The 2026 GLP-1 Business Blueprint

The 2026 GLP-1 Business Blueprint: The 90-Day Critical Path

What this document is

This is an operating blueprint, not a pitch deck. It describes how a cash-pay metabolic health business is assembled in 2026: the legal structures available, the clinical staffing question, the technology and supply chains underneath it, how members are acquired and kept, and the specific ways the model can fail.

It is written to be useful if you never speak to us. If you read all fifteen pages and go build this yourself, or hire a lawyer and a fractional operator and assemble it piece by piece, the document has done its job. We would rather be judged on whether the analysis holds up than on whether the brochure was persuasive.

Atlas Metabolic licenses a complete white-label metabolic health build to a partner. That is what we sell, and page fifteen tells you how to start that conversation. Pages two through fourteen are the category, not the company.

• Fifteen discrete pages, each self-contained
• Every statistic carries a named public source
• Claims we reasoned to rather than sourced are labelled INFERENCE
• Nothing here is legal, medical, tax or financial advice

What this document is not

There are no income numbers in this blueprint. Not a revenue projection, not a margin table, not a payback period, not an illustrative example with figures attached. That is deliberate and it is not modesty.

A company selling a business build that shows you earnings figures without a Franchise Disclosure Document is showing you a number nobody audited, nobody substantiated, and nobody is accountable for. Atlas has no FDD. So Atlas shows no earnings figures. If you have been shown such numbers by anyone in this category, page fourteen tells you exactly what to ask for next.

There is also no price in this document. Where a fee would naturally appear you will see The license fee is one-time. It carries 0% of revenue and no ongoing partner fees. The figure is not published anywhere, by design — it is disclosed in full on the fit call, where it can be put next to what it covers instead of floating on its own.. The license fee is disclosed on a fit call, in context, against what is actually being delivered. We explain the reasoning on page fifteen rather than asking you to accept it here.

How the pages are organised

Pages two through four establish whether you are the right reader and whether the market is real. Pages five through nine are structure: franchise versus license, the entity question, clinical staffing, technology, and supply. Pages ten and eleven are the commercial engine, acquisition and retention. Page twelve is the sequence. Pages thirteen and fourteen are the ones that matter most.

If you only have ten minutes, read page thirteen and page fourteen. Page thirteen states the risks in this category without softening them. Page fourteen is a diligence instrument you can point at any company selling you a build, including this one.

On sources

Every figure in this blueprint traces to a named publisher: CDC and the National Center for Health Statistics for prevalence, KFF for adoption and coverage, FDA and the Federal Register for compounding policy, Morgan Stanley, Goldman Sachs and Grand View Research for market forecasts, the AMA for liability costs, the Milbank Memorial Fund for corporate practice of medicine.

Where the widely repeated number in this industry could not be traced to a real source, we say so and omit it. Several commonly cited startup-cost and prescription-volume figures fall into that category. A blueprint that quietly repeats unsourced numbers is worth less than one that admits the gap.

Where to start if you are new to the category

If the vocabulary is unfamiliar, three of our library pieces set the baseline before you continue: what turnkey actually means when a company uses the word, how to start a wellness business as a general matter, and a plain overview of the metabolic wellness market. None of them require you to have read this blueprint first.

Otherwise, turn the page. Page two is about disqualification, and it is honest about it.

Who This Is For, and Who It Is Not

The four profiles this fits

First, the operator with capital and no clinical background: someone who has run a business, understands acquisition cost and service delivery, and wants a cash-pay category with durable demand. This profile succeeds when the clinical side is held by a separately licensed medical entity and the operator stays firmly on the business side of that line.

Second, the healthcare professional with an existing member or patient base: a functional medicine practice, a chiropractic office, a medspa. The advantage here is distribution, not clinical skill. You already have people who trust you and who are in the demographic. The constraint is that adding a metabolic program to an existing practice raises entity and scope questions that page six covers.

Third, the investor-operator pairing: capital from one party, day-to-day operating from another. This works when the operating partner is genuinely accountable and genuinely present. It fails when the capital partner assumes a build runs itself.

Fourth, the multi-site operator adding a line. If you already run a service business with front-desk staff, scheduling and local marketing, most of the hard operating muscle exists.

The five profiles this does not fit

It does not fit someone looking for passive income. This is an operating business with staffing, compliance obligations and a marketing spend that requires management. Nobody hands you a machine that runs without you.

It does not fit someone whose capital is fully committed to the license fee with nothing behind it. Beyond any build fee, you carry working capital for acquisition, staffing and the clinical entity's own costs. Committing your last dollar to the front end is the most common way this fails.

It does not fit someone who wants earnings guarantees. There are none, from us or anyone honest in this category, and page thirteen explains why the absence is the credible signal.

It does not fit someone who wants to avoid the compliance layer. Corporate practice of medicine, state licensure, advertising rules and supply provenance are permanent operating features here, not launch-week paperwork.

It does not fit someone who wants to sell a product they will not stand behind. Retention decides this business, and retention is downstream of whether the program is genuinely good.

The honest capital question

The right question is not whether you can afford the entry. It is whether you can fund the first several months of operating and marketing without the business needing to carry itself immediately. Commonly cited first-year startup and operating ranges for a medical practice run roughly $70,000 to $500,000, with build-out at $20,000 to $60,000, clinical equipment at $15,000 to $75,000, and IT and EHR infrastructure at $10,000 to $25,000. Those are cost ranges only and we label them INFERENCE: they recur across practice-management publishers, the widely repeated attribution to a specific industry survey could not be traced, and no revenue figure accompanies them or may be derived from them.

A telehealth-first or hybrid model changes the composition of that cost rather than reducing a stated total: it removes build-out and clinical equipment lines while adding multi-state licensure and technology lines. That is a structural statement, not a magnitude claim.

The temperament question

The people who do well in service categories with a clinical adjacency tend to share three traits. They are comfortable saying no to a prospective member who is not a fit. They treat compliance as an operating discipline rather than an obstacle. And they are willing to be the accountable adult in a business where a licensed clinician, a marketing channel and a supply chain all have to hold at once.

If reading page thirteen makes you want to close the document, that is useful information and it is not a failure. This category rewards people who read the risk page twice.

