190 high-earning drugs are likely to lose patent protection by 2030, clearing the path for generics and biosimilars to capitalize—but not all R&D teams are set up to succeed.
In the age of AI, research intelligence has to begin with trusted data. This series, Inside Our Data, explores the foundational data infrastructure that makes trust possible.
According to Deloitte, 190 drugs—36% of which are considered “blockbusters”—are set to lose patent protection by 2030.
This means pharma companies will be able to compete with these products, developing and distributing generic and biosimilar versions. At the same time, those firms who have held exclusivity will be looking for new opportunities, as greater availability will swiftly drive the prices of those drugs down.
With an estimated $236B in revenue at risk, pharmas are racing to ensure swift and sound development work, often by establishing trustworthy, reliable, and explainable AI workflows.
But when it comes to AI, garbage in equals garbage out.
In order to minimize risk and capitalize on market opportunity, there’s one thing R&D teams must have: rich up-to-date patent data.
Patents can hold both risk and opportunity
A patent record contains myriad details that set out where and how an enterprise can operate, but what matters most is its legal status—who owns it, is it in force, and how might this vary across jurisdictions? Depending on the answers to these questions, the same record can mean protection, opportunity or risk for an R&D team.
Parsing the status of patent records is made more complex because status can change subtly and between jurisdictions. Status differs by country, so single patents active in some designated states may be expired in others.
A patent’s term can also vary by country. While pharmaceutical patents typically last for 20 years from filing, they can be extended up to five years in the EEA and US. For example, the semaglutide patent expired in March 2026 in many countries including China, while extensions stretch to 2031 in Europe and to at least 2032 in the US. These extensions are available because drugs can’t be sold until they clear years of clinical trials and regulatory review, such as FDA approval, which uses up part of the 20-year patent term before they some of that time, the US and Europe let companies extend their patent protection by up to five years.
This means—even though a patent’s legal status is its most important characteristic—there’s often no simple answer to “is it in force?”
And each patent office, across the world, has its own systems: they use their own delivery structure, terminology and own original language, meaning raw records are nearly impossible to automatically compare.
With millions of changes to record legal statuses every week globally, it can be difficult for R&D teams to be certain of their freedom to operate. But the cost of getting this wrong can be massive.
Stale patent records can create existential risk
The first step of R&D is defining the freedom to operate; teams must search for in-force patents and published applications in their target jurisdictions. A lack of patents, or an expiring patent, can signal a new opportunity for a firm to expand.
This is the first instance where stale patent data creates risk. A stale record can show a space in the market that no longer exists, or can show a patent to be in force when it isn’t.
The risks associated with relying on stale data can be existential for firms.
The median cost of bringing a drug to market is estimated at ~$985M—if a firm invests even a portion of this expenditure before realizing they lack freedom to operate, those costs may be sunk.
This same logic can protect, or expose, a firm’s own portfolio. An expiry in one jurisdiction is visible to competitors, and patents are costly to obtain and maintain, so knowing what to renew or let lapse depends on having a clear view of current records.
AI workflows make bad patent data more expensive
AI workflows are allowing R&D teams to work faster and smarter all over the world. These workflows can form a critical part of setting a strategic course—and anticipating the direction of travel for competitors.
But stale data means bad outputs and material enterprise risk. An agent which accepts an outdated record can tell a company