55%. Quoting one of them is quoting a vendor's model.

"200 to 500 million people meditate every day." It appears in vendor marketing and dozens of secondary articles and traces back to no primary source.

60.53 million American adults meditate. The figure is real and the base is wrong: it applies an adult prevalence to the whole US population. Roughly 49 million adults is the defensible reading.

That the market-creation model is a forecast. It is a model. Its inputs are named above and its largest assumption has never been tested by anyone.

That an analogy is evidence. Language learning and software were remade by removing a scarce expert, and meditation is built the same way. Analogy is a reason to look. The disanalogy is named in section eight.

Market-mill TAM figures presented as authoritative. Grand View, Technavio, Persistence and HTF are useful for showing that the category is poorly measured, and for nothing else.

"Millions of paying subscribers" for any app other than Calm and Headspace. Those two are the only ones clearly at that scale.

Method

Desk research against sixteen questions, with every figure graded before it was used.

The questions covered participation, digital usage, paid conversion, in-person instruction, growth rates, daily session volume, personalization as a value driver, the personalization depth of competing apps, the coach and retreat economy, per-person annual spend, luxury wellness, and the size of the spiritually unaffiliated population. The habit-formation, barrier, retention and buyer-profile evidence in sections two, four and five was added in a second pass, once the question turned to the population outside the market. Where two credible sources disagreed we report both and say why they differ. Where only vendor estimates existed we say so.

Absolute population figures use a US adult population of 267 million (Census Bureau, 2024 vintage). Where a cited paper used a different base, we say so.

Three things would change our reading. A second representative survey replicating the 19.1% Preparation figure would move the central input of this report from one study to two. A published free-to-paid conversion rate from any meditation app other than the two leaders would tell us whether four to six percent is the category's ceiling or its average. And a measured activation rate — anyone converting intenders into sustained practitioners at any rate at all — would replace the assumption the whole model rests on with a result.

Sources, and how to read them · four tiers, one figure we refuse to repeat

How we grade a source

Every number in this report sits in one of four tiers, and the tier travels with the number wherever we use it.

Cite freely.

Government surveys, peer-reviewed journals, Pew, the Global Wellness Institute, the International Coaching Federation, McKinsey, and direct company disclosures.

Cite with attribution and a date.

Sensor Tower, SBI Growth, Kajabi, platform press releases. Real, and self-interested.

Do not cite as fact.

Market-report vendors and content-marketing aggregators. Used here only to show that the category is badly measured, which is itself a finding.

Never cite.

Figures that circulate widely and trace back to no primary source. One of those is named at the bottom of this list.

Linked

Captured at retrieval time. These resolve.

• Bowen & Beam — Mindfulness (2025), the Sussex Mindfulness Meditation model

  The load-bearing figure of this report: 69.7% of US adults in Pre-Intention, 19.1% in Preparation, 2.3% in Action and 9.0% in Maintenance, from a representative Ipsos KnowledgePanel survey of 2,000 adults.

• Bowen, Beam, Semb & Yoo — JMIR Mental Health (2025), theory-driven messaging field experiment

  92% awareness, 49.8% ever tried, 15.1% in the past seven days; the four-barrier taxonomy; and the randomized result across 1.33 million Spotify Premium users that practical-friction messaging lifted activation while knowledge reassurance underperformed the control.

• Baumel, Muench, Edan & Kane — Journal of Medical Internet Research (2019)

  The retention cliff, measured from an independent usage panel across 93 mental health apps: 69.4% open at day 1, 3.9% at day 15, 3.3% at day 30, and 4.7% at day 30 for mindfulness and meditation apps specifically.

• Huberty, Vranceanu, Carney, Breus, Gordon & Puzia — JMIR mHealth and uHealth (2019)

  The only good portrait of who actually pays for meditation software: 12,151 paying Calm subscribers. Mean age 48.6, 80% female, 39% earning over $100,000; sleep is the top reason at 63%, 77% report sleep difficulty and 41% carry a mental health diagnosis.