If you are already in practice

Existing practices carry a different set of questions than greenfield operators. A functional medicine practice already knows how to enrol members into a high-ticket program and is mostly solving supply, technology and program structure. A chiropractic office expanding into metabolic wellness is usually solving scope, entity structure and clinical staffing. A medspa integrating a weight-loss program is usually solving clinical oversight and member experience continuity.

In all three cases the practice's own clinical patients remain patients of that practice. Members of a metabolic program are members. Keeping those two populations, and their records, cleanly separated is an entity question, and it is the subject of page six.

The Market, With Sources

Prevalence: the addressable population

US adult obesity prevalence was 40.3% during August 2021 through August 2023, with severe obesity at 9.4%, according to CDC and the National Center for Health Statistics. Prevalence was 39.2% in men and 41.3% in women, and highest among adults aged 40 to 59 at 46.4%. Including overweight, 72.4% of US adults aged 20 and over were affected over the same period.

The trend line matters as much as the level. Age-adjusted adult obesity prevalence did not change significantly from 2013-2014 through August 2021-August 2023, while severe obesity rose. This is a large, stable, slowly worsening population, not a spike.

The 40-to-59 concentration is the single most operationally useful number on this page, because it tells you who your marketing is actually speaking to.

Adoption: who is actually using these drugs

KFF's health tracking poll, fielded October 27 to November 2, 2025, found 18% of US adults say they have ever used a GLP-1 drug and 12% are currently using one, up six percentage points in current use since May 2024.

Current use concentrates where you would expect: 45% among adults with diabetes, 29% among those with heart disease, 23% among adults diagnosed as overweight or obese in the past five years, 22% among ages 50 to 64, and 9% among those 65 and over.

The affordability picture inside that user base is the part most operators underweight. Among users, 56% reported difficulty affording the drug and 25% said it was very difficult. Twenty-seven percent had insurance but paid the entire cost themselves. Fourteen percent stopped because of cost. That last figure is a retention problem before it is anything else, and page eleven treats it as one.

Forecasts: read the spread, not the headline

Morgan Stanley Research projects global GLP-1 sales across diabetes and obesity at approximately $190 billion by 2035, more than double 2025 levels. Goldman Sachs Research forecasts the global anti-obesity drug market at approximately $95 billion by 2030, revised down from an earlier estimate of approximately $130 billion. Grand View Research projects the narrower GLP-1 weight-loss segment at approximately $48.8 billion by 2030, on an 18.5% compound annual growth rate from 2025.

These are not contradictions. They are different definitions across different horizons: class-wide including diabetes, anti-obesity drugs generally, and the GLP-1 weight-loss segment specifically. Anyone quoting only the largest of the three at you is selling, not analysing.

The Goldman downward revision is the most instructive item here. A serious forecaster cutting a category estimate by roughly a quarter is a signal about pricing pressure and adherence assumptions, and it belongs in your planning more than the maximum headline does.

What this data does not tell you

It does not tell you what a business in this category will produce. Market size is the size of a category's drug sales, not of your service revenue, and there is no defensible arithmetic that bridges from one to the other for a single operator.

We will not build that bridge, and you should treat anyone who does as having told you something important about themselves. Atlas has no Franchise Disclosure Document, and without one any figure describing what a partner might earn is an unsubstantiated earnings claim regardless of how it is labelled.

What the data does tell you is that the population is large and stable, adoption is rising, use is concentrated in identifiable segments, and cost is the dominant friction inside the existing user base. Those four facts shape a service model. They do not forecast one.

Numbers we deliberately omit

Three figures circulate widely in this category that we will not repeat. A frequently quoted 2025 category sales total with per-product breakdowns could not be confirmed against IQVIA or company filings. No reliable public national count of semaglutide or tirzepatide prescriptions dispensed exists, and retail-discount fill-growth percentages describe one publisher's own fill base rather than national volume. And the commonly stated count of states with corporate practice of medicine laws rests on inconsistent methodology; the Milbank brief itself declines to state a count.

Omitting a number you cannot source is not a gap in a market analysis. It is the analysis.

Why Cash-Pay

Mechanism one: the Medicare exclusion is statutory

Medicare Part D has been prohibited by statute from covering drugs used for weight loss since the program was created under the 2003 Medicare Modernization Act. Access under an obesity indication exists only where another medically accepted indication applies. This is not a coverage policy that a plan can revise; it is legislation.

Two temporary programs sit on top of that exclusion. The Medicare GLP-1 Bridge runs July 1, 2026 through December 31, 2027 at a $50 monthly copay, and those copays do not count toward Part D deductibles or out-of-pocket maximums. The BALANCE model began for Medicaid in May 2026 and was delayed indefinitely for Medicare Part D after plan participation fell short of an 80% enrollment threshold. State and plan participation in both is voluntary.

Read those two paragraphs together. The permanent structure is exclusion. The relief is temporary, partial, voluntary, and in one case already stalled.

Mechanism two: employer coverage is a minority position below the largest firms

KFF's 2025 Employer Health Benefits Survey, fielded January through July 2025 across 1,862 firms, found that among firms offering health benefits, 16% of firms with 200 to 999 workers cover GLP-1 agonists used primarily for weight loss, rising to 30% at 1,000 to 4,999 workers and 43% at 5,000 or more.

Coverage at the largest firms is growing fast, from 28% in 2024 to 43% in 2025. But the same survey shows why that growth is not safely extrapolated: 59% of the largest covering firms reported utilisation above projections and 66% reported significant impact on prescription drug spending. Thirty-four percent of covering firms required a provider visit or lifestyle program participation as a condition.

That last requirement is worth pausing on. Employers who do cover are increasingly attaching a clinical-touch or program condition, which is a demand signal for structured programs rather than for drugs alone.

What the two mechanisms produce

Cash-pay demand in this category is produced by a statutory Medicare exclusion addressed only through temporary voluntary programs, plus majority non-coverage of the weight-loss indication among mid-sized employers. We label this INFERENCE: no single publisher states it as a conclusion, and it is a characterisation of the coverage landscape, not a demand forecast and not a statement about any operator's results.

The distinction matters because a demand argument built on a temporary supply shortage expires when the shortage does. That is precisely what happened to a large part of this category in 2025, and page nine covers it. A demand argument built on statute and benefit design has a different half-life.

The counterweight: direct-to-consumer pricing is falling

Manufacturers now sell directly. Wegovy self-pay pricing through NovoCare Pharmacy starts at $149 per month for the 1.5 mg or 4 mg oral doses for new patients, with the standard pen starting at $199 per month for the first two months as a limited-time offer and the HD 7.2 mg pen from $399 per month after the introductory period, all stated as subject to change and checked August 2026.

