Which Drugs Open Up Next: A Practical Guide to Patent Expiries, Generic Launches and API Sourcing

About $236 billion in drug sales loses patent protection by 2030. Here is how I work out which molecules open up, in which countries, and who can supply the API — plus a note on the tool I'm building to do it at scale.

For the past few weeks I've been working on one question: which drugs come off patent next, where do they open up first, and who can supply the API? This post collects what I've found so far in one place. Part of it is a method you can follow yourself; part of it describes a tool I'm building to do the same work automatically.

A short note on where I'm coming from. I'm a computational biologist, with a PhD in computational genomics from NIT Warangal and an M.Tech from IIT Mandi. Most of my work involves taking messy public data and turning it into something people can act on. Patent and regulatory data is exactly that kind of problem, which is part of why it interests me.

This is written for people in generics, API manufacturing, portfolio planning, business development or regulatory affairs. It should also be readable if you simply want to understand how a drug moves from being patented to being available from several suppliers at a fraction of the price.

One caveat up front: this is a work in progress. I have the core data and a working first version. What I'm looking for now is feedback from people in the industry on what is useful and what is missing. More on that at the end.

The opportunity in numbers

Drugs with roughly $236 billion in annual sales are expected to lose exclusivity between 2026 and 2030. On the branded side this is called the patent cliff. For a generic or API company, the same event is an opening rather than a loss.

A few examples give a sense of the range:

Drug	Scale	Core protection ends	Note
Keytruda (pembrolizumab)	$31.7B sales (2025)	US — 2028	Biologic, so entrants make biosimilars: a slower, costlier route
Semaglutide (Ozempic)	—	India — expired Mar 2026; US — listed patent to 2031	40+ Indian launches, priced 50–60% below the innovator
My 25-drug watchlist	$78.9B US Medicare Part D spend	2026–2030	Small molecules; that figure is one payer, in one country

The point worth keeping in mind is that this value does not accrue to whoever is simply present in the market. It tends to go to companies that are early, prepared, and not entering an already crowded field. Timing is central.

Patents 101: what actually protects a drug

A drug is rarely protected by a single patent. It sits behind several layers, and the date that matters is when the last layer capable of blocking a generic expires.

A common mistake is to look up "the patent expiry," take the last date on the list, and stop there. That last date is usually a formulation or use patent, which generics can often design around. The more relevant date is the compound patent plus its extensions, or the regulatory exclusivity, whichever ends later. I refer to this as the core protection end, and it is the date my engine works from.

How to tell when a drug really opens up in the US

The US is a good place to start because most of the relevant information is public. One free FDA database, the Orange Book, lists every patent and exclusivity the innovator has claimed for each approved small-molecule drug. The steps I follow are roughly these:

Find the drug in the Orange Book and record every listed patent with its expiry date, noting which are flagged as drug substance (DS), drug product (DP) or method of use. The flag indicates how difficult each patent is to work around.
Separate the core patent from the rest. The earliest drug-substance patent, plus any pediatric extension, is usually the compound patent. Later patents are generally secondary and more easily avoided.
Check the exclusivities. A new chemical entity (NCE) receives 5 years of exclusivity; a generic may file at year 4 if it challenges a patent. Other exclusivities — orphan, new indication, new patient population — usually cover only a single use rather than the molecule itself.
Look for Paragraph IV challengers. When a generic company asserts that a patent is invalid or not infringed, that is a Paragraph IV filing. The FDA publishes the Paragraph IV list with the first filing date for each drug. A large number of challengers suggests the market will become crowded.
Understand the 180-day exclusivity. The first company to file a successful Paragraph IV challenge can obtain 180 days during which no other generic is approved. This window carries the best margins. Note that a lawsuit filed within 45 days triggers an automatic 30-month stay on approval, which can extend the timeline.
Consider carve-outs. If only one indication remains patented, a generic can launch for the other indications by omitting the patented use from its label. This is known as a "section viii" carve-out.
Watch for settlements. Many patent cases end in a settlement with an agreed entry date, which can be earlier or later than the patent dates. These agreed dates are often not public, which is why the date on paper and the actual launch date can differ. Apixaban (Eliquis) is a well-known example.

If you do only part of this, the first four steps give you most of the answer for most drugs.

Beyond the US: the same drug opens at different times

A patent is national. The same invention must be filed separately in each country, as a patent family, and each country sets its own rules for extensions and data protection. This is why semaglutide opened in India in 2026 while the US remains several years away.

Under TRIPS, most countries grant at least 20 years from filing, and through the PCT a single international application can lead to patents in many countries. Searching a family once shows both where it was filed and where it was not. A country with no filing is open from the start, and these gaps are easy to overlook when attention stays on the largest markets.

The API side: who can actually supply it

A launch date is of limited use without qualified API available in time. In the US this is relatively straightforward to check.

An API maker files a confidential Type II Drug Master File (DMF) with the FDA describing how it produces the drug substance. The DMF itself is not public, but the FDA publishes the list of DMFs each quarter, with the holder's name, the subject, and whether the file is active.

