RonanRx (YC S26): the software is in the supply chain
RonanRx's founder ran a mask factory that hit $50M and then evaporated. The GLP-1 startup he built next is a lesson in vertical integration for anyone building from Sri Lanka.
RonanRx (YC S26) launched this week selling compounded GLP-1s, and the interesting part has almost nothing to do with the drug. Per the company's Launch HN post and its site at ronanrx.com, one founder, Lloyd, previously built a mask factory that made a million masks a day and reached $50M in revenue, then watched demand disappear.
That arc contains two lessons worth more to a builder in Colombo than any funding headline.
🏭 The product is the supply chain, not the molecule
Semaglutide and tirzepatide are not secrets. Anyone can look up the compounds. So what exactly is RonanRx selling?
Read the seven steps the site describes and the answer becomes obvious:
| Step | What happens | Where software lives |
|---|---|---|
| 1. Intake | Patient history collected by text | Structured capture, triage |
| 2. Prescribe | Licensed physician approves or declines | Clinical decision workflow |
| 3. Verify | Pharmacist checks at the selected pharmacy | Routing, credential checks |
| 4. Compound | Made per patient, lot-traced ingredients | Traceability, batch records |
| 5. QA release | Pharmacist sign-off | Audit trail |
| 6. Ship | Cold chain with temperature tracking | Sensor telemetry |
| 7. Refill | Tied to continued physician review | Recurring compliance |
Every single step is a hand-off between a regulated party, and every hand-off is where a normal telehealth company loses money, time, or compliance. The site names its pharmacy partner explicitly (Elite Care Pharmacy LLC, TSBP #36002, in Hockley, Texas) rather than hiding behind a vague "partner network".
Key takeaway: the defensible thing is rarely the thing you sell. It is the number of expensive hand-offs you removed between raw input and delivered output.
The founder's own words about the mask factory make the pattern explicit: raw polypropylene pellets in, finished pallets out, with machine learning and computer vision used to optimise production. Same playbook, different regulated industry.
📉 The $50M lesson nobody puts on a pitch deck
Here is the part I keep thinking about. The mask business worked. It scaled. Then, in the founder's framing, people simply stopped caring about masks.
A business can be operationally excellent and still die because demand was a moment, not a market.
- Masks in 2020: demand created by an event, not a habit.
- Test kits, sanitiser, webcam mounts, ring lights: same shape.
- Every "AI wrapper" shipped in the last 18 months: possibly the same shape.
For a small Sri Lankan team, this matters more than it does for a funded US startup, because you do not have a runway to absorb a demand cliff. The practical defence is not prediction. It is asking two questions before you commit six months:
- If the triggering event vanished tomorrow, does anyone still need this?
- What of what I built survives the pivot? For Lloyd, it was the manufacturing-plus-software muscle, which transferred straight into pharmaceuticals.
If your answer to the second question is "nothing, I'd start from zero," you are not building a company. You are riding a wave and hoping it lands somewhere soft.
⚖️ The fine print is the business model
I want to be precise here, because health content deserves precision. This is not medical advice and I am not endorsing anything. But the disclosures on the page are genuinely instructive about how the company is built:
| What the site claims | What the site also states |
|---|---|
| Personalised, patient-specific formulations | "Compounded medications are not FDA-approved" |
| Can cost less than cash-pay brand list prices | Requires a valid patient-specific prescription |
| Physician reviews your history | The physician "either prescribes or declines" |
| Tirzepatide and semaglutide available | Compounded versions "should not be evaluated using branded-drug trial data" |
That last line is the honest one, and it is doing a lot of work. Compounded medicine sits in a legal lane that exists specifically because a patient sometimes needs something the mass-market product does not provide. The whole model depends on that lane staying open, which is a regulatory bet, not a technical one.
Sri Lanka's medicines regulation is a different regime entirely, and none of this maps across. Do not read a US compounding model as a template for anything here.
🇱🇰 Where this pattern actually transfers
Strip out the pharmacy and what remains is a repeatable structure: take a physical process that is already regulated and paper-based, and be the software layer that makes each hand-off traceable.
Sri Lanka is full of these. Apparel, tea, spices, cinnamon export, cold-chain fish, ayurveda manufacturing, freight clearance. All of them run on WhatsApp messages, Excel files and PDF certificates. All of them have buyers overseas who increasingly demand traceability they can audit.
You do not need a factory to start. You need:
- One process owner who will let you sit and watch the paperwork for a week.
- One hand-off that currently fails often enough to cost real money.
- A unit-economics model before you write code. Our break-even calculator is enough to sanity-check whether the fix pays for itself at the volume that business actually does.
A software layer over a regulated physical process is harder to copy than an app, because the moat is the relationships and the compliance record, not the code.
The unglamorous truth is that this work is slow and involves a lot of phone calls. That is exactly why it is defensible. Nobody spins one up over a weekend.
💡 What this means for you
If you are a student or a small-team builder in Sri Lanka reading yet another YC launch:
- Stop reading the product, read the hand-offs. The pitch is the drug. The company is the chain of custody.
- Assume your current market can vanish. Then ask what skill survives that. Build the skill deliberately.
- Go where the paperwork is. Regulated, physical, boring industries are underserved precisely because they are unpleasant to sell into.
- Do not confuse funding with validation. "Backed by Y Combinator" is on the RonanRx site. It tells you a partner believed the team. It tells you nothing about whether the market holds.
And if the health angle is what caught your eye rather than the engineering, the useful move costs nothing: understand your own numbers first with a TDEE calculator, and talk to an actual doctor before anything else.
The founder's story is not "I found a hot market." It is "I got very good at converting raw input into a finished product with software in the loop, and then I did it again somewhere the stakes were higher." That skill is portable. Almost nothing else in a startup is.