On November 6, 2025 the White House announced agreements with Eli Lilly and Novo Nordisk placing starting doses at $350 per month via the TrumpRx direct-to-consumer site that launched in January 2026, trending toward $245 over two years, with Medicare prices set at $245 and a $50 beneficiary copay, and approximately $150 per month for oral GLP-1s if approved.

If your model's value proposition is access to a molecule at a price, that proposition is being competed away by the manufacturers themselves. This is the single most important strategic fact on this page and it argues for building a program, not a pharmacy counter.

What survives the pricing pressure

What manufacturers are not selling direct is the wrapper: clinical oversight through a separately licensed medical entity, structured nutrition and behaviour support, lab work and monitoring, adherence coaching, and a member experience someone stays inside for a year rather than a quarter. The affordability data on page three is a strong hint that people leave for cost reasons long before they leave for clinical ones.

A cash-pay metabolic business in 2026 is a program business with a medication component, not a medication business with a program veneer. Page eleven argues that retention is where that distinction gets paid for.

Franchise or License

What a franchise actually is

Franchising is a regulated way of selling a business system. In broad terms, regulators look at whether the buyer operates under the seller's trademark, whether the seller exerts significant control over or provides significant assistance to the buyer's method of operation, and whether the buyer pays a required fee. Whether a specific arrangement meets the definition in your state is a legal question, and it is one to put to your own counsel rather than to the company selling you the arrangement.

Where the definition is met, the seller must deliver a Franchise Disclosure Document before you pay. The FDD is the single most valuable artifact in this entire comparison, and most buyers never read it properly.

What the FDD gives you, and what it costs

An FDD forces disclosure across twenty-three items: the seller's litigation and bankruptcy history, the identity of its principals, initial and ongoing fees, your obligations, restrictions on what you may sell, termination and renewal terms, a list of current and former franchise buyers you may contact, audited financial statements, and, if the seller chooses to make one at all, a financial performance representation in Item 19 that must be substantiated.

Item 6 is the one that decides the arithmetic. It lists every recurring fee: royalty, brand fund, technology fee, required purchases, transfer fees. A royalty stated as a percentage of gross revenue is charged on money that arrives before your costs are paid, which means it is charged regardless of whether the underlying activity was profitable at all.

The price of that disclosure regime is what it discloses: an ongoing claim on your business, a set of operating restrictions, and typically a granted territory with terms attached to it.

What a license is

A license is a commercial agreement, not a regulated disclosure regime. You get whatever the contract says you get, and you get no FDD, no mandated litigation history, no audited financials and no mandated list of prior buyers to call.

That cuts both ways with unusual force. The absence of ongoing fees can be genuinely better arithmetic. The absence of mandated disclosure is genuinely worse protection. Anyone who tells you a license is simply safer than a franchise is describing one half of the trade.

The correct response to a license structure is not to relax your diligence but to run manually the diligence an FDD would have run for you. Page fourteen is that checklist.

Where Atlas sits

Atlas Metabolic licenses a build. The license fee is one time. There is no revenue share, no royalty, and no ongoing partner fee. Atlas does not sell geographic exclusivity, and no clause purports to protect a partner's market from other partners. Whether that structure meets the franchise definition in your state is a question for your counsel, and we would rather you ask it than take our characterisation.

The license fee is one-time. It carries 0% of revenue and no ongoing partner fees. The figure is not published anywhere, by design — it is disclosed in full on the fit call, where it can be put next to what it covers instead of floating on its own.

Ask Atlas to put this in writing: the exact market-selection language in the current Atlas license agreement, so the blueprint can quote the clause rather than paraphrase it

Because there is no FDD, there is no Item 19, and therefore there are no earnings figures anywhere in Atlas materials. That is not a marketing choice we made reluctantly. Without substantiation, an earnings figure is the red flag, not the proof.

How to compare the two honestly

Do not compare a one-time fee to a one-time fee. Compare the total claim each structure has on your business over the period you intend to operate it, including the restrictions each places on what you may sell, who you may buy from, whether you may sell the business, and what happens at renewal or termination.

Then compare the protection. A franchise gives you a disclosure document backed by a regulatory regime. A license gives you a contract. If you take a license, you have to manufacture the protection yourself, in the contract and in your diligence, and you should assume nobody else will do it for you.

Both structures have honest sellers and dishonest ones. The structure tells you what questions to ask. It does not tell you the answer.

The Entity and MSO Question

The doctrine, stated plainly

The corporate practice of medicine doctrine is a body of state-level rules barring unlicensed persons and corporations from owning or controlling a medical practice, or from employing physicians to deliver clinical care. Its stated purpose is to preserve independent medical judgment from commercial interference.

It varies by state, materially. We deliberately give no count of how many states have it, because the counts circulating in this industry rest on inconsistent methodology and the leading published brief declines to state one. Anyone quoting you a confident number of CPOM states is repeating something they did not check.

Nothing on this page is legal advice. Treat it as a list of things to verify with your own counsel in the states where you intend to operate.

The standard structure

The conventional response is separation. A professional corporation or professional entity, owned by a licensed clinician, holds the clinical practice: it employs or contracts the clinicians, owns the clinical relationship with patients, and holds clinical records. A management services organisation, which may be owned by non-clinicians, provides non-clinical functions under a management services agreement: administration, scheduling infrastructure, marketing, technology, facilities, billing support, human resources.

This is well-trodden. It is also where enforcement attention lands, because the structure can be used to separate business from medicine or to disguise business control of medicine, and the paperwork looks similar in both cases.

The control test is the whole game

Enforcement risk turns on the degree of control the MSO exercises over clinical operations and professional judgment. That is the axis regulators examine, and it is examined on substance rather than on what the agreement is titled.

In practice the questions are concrete. Who decides clinical criteria for enrolling a member into a program? Who sets visit length and clinician workload? Who can hire and fire the clinicians? Who owns the clinical records? Can the MSO override a clinician's decision to decline a member? Is the management fee structured in a way that survives scrutiny, and is it defensible as fair market value for actual services?

If the answer to several of those is the non-clinical entity, the structure is a label rather than a separation, and no amount of contract drafting fixes it after the fact.