A generic company cannot read a DMF directly. It requests a Letter of Authorization (LoA) from the API maker, which allows the FDA to review the DMF as part of the generic's ANDA. In practice, "who can supply me API" means "who holds an active DMF and will provide an LoA."

The counts are informative:

Many active DMFs for a molecule indicate easy sourcing, but also a crowded market later.
Two or three, mostly from one country, indicate a supply risk — or an opportunity for a new API maker to enter.
Because of Bolar-type exemptions, API makers typically file years before expiry, so a rise in DMFs is an early sign that preparation has begun.

For India-focused companies this is the useful link: you can see which Indian API makers are already positioned for a given molecule, and which molecules have no Indian supplier yet.

How crowded will it get?

Prices fall quickly as more generics enter. The FDA's own analysis of US prices puts rough numbers on it:

Generic manufacturers	Approx. price drop vs brand
1	~39%
6 or more	>95% (prices fall toward ~20% of the brand)

So the relevant question is not only when a drug opens, but how many others will be entering at the same time. Being among the first two suppliers is a different proposition from being the eighth.

The likely level of competition can be estimated in advance from public signals: the number of Paragraph IV filers, the number of active API DMFs, the number of generics already approved, and the manufacturing complexity of the product. Injectables, inhalers and modified-release forms are harder to make, which tends to keep the field smaller.

Do it yourself: a checklist for one molecule

All of this can be done for free, for a single molecule, in an afternoon:

Market size — look up US spending in the CMS Medicare Part D dashboard, a reasonable proxy for the molecule's value.
Patents and exclusivities — use the Orange Book; mark the compound patent, the secondary patents and the NCE exclusivity.
Core protection end — take the compound patent date, including pediatric and PTE extensions, or the NCE exclusivity date, whichever is later.
Challengers — check the FDA Paragraph IV list for the number of filers and the first filing date.
Generics already approved — search the Orange Book for ANDA products on the same ingredient.
API suppliers — filter the FDA DMF list by ingredient, keep the active Type II files, and note which holders are Indian.
Other countries — look up the patent family on Google Patents to see where it was filed; for India, check status on InPASS.
Crowd estimate — combine the number of challengers, DMFs and approved generics, and read them against the price-drop figures above.

Doing this once is a useful exercise. Doing it for 25 molecules every week, and catching each change as it happens, is closer to a full-time job. That is the part I decided to automate.

What I'm building

I have a first version of an engine that runs the checklist above automatically, for every innovator drug in the Orange Book. The underlying data is already in place:

FDA Orange Book — products, listed patents with their type flags, exclusivities, approved generics
FDA Paragraph IV list — challengers, first filing dates, 180-day status
FDA DMF list — active API suppliers per molecule, with an estimate of which are Indian
CMS Medicare Part D — market size
PubChem — chemical identity, so salts and forms of the same molecule are grouped together
Google Patents families (being added) — filings and estimated expiry in India, the EU, China and Brazil

For each molecule the engine computes the core protection end, the length of the secondary-patent tail, an India Section 3(d) flag, and an opportunity score from 0 to 100. The score weights market size at 35%, timing at 25%, expected crowding at 25% and API availability at 15%. It runs weekly and flags what has changed, such as a new challenger, a new DMF or a new generic approval.

The output is a watchlist plus a one-page snapshot for each molecule, and a scientist reviews every report before it goes out. A few results from the current top 10 are worth noting:

Molecule	Public signal	What it implies
Ruxolitinib	1 challenger, 11 Indian API makers already positioned	An open lane with easy sourcing
Empagliflozin	Large market, many entrants	Big, but already crowded
Linaclotide	Only 2 Indian API suppliers	A supply gap, or an opening

Each represents a different business decision, which is the main value of looking at them side by side.

Where it stands

The US engine works. These parts are still in progress:

Settlements and litigation outcomes are not yet modelled, so a settled entry date can differ from the date on paper.
India and other-country dates depend on patent-family mapping plus a manual check on InPASS, because India has no patent API.
The Indian DMF flag is an estimate based on company names.
Biologics and biosimilars (the Purple Book) are not covered yet; Keytruda-type molecules are the next step.
Claim-level reading of patents — for example, judging whether a specific polymorph patent can be designed around — is planned, with expert review.

To be clear about its scope: this is market intelligence derived from public regulatory data. It is not a legal freedom-to-operate opinion.

How you can be part of it

I would rather build this with people from the industry than guess at what they need. If any of this is close to your work, here are a few ways to connect:

Try it. If you work at a generic or API company, send me one or two molecules you are tracking and I will return a free snapshot — dates by country, the expected level of competition, and the API suppliers. In return I ask only for honest feedback on what is useful and what is missing.
Contribute. If you know patents, regulatory, sourcing or BD, I would welcome your input on the logic, particularly on settlements, India dates and biologics.
Tell me what you need. If you would want something like this built around your own portfolio, I'm happy to talk. The data and the engine exist; they can be shaped to a specific use case.
Introduce someone. If you know someone in pharma BD, portfolio planning or regulatory who would find this useful, an introduction would be appreciated.

You can reach me through a comment, a message on LinkedIn, or my site, jyotsnapriyam.com.

Sources