• Global Wellness Institute — Defining the Mental Wellness Economy

  The four mental wellness sub-sectors and what the meditation and mindfulness segment counts inside it — classes, teachers, retreats, books, cushions, beads, chimes, journals, coloring books and biofeedback gadgets — plus GWI's own line that only a small fraction of the people who practice spend anything on it.

• Waking Up — subscription pricing

  $129.99 a year or $19.99 a month: the category's depth-positioned product, at close to double Calm's annual rate.

• Lam, Xie & Goldberg — JMIR Mental Health (2023), situating meditation apps

  Barriers among 434 meditators screened from 953 adults: cost 42.4%, doubt that it works 35.5%, time 21.2%, insufficient interest 20.5%. Also the finding that only 21.7% of meditators use an app weekly or daily.

• Davies, Faschinger, Galante & Van Dam — Scientific Reports (2024)

  18.3% of US adults used meditation in 2022, up from 7.5% in 2002, weighted across five NHIS cycles and 134,959 respondents. Also the source of the 60.53 million count this report corrects — the paper projects an adult prevalence onto the whole-population census estimate.

• US Census Bureau — 2024 vintage population estimates by age

  The 267 million US adult population base used for every absolute figure in this report: 205.7 million aged 18 to 64 plus 61.2 million aged 65 and over.

• Apple — first quarter 2026 results and WWDC 2026

  More than 2.5 billion active devices as disclosed in January 2026, and the rebuilt assistant layer announced in June — the distribution premise in the opening section.

• GitHub — Octoverse 2025

  36 million new developers joined GitHub in a year, up 23% year over year, to a total above 180 million — more than one a second — and 80% of new developers use Copilot in their first week.

• TechCrunch — Cursor's Anysphere nabs $9.9B valuation, soars past $500M ARR (June 2025)

  Cursor's revenue and valuation, as the scale marker for AI coding tools.

• Bloomberg — AI coding startup Replit valued at $3 billion with new funding (September 2025)

  Replit's valuation after its pivot away from professional developers.

• Semafor — Replit's CEO on companies and software developers (May 2025)

  Replit's stated position that its audience is now people who do not write code.

• Duolingo — third quarter 2025 results

  50.5 million daily active users, 11.5 million paying subscribers, and bookings past a billion dollars — the scale of a market built rather than captured.

• Global Wellness Institute — Global Wellness Economy Monitor

  The $6.8T wellness economy, the $181B mental wellness market, meditation and mindfulness spending at $7.1B growing 18.9% a year, wellness tourism, per-capita spend, and the global spa count.

• International Coaching Federation — Global Coaching Study

  $5.34B global coaching revenue across 122,974 practitioners, average fees, and the finding that 53% of coaches earn under $30,000 a year from coaching.

• International Coaching Federation — Professional Coaching Continues Global Expansion

  Practitioner counts and the growth of the coaching population since 2019.

• US Bureau of Labor Statistics — Fitness Trainers and Instructors

  The closest official occupational proxy for a meditation teacher: 370,100 jobs in 2024, median pay $46,180. There is no BLS code for the job itself.

• Sounds True — Mindfulness Meditation Teacher Certification Program

  Over 8,000 certified teachers across more than 70 countries.

• Insight Timer — Google Cloud customer case study

  User base and infrastructure scale.

• Insight Timer — App Store listing

  Library size and the claims the product makes about itself today.

• Pew Research Center — on religious nones, and on leaving and joining

  The unaffiliated share of US adults and its age distribution.

• Simply.Coach — ICF coaching statistics

  Secondary summary of ICF data, including the income distribution. Used to cross-check, not to originate.

• Mighty Networks

  Host earnings and community-platform economics, as the platform reports them.

• StriveCloud — Headspace gamification features

  Headspace download and retention figures. Secondary.

• BusinessDojo — retreat pricing

  Retreat pricing benchmarks. Secondary; used directionally.

• StillMind — meditation statistics

  Aggregated participation and frequency data. Orientation only; nothing in this report rests on it.