The direction of travel is tightening

State activity in 2025 moved toward tightening rather than loosening. Massachusetts enacted MSO ownership transparency requirements. Oregon, Washington and California saw CPOM-strengthening bills introduced.

Plan on the assumption that disclosure obligations increase and that structures which currently sit in grey areas get less grey. A structure built to survive scrutiny is cheaper than a structure retrofitted under it.

What this means for a licensed build

A company selling you a build cannot be your clinical entity. Atlas provides no medical services and employs no clinicians. Clinical care in an Atlas-built business sits with a separately licensed medical entity, and the relationship between your business entity and that clinical entity is one you form with your own counsel, in your state, on your terms.

That means part of what you are buying in any build is the non-clinical apparatus, and part of what you must still assemble is the clinical side. A seller who is vague about which side of that line each deliverable sits on is telling you something. Ask them to draw the line on paper, per deliverable.

Ask Atlas to put this in writing: the specific entity-formation support Atlas provides versus what the partner's own counsel must complete, itemised for the fit call

The Clinical Staffing Model

The rule that governs everything else

The practice of medicine is generally deemed to occur where the patient is located. A clinician therefore typically must be licensed in the patient's state, not merely in their own. For a telehealth-forward metabolic business this is the binding constraint on geographic reach, and it is a licensure fact rather than a technology fact.

The Interstate Medical Licensure Compact provides a streamlined multi-state pathway and covered up to 43 member states as of March 2026. Streamlined is not the same as instant or free, and the compact does not cover every state or every licence type.

Note the vocabulary carefully here. When a clinical entity delivers care, the people receiving it are that practice's patients. People enrolled in the metabolic program are members. Both words are correct in their own place and confusing them muddies both the compliance picture and the marketing.

Who does what

A typical staffing model separates four functions. Clinical decision-making sits with licensed clinicians in the professional entity. Clinical support, intake review, lab ordering workflows and follow-up sit with clinical staff under supervision consistent with state scope rules. Member coaching, which is education and behaviour support rather than clinical care, sits on the program side. Administration, scheduling and acquisition sit with the business entity.

The most common structural error is letting the coaching layer drift into clinical territory. A coach discussing dose, side-effect management or medication changes has crossed a line, and the fact that they were well-intentioned does not un-cross it. Scripts, escalation rules and a hard handoff to the clinical entity are not bureaucracy; they are the mechanism that keeps the model defensible.

What clinical staffing costs to carry

Two cost lines are worth stating with real sources because they surprise people. DEA practitioner registration is $888 per three-year registration period under 21 CFR 1301.13, per practitioner, and applies where controlled substances are involved.

Medical liability insurance is rising and has been for some time. The AMA, using Medical Liability Monitor data, reported in May 2026 that premiums rose for a seventh consecutive year and that the share of reported premiums increasing year over year climbed from 13.7% in 2018 to nearly 40% in 2025, the highest share since 2005.

Premiums are intensely market and specialty specific. The AMA reported a 2025 manual premium for an internal medicine physician in Miami-Dade County, Florida of $59,736, against $243,988 for ob-gyn and general surgery in the same market. Treat that internal medicine figure as a high-cost outlier illustrating spread, never as a typical number for your market. What is generalisable is the direction, not the level.

State licensure and registration costs, including state medical licences, professional entity registration and state controlled-substance registration where applicable, are jurisdiction-specific and cannot be generalised. We label that INFERENCE and decline to invent a range.

Employed, contracted, or partnered

There are three common arrangements and each trades off differently. An employed clinician inside the professional entity gives you the most continuity and the most fixed cost. A contracted clinician gives flexibility and lower fixed cost but weaker availability guarantees and more turnover risk in a member-facing relationship. A partnership with an existing clinical entity gives you speed, because their licences, credentialing and insurance already exist, at the cost of dependence on a party you do not control.

For a build targeting a ninety-day path, the third option is usually what makes the timeline real, because licensure and credentialing timelines are the slowest items on page twelve and they are almost entirely outside your control.

Ask Atlas to put this in writing: whether Atlas introduces partners to established clinical entities, or whether the partner sources the clinical relationship independently, and what the build includes either way

The question to settle before anything else

Decide your service footprint before you decide your staffing. A single-state, single-location model has a narrow licensure burden and a heavier facility burden. A multi-state telehealth model inverts both. A hybrid carries some of each.

Everything on pages eight, nine and twelve depends on which of those three you choose, and changing your mind at day sixty is expensive in a way that changing it at day five is not.

The Technology Stack

The seven systems

A working metabolic health business runs on roughly seven things. An EHR or clinical documentation system inside the professional entity. A telehealth or visit platform. A scheduling system. A CRM and marketing automation layer for acquisition and follow-up. A payment and subscription billing system, because cash-pay programs are usually recurring. A member experience layer, which is where coaching, education, tracking and support live. And a reporting layer that tells you what is actually happening.

Commonly cited IT and EHR infrastructure costs of $10,000 to $25,000 with EHR and technology subscriptions of $500 to $2,000 per month recur across practice-management publishers, and we label them INFERENCE because the widely repeated attribution could not be traced. Use them as an order-of-magnitude planning frame, not a quote.

The line that must not blur

Clinical records belong inside the professional entity's systems, under its control, governed by its obligations. The member experience layer on the business side is a wellness product: education, coaching, adherence support, tracking, community.

Atlas builds the member experience layer as a wellness experience that does not hold clinical records and is not a clinician write-path. That is a deliberate scope decision, not a feature gap. Pulling clinical data into a consumer-facing product owned by a non-clinical entity creates obligations and control questions that page six explains you do not want.

The practical test: if a regulator asked who controls the record of a clinical encounter, the answer should be the clinical entity, without qualification.

Where builds actually fail

Not on any single system. On the seams. The five failures that recur are: a lead captured in the CRM that never reaches scheduling; a member who pays but whose program status never updates; a clinical visit that happens but is never reflected in the member's journey; a cancelled subscription that leaves access on; and a reporting layer built on numbers that three systems each define differently.

Every one of those is an integration and definitions problem rather than a software-selection problem. Which is why the useful diligence question about any build is not which tools does it use, but what happens end to end when a member signs up, pays, is declined by the clinician, and asks for a refund. Ask a seller to walk that path with you, live, in the actual systems.

What to insist on owning

Your domain, your brand assets, your member list and contact data, your ad accounts, your payment processor relationship, and an export path for everything in the CRM. If any of those sit in the seller's name, you are renting a business rather than owning one.

This is the difference between a stack built for you and a stack you are a tenant inside. Ask, in writing, what happens to each of those assets if the relationship with the seller ends tomorrow. Page fourteen makes it a formal question.

Ask Atlas to put this in writing: itemised list of accounts and assets that are created in the partner's name at build, versus any held by Atlas, so the answer can be published rather than described

Build, buy, or license

Building this stack yourself is entirely possible and it is mostly an integration project rather than a software project. The cost is time and the risk is that the seams above are exactly the part that takes longest to get right, because you only discover them under live load with real members.

Buying assembled tools individually gets you further faster but leaves the integration and the definitions to you. Licensing a build transfers the assembly, which is the whole value proposition and also the whole thing to interrogate: ask precisely which of the seven systems are delivered configured, which are delivered as accounts you must configure, and which are not included at all. A build that includes six of seven and is silent about the seventh is a schedule risk you will meet at day seventy.

Supply and Fulfilment

What happened in 2025

FDA removed tirzepatide from the shortage list in December 2024, with compounders expected to cease by March 2025. FDA declared the injectable semaglutide shortage resolved on February 21, 2025; enforcement discretion for 503A compounding pharmacies ended April 22, 2025 and for 503B outsourcing facilities on May 22, 2025.

A large number of businesses in this category had built their entire supply position, and therefore their entire price position, on shortage-era compounding. When the shortage ended, so did the legal basis for the supply. That is not a market shock. It was a foreseeable end date on a temporary permission, and it was foreseeable from the day the permission was granted.

The lesson generalises past GLP-1s: a supply position that exists because of a temporary regulatory condition is a countdown, not a moat.

What FDA says about compounded product

FDA states that a compounded product may contain the same active ingredient but is not the drug FDA reviewed and approved, and that FDA does not evaluate specific compounded products for safety, effectiveness, manufacturing quality or consistency.

FDA has received hundreds of adverse event reports associated with compounded semaglutide and tirzepatide, including reports linked to dosing errors, and has warned about fraudulently labelled product naming pharmacies that do not exist. In September 2025 FDA issued more than 50 warning letters to companies compounding or manufacturing semaglutide and tirzepatide over statements it deemed false or misleading.

That last item is a marketing enforcement fact as much as a supply fact. Read page ten with it in mind.

The live risk, with its dates

On April 30, 2026 FDA proposed excluding semaglutide, tirzepatide and liraglutide from the 503B bulk drug substances list, finding no clinical need for outsourcing facilities to compound them from bulk substance and expressly rejecting affordability and insurance access as constituting clinical need. The comment period was extended, with comments due July 30, 2026.

Note the reasoning, not just the outcome. FDA specifically declined to treat cost and coverage as a clinical justification. Any business model whose supply argument is that compounded product is cheaper for people who cannot afford branded product has been addressed directly by the regulator and told that argument does not count.

This is the single most consequential open item in the category as of this writing, and any seller of a build who cannot discuss it by name and date is not current.

What the market did anyway

IQVIA reported in October 2025 that compounded GLP-1 prescribing did not abate after delisting. More than 80% of compounded prescriptions include supplemental ingredients such as B vitamins or levocarnitine. Anti-obesity patients were approximately 83% of the compounded GLP-1 market during the shortage period, and roughly 2% of compounded patients switched to branded products monthly.

The supplemental-ingredient figure is the tell. Adding an ingredient is a common route to arguing a compounded product is not essentially a copy of an approved drug. Whether any specific formulation is permissible is a legal and regulatory question for the compounding entity, the prescriber and their counsel, not a marketing question and not one a build seller should be answering for you.

The 2% monthly switching figure tells you something else: this is a sticky channel, which means a supply disruption inside it does not resolve itself gently.

What a durable supply position looks like

Three properties. First, it does not depend on a temporary regulatory condition. Second, it survives price compression, because manufacturers are now selling direct at falling prices as page four describes, so any margin that exists solely because of a price gap on a molecule is borrowed. Third, it is documented: you can name the entity, see its registrations, and produce provenance for what a member receives.

Supplements and non-prescription components have their own supply chain with its own failure modes, and our library covers how those chains actually work, including who manufactures, who labels and where quality claims come from.

We make no efficacy claims about any medication or supplement in this document, and neither should any material you publish. Ask Atlas to put this in writing: the specific fulfilment relationships included in an Atlas build, named entity by entity, so partners can verify registrations themselves

Member Acquisition

Match the asset to the temperature

The failure mode in this category is sending cold traffic to an application form. Someone who has never heard of you, has been marketed at by every telehealth brand on the internet, and is carrying a decade of failed attempts is not going to fill in a qualification form because your headline was bold.

Cold traffic needs a briefing: an editorial asset that teaches something true and earns the right to the next step. Warm traffic, someone who read the briefing, attended the session, or came from a referral, is the audience for an application. The document you are reading is itself the pattern applied to a different audience.

Between those sits the middle: retargeting, email sequences, a second asset that goes deeper. Most operators build the top and the bottom and skip the middle, then conclude the channel does not work.

Where members actually come from

Four sources, in roughly descending order of reliability for a new business. Existing relationships, if you already operate a practice, which is the fastest and cheapest and is the reason page two treats existing practices as an advantage. Local presence and referral, including other practitioners who see the population but do not offer a program. Paid acquisition, which is the most scalable and the most expensive to learn on. And organic content, which is the slowest to start and the cheapest to sustain.

A ninety-day path realistically leans on the first two, while the third is being instrumented and the fourth is being seeded. Any plan that has paid acquisition carrying the business in month one is a plan that has never run paid acquisition.

The enrollment conversation

Cash-pay programs are sold in a conversation, not on a page. That conversation has a shape: understand the person's history and what has already failed, be explicit about what the program is and is not, be explicit about cost, be explicit about what is required of them, and be willing to decline people who are not a fit.

Declining is not a sales technique. In a program with a clinical component the clinical entity decides eligibility on clinical grounds, full stop, and the business side must never treat a clinical decline as a sales objection to be handled. Building your acquisition process so that the clinical gate is real, and is respected, is both the compliance position and, as page eleven argues, the retention position.

Our library piece on high-ticket enrollment in functional medicine covers the mechanics of the conversation itself in more depth than fits here.

What you cannot say

This category has an enforcement record and it is recent. In September 2025 FDA issued more than 50 warning letters to companies compounding or manufacturing semaglutide and tirzepatide over statements it deemed false or misleading. Marketing claims are where regulators reach businesses that are otherwise structured correctly.

Practical constraints: make no efficacy claims about any medication or supplement, make no guarantees of results, do not present a compounded product as equivalent to an approved drug, do not imply clinical outcomes your program has not measured, and keep clinical claims inside the clinical entity's voice where they belong.

Social proof is legitimate and powerful, but only if it is real and consented. Where this document would carry a testimonial, it carries instead. To fill that slot you need a named, consenting person describing their genuine experience, with any required disclosure, and with documentation on file. A composite, an actor, or an invented name is a fabricated endorsement, and it is the cheapest possible way to convert a marketing problem into a legal one.

Instrumentation, not vibes

From day one you want four numbers defined once and reported consistently: cost per lead, lead to consultation, consultation to enrollment, and enrollment to activation. Every seam failure on page eight shows up as a broken number in that chain.

We deliberately publish no benchmark figures for those metrics. Any benchmark we quoted would function as an implied earnings claim, and we do not have substantiated data to offer. Measure your own, weekly, and compare them to your own prior weeks rather than to numbers someone put in a deck.

Ask Atlas to put this in writing: which acquisition assets, ad accounts and funnel components are delivered in an Atlas build versus configured by the partner, itemised

Retention Decides the Business

The one number to design around

KFF found that among GLP-1 users, 56% reported difficulty affording the drug, 25% found it very difficult, 27% had insurance but paid the entire cost themselves, and 14% stopped taking it because of cost.

Read that as an operating fact rather than a market fact. In a cash-pay program, cost pressure is the leading documented reason people leave, and it is not primarily a clinical failure. It is an affordability and perceived-value failure, which means it is addressable by design rather than only by discounting.

We are not going to convert that into a churn projection, a retention curve, or any figure attached to your business, because any such number would be an earnings representation dressed as an operating metric.

Why retention carries more weight here than in most service businesses

Three structural reasons. Acquisition in this category is expensive and getting more competitive as manufacturers market directly. The program's value compounds over time, so a member who leaves in month two never experiences the thing you actually sold. And referral, which is the cheapest acquisition source on page ten, is generated almost entirely by members who stayed long enough to have an experience worth describing.

That makes retention the input to acquisition rather than a separate department. A business with weak retention is a business that must buy every member it will ever have, forever, at a rising price.

The first thirty days decide most of it

The pattern across service programs with a clinical component is that early experience predicts duration disproportionately. Practically, the first thirty days should do four things: set an accurate expectation of what happens and when, deliver a visible early win that is not weight on a scale, make contact frequent enough that nobody feels abandoned between clinical visits, and give the member a named human who knows their situation.

The failure pattern is equally consistent: an intensive sales conversation, a clinical visit, a shipment, and then silence. The member's next meaningful interaction is a card charge, which is the worst possible next interaction.

Note that KFF found 34% of employers covering these drugs required a provider visit or lifestyle program participation. Sophisticated payers have concluded that a structured program alongside medication is worth mandating. That is a demand signal for exactly the wrapper described here.

Design decisions that affect duration

Program length and how it is framed. A defined program with phases and a completion point behaves differently from an indefinite subscription, both in how it is sold and in how people leave it.

What is included beyond the medication component: labs and monitoring cadence, nutrition support, movement, sleep, behaviour work, and the coaching layer described on page seven. The more of the member's actual problem you address, the less the program's perceived value is pinned to one line item whose price is falling.

Cost structure and transparency. Surprise charges are a retention event. So is a member discovering that the same molecule is available direct at a lower price, if you never explained what else they were paying for.

And a real offboard. People leave. A program that ends well produces referrals and returns; one that ends in a failed card and an ignored email produces neither, and occasionally produces a complaint.

What to measure

Duration by cohort, stated reason for departure captured at the moment of departure rather than reconstructed later, contact frequency in the first thirty days, and the proportion of departures citing cost specifically. That last one tells you whether your problem is price, perceived value, or program design, and those three have completely different fixes.

We publish no benchmarks for any of them. Measure your own, compare to your own prior cohorts, and be suspicious of anyone in this category who hands you a retention figure they cannot substantiate.

The 90-Day Critical Path

The three items that determine whether 90 days is real

Before any schedule, understand what actually gates it. Clinical licensure and credentialing timelines are outside your control and vary by state, by board and by the individual's history. Entity formation and the professional entity structure depend on your counsel's availability and your state's filing times. Supply and fulfilment relationships depend on the counterparty's onboarding of you, not on your enthusiasm.

Everything else, brand, technology, content, funnel, program design, operating procedures, is inside your control and can be compressed. If the three items above do not start on day one, ninety days becomes a hundred and fifty and the reason is always the same.

This page describes a sequence, not a promise. Where a specific Atlas delivery date belongs, we mark it Ask Atlas to put this in writing: build milestone dates by phase, per the current delivery schedule rather than invent one.

Phase one, days 1 to 30: structure and decisions

The work in this phase is mostly legal, structural and decisional, and it is unglamorous enough that people skip ahead. Do not.

Decide the service footprint: single-state in-person, multi-state telehealth, or hybrid. Page seven explains why this decision governs everything downstream. Engage your own counsel on entity structure and the professional entity relationship. Begin licensure and credentialing for the clinical entity, whatever form it takes. Open banking, payment processing and insurance conversations, all of which have their own underwriting timelines. Lock program design: what the member receives, over what period, at what cadence. Lock brand and secure domains and handles in your own name.

Gate to phase two: entity structure decided and filed, clinical relationship identified with licensure in progress, program scope frozen.

Phase two, days 31 to 60: build and instrument

Now the assembly. Configure the seven systems on page eight and, more importantly, connect the seams between them. Build the member journey end to end and walk it yourself as a member, including the paths where someone is declined or asks for a refund.

Write the operating procedures: intake, the clinical handoff, coaching scripts with escalation rules, the refund and cancellation policy, the complaint path. Produce the acquisition assets in the temperatures page ten describes, at minimum one cold briefing asset and one warm application path. Establish supply and fulfilment relationships and verify the counterparties' registrations yourself rather than accepting a summary.