• Amra & Elma — meditation platform marketing statistics

  App download and engagement aggregation. Marketing-agency aggregator, low confidence, listed for completeness.

Named, not linked

These carry several of the report's strongest numbers. The publications and studies are real and the figures came from them, but the exact URL was not captured when the figure was taken — so they are named without links. A plausible-looking URL invented after the fact turns a real study into an unverifiable claim. Verify each before quoting it anywhere it matters.

• Nahin, Rhee & Stussman — JAMA, 2024

  The alternate 17.3% reading for 2022, adopted by NCCIH. The difference from the figure above is population weighting.

• CDC National Health Interview Survey — 2002, 2007, 2012, 2017, 2022 cycles

  The underlying dataset for both papers above, and the 7.5% and 14.2% readings for 2002 and 2017.

• Simonsson et al. — PLOS ONE, May 2024

  UK mindfulness: 16% of adults had learned a practice by 2021, of whom 25% still practice regularly, 32% occasionally, and 43% have stopped.

• Pew Research Center — Spirituality Among Americans, December 2023, n=11,201

  70% of US adults describe themselves as spiritual in some way; 22% are spiritual but not religious; 83% believe people have a soul or spirit.

• Pew Research Center — Religious Landscape Study, 2023–24

  28–29% of US adults religiously unaffiliated, and the finding that 69% of them are under 50.

• McKinsey — The Next in Personalization (2021) and What is Personalization (2023)

  A 10–15% revenue lift from personalization, and the finding that faster-growing companies take 40% more of their revenue from it.

• SBI Growth — meditation app willingness-to-pay survey, n=5,391

  Stated willingness to pay of $4.90 a month for weekly meditators and $5.90 for near-daily ones.

• Sensor Tower

  App revenue and download figures for the smaller paid apps, and the retention comparison across the category.

• tm.org — published course fee tiers

  The $420 / $540 / $780 / $880 / $980 income-based scale for four sessions of Transcendental Meditation instruction.

• Calm, Headspace, Insight Timer, Duolingo — company disclosures and investor materials

  Subscriber counts, list pricing, session and minute totals, and Duolingo's paid-subscriber and DAU figures.

• Kajabi press release, October 2023; Teachable creator earnings

  Creator-economy proxies for what an independent teacher earns from a cohort or a course.

• Duolingo — 2021 S-1 survey, roughly four in five US users not learning a language before Duolingo

  Quoted consistently across secondary coverage of the IPO filing and attributed to Duolingo's own survey, but the filing would not serve to us and we have not read the sentence in the original. Strong and unverified; the report says so where it uses it.

• Lovable — Build Economy report

  Roughly four in five people building with AI tools have no technical background; about 6% of Lovable's users are engineers and 45.7% are founders. Company-published; we have not seen the underlying method.

• Vercel — v0 usage report, 2025

  63% of v0's stated 4M+ users are not developers. Company-published, same caveat.

• Replit — revenue growth from $10M to $150M ARR

  Widely reported across trade press and consistent between outlets, but we did not find a company filing or first-party statement to anchor it.

• Rosetta Stone — peak annual revenue

  About $250M to $273M at its height, depending on which financial aggregator you read. Used only to establish an order of magnitude against Duolingo's billion in bookings.

Cited only for their disagreement

Four vendors sizing the same market, quoted here to show the spread and for nothing else. None of these is treated as authority anywhere in this report.

• Grand View Research

  Meditation apps $2.2B in 2025 to $6.99B by 2033, a 14.67% CAGR.

• Persistence Market Research

  9.9% CAGR — the low end of the spread.

• Research and Markets

  29.6% CAGR.

• Technavio

  55% CAGR — the high end, and 5.6 times the low one.

• HTF Market Intelligence

  Healthy living retreats at roughly $46B in 2025 rising to $73B by 2033. Same caution.

Never cited

“200 to 500 million people meditate every day.” It appears in vendor marketing, it is repeated in dozens of secondary articles, and it traces back to no primary source. It is not in this report.