Gate to phase three: a live end-to-end walkthrough completed with a real payment and a real refund, and the clinical entity operational.

Phase three, days 61 to 90: pilot, then open

Do not open to paid acquisition on day sixty-one. Run a small pilot cohort from existing relationships and referral, few enough that you can speak to every one of them personally, long enough to see the first thirty days of member experience described on page eleven.

Fix what the pilot breaks. It will break the seams, the timing of communications, and at least one assumption in your program design. Then instrument acquisition, start paid spend small and deliberately, and let the four numbers on page ten accumulate before you scale anything.

Ninety days gets you legitimately open with real members and a measured funnel. It does not get you to a mature business, and any seller implying otherwise is describing a launch as though it were an outcome.

How this slips, every time

Five recurring causes. Licensure started in week three instead of week one. Entity structure revisited in week seven because counsel raised something that should have been raised in week one. Program scope changed after the technology was configured to the old scope. Payment processing underwriting delayed because the category triggered additional review and nobody started it early. And the pilot skipped, which does not delay the launch but reliably damages the first cohort.

The fix for four of the five is the same: start the slow, externally controlled items immediately, even when the exciting items are tempting. The fix for the fifth is discipline.

The Risks, Stated Plainly

Regulatory risk is live and moving toward you

Corporate practice of medicine enforcement turns on how much control a non-clinical entity exercises over clinical judgment, and 2025 moved toward tightening: Massachusetts enacted MSO ownership transparency requirements and Oregon, Washington and California saw CPOM-strengthening bills introduced.

A structure that is defensible today may face additional disclosure obligations tomorrow. If your structure only works because nobody is looking closely, plan for someone looking closely.

Marketing enforcement is equally live. FDA issued more than 50 warning letters in September 2025 to companies compounding or manufacturing semaglutide and tirzepatide over statements it deemed false or misleading. Businesses in this category get reached through their advertising more often than through their operations.

Supply risk has a docket number and a date

On April 30, 2026 FDA proposed excluding semaglutide, tirzepatide and liraglutide from the 503B bulk drug substances list, finding no clinical need and expressly rejecting affordability and insurance access as constituting clinical need. Comments were due July 30, 2026.

This is not speculative risk. It is a proposed action with dates, and its reasoning has already dismissed the argument most commonly used to defend the compounded channel. Anyone building a model that depends on that channel should be reading the docket, not a summary of it.

The 2025 precedent is instructive: enforcement discretion ended April 22, 2025 for 503A and May 22, 2025 for 503B, and businesses whose supply and pricing depended on it had no fallback.

Pricing risk comes from the manufacturers

Wegovy self-pay through NovoCare starts at $149 per month for certain oral doses for new patients, with the standard pen from $199 per month for the first two months and the HD pen from $399 per month after introduction, all subject to change. Under the November 6, 2025 agreements, starting doses are offered at $350 per month via TrumpRx, trending toward $245 over two years, with roughly $150 per month for oral GLP-1s if approved.

If your program's value proposition rests on price on a molecule, the counterparty compressing that price is the company that makes it. This is a permanent competitive condition, not a phase.

Coverage risk cuts both ways

The Medicare GLP-1 Bridge runs July 1, 2026 through December 31, 2027 at a $50 monthly copay. Large-employer coverage rose from 28% in 2024 to 43% in 2025. Both trends reduce the cash-pay population at the margin.

At the same time, 59% of the largest covering firms reported utilisation above projections and 66% reported significant prescription drug spending impact, which are the classic preconditions for employers restricting coverage again. And the BALANCE model was already delayed indefinitely for Medicare Part D after plan participation fell short of an 80% threshold.

The honest statement is that coverage is volatile in both directions and neither direction is safe to plan on.

The risks that attach to buying a build, including from us

There is no FDD. A license structure carries no mandated disclosure of litigation history, no audited financial statements, no mandated list of prior buyers to contact, and no substantiated performance representation. You must generate that protection yourself through the contract and the diligence on page fourteen. We would rather say that plainly than let the absence pass unmentioned.

Counterparty risk: if the seller ceases operating, what continues? Your entity, brand, member list, clinical relationship and supply relationships should all survive independently of the seller. If any of them do not, you have concentration risk you may not have priced.

Execution risk: a build removes assembly work. It does not remove operating work. Acquisition, staffing, compliance and retention remain yours, and no build quality compensates for an operator who is not present.

Category risk: this is a competitive, well-capitalised space with direct-to-consumer entrants who have advertising budgets you will not match. Differentiation has to be real, and local presence, clinical relationship and program depth are where it is available.

Concentration risk: a business built around one molecule class is exposed to that class's regulatory, supply and pricing conditions simultaneously. Breadth of program is a risk control, not just a value proposition.

What we cannot tell you

We cannot tell you what you will make. Not a range, not a scenario, not a labelled hypothetical. Atlas has no FDD, so there is no substantiated basis for any earnings figure, and publishing one anyway would be exactly the practice that page fourteen teaches you to treat as disqualifying.

We cannot guarantee supply, because a channel's regulatory basis is under active proposal. We cannot promise results, clinical or commercial. And we cannot advise you on the law in your state, which is why every regulatory passage in this blueprint is framed as something to verify with your own counsel.

If those absences make Atlas less attractive than a company that promises all four, that is a decision you are entitled to make with full information. We think the promises are the risk.

Questions to Ask Any Seller

On money and structure

Ask: What is the total fee, and is any portion recurring, ever? Is there any revenue share, royalty, technology fee, brand fund, required purchase or transfer fee, now or under any future circumstance? What triggers any additional charge? What is the refund policy, in the contract, in writing? May I see the full agreement before I pay anything, including any deposit?

Ask: Is this a franchise under my state's law, and what is your basis for that position? A seller who answers with certainty about your state's law rather than referring you to your own counsel is overreaching. The correct answer sounds like an explanation of the structure plus an invitation to have your lawyer review it.

Atlas's answers: the license fee is one time, The license fee is one-time. It carries 0% of revenue and no ongoing partner fees. The figure is not published anywhere, by design — it is disclosed in full on the fit call, where it can be put next to what it covers instead of floating on its own., with no revenue share, no royalty and no ongoing partner fee. The agreement is available for review before payment. Ask Atlas to put this in writing: the refund and cancellation terms as written in the current license agreement, quoted verbatim We do not offer a view on whether the structure is a franchise in your state; ask your counsel.

On numbers and proof

Ask: Have you shown me any figure describing what I might earn? If yes, ask for the Franchise Disclosure Document and the Item 19 substantiation behind it. If they have no FDD and are still showing you numbers, you have learned everything you need to know, and the correct next step is to leave.

Ask: Are your testimonials real, named, consented people describing their own experience? May I contact them directly, chosen by me from a full list rather than selected by you? Are any of them compensated, and is that disclosed?

Ask: How many partners have you built, how many are still operating, and how many stopped? What happened to the ones who stopped? A seller who cannot or will not answer the third question has given you the answer.

Atlas's answers: we publish no earnings figures anywhere, by policy, because we have no FDD and therefore no substantiated basis. Where consented social proof belongs in our materials you will see , and it stays empty until a real named partner consents in writing. Ask Atlas to put this in writing: current count of builds delivered, count still operating, and the process by which a prospective partner may contact them

On what is actually delivered

Ask for the deliverable list, itemised, with dates. For each item ask three follow-ups: is it delivered configured, delivered as an account I must configure, or not included? Who performs the work? What is the remedy if it is late?

Ask: Which parts are on the business side and which require the clinical entity? Who supplies the clinical entity, and is it separately licensed? Do you employ clinicians or provide any medical service? A seller who is vague here is either confused about corporate practice of medicine or hoping you are.

Ask: Walk me through, live in your real systems, a member who signs up, pays, is declined by the clinician, and requests a refund. Sellers with a real build can do this in ten minutes. Sellers with a slide deck cannot do it at all.

On ownership and exit

Ask: In whose legal name are the domain, brand assets, ad accounts, payment processor, CRM and member data? Can I export all member data at any time, in a usable format, without permission? Do you retain any right to use my member list or my brand?

Ask: What survives if you cease operating tomorrow? Which of my systems stop working? Which relationships are mine directly versus routed through you? Can I sell my business, and does that require your consent?

Ask: Do you restrict what else I may sell, who I may buy from, or where I may operate? Is any geographic exclusivity offered, how is it enforced, and what is the remedy if it is breached? Note that a seller offering protected geography is offering a promise that only means something if there is an enforcement mechanism attached to it.

On risk and the uncomfortable ones

Ask: What is your position on the FDA proposal of April 30, 2026 regarding the 503B bulks list, and what happens to my supply arrangement under each outcome? A seller who does not know what you are referring to is not current in this category.

Ask: What happened to your buyers when compounding enforcement discretion ended in 2025? Ask: What is the single most common reason your builds underperform? Ask: Who should not buy this? Ask: What do you get wrong most often?

Ask: What are the three biggest risks to this business, in your own words? Then compare their answer to page thirteen of this document. A seller whose risk list is shorter than the one their own competitor published is either less informed than you now are, or less candid.

Finally, ask everything above in writing and keep the replies. The pattern of what a company will put in writing, and how fast, is the most reliable signal you will get. If a seller answers freely on a call and goes quiet when asked to confirm by email, that is the finding.

The Next Step

What Atlas actually is

Atlas Metabolic licenses a complete white-label metabolic health build to a partner: the telehealth stack, the ordering system, the member AI coaching layer, the brand kit, and the operational playbooks. One-time license fee. Zero percent of your revenue. No ongoing partner fees.

Atlas provides no medical services and employs no clinicians. Clinical care in a business built with Atlas sits with a separately licensed medical entity. That is a hard boundary, not a technicality, and page six explains why it protects you as much as it constrains us.

What we sell is assembly and sequence. What remains yours is the operating business: acquisition, staffing, compliance, retention, and every decision that determines whether it works.

Why the fee is not on this page

The license fee is one-time. It carries 0% of revenue and no ongoing partner fees. The figure is not published anywhere, by design — it is disclosed in full on the fit call, where it can be put next to what it covers instead of floating on its own.

A number without a scope is noise. What is included varies with what you already have; an existing practice with a clinical entity, an EHR and a member base is buying a different build than a greenfield operator, and quoting one figure into both situations would misinform both.

There is also a positioning reason and we will state it rather than hide it: publishing a price invites comparison against packages that are not comparable, usually ones carrying an ongoing royalty that makes the headline look smaller. Page five explains why a one-time figure and a percentage figure cannot be compared on the front end alone.

On the call you get the actual number, in writing, against an itemised scope, before you decide anything.

What happens on the fit call

Roughly forty-five minutes, and the first half is us asking questions. What you already operate, your capital position, your intended service footprint, your timeline, and whether you have counsel engaged. If the answers say this is not a fit, we will say so on the call rather than after your deposit.

The second half is yours. Bring page fourteen and use it. The fee, the scope, the delivery schedule with dated milestones, the refund terms, and the ownership questions all get answered on that call, and everything material gets confirmed in writing afterwards.

We will not show you earnings figures, because we have none that are substantiated. We will not guarantee supply, because the regulatory basis of one channel is under active FDA proposal. We will not advise you on your state's law. If those are dealbreakers, this page has saved you a call.

Who we decline

People looking for passive income. People whose capital is fully committed to the front-end fee with no working capital behind it. People who want an earnings guarantee, because we cannot honestly give one and would not want a partner who bought on the strength of one. People who want us to hold the clinical side, which we cannot and will not do. And people who are not prepared to operate under the compliance layer that pages six, seven, nine and ten describe.

That list is not gatekeeping theatre. Each of those profiles produces a failed build, and a failed build costs the buyer far more than it costs us.

- this slot stays empty until a named Atlas partner provides written, consented permission to describe their own experience, with any required disclosure. We will not fill it with a composite, an actor, or an invented person, and if you see such proof anywhere in this category, page fourteen tells you what to do about it.

If you would rather build it yourself

That is a legitimate outcome of reading this blueprint, and it is not a failure of the document. Everything in these fifteen pages is either publicly sourced or structural. Take it to your own counsel, your own clinical partner and your own operator, and assemble it.

If you do, three of our library pieces are the right next reads: the full due diligence checklist for anyone selling into this category, why owning your own brand changes what the business is worth to you, and what sovereign ownership of an independent clinic actually means in practice.

And if you would rather not spend the next year on assembly, the application is the next step. It takes a few minutes, and it exists so that the call is a real conversation rather than a qualification exercise. Ask Atlas to put this in writing: application URL and the current response time Atlas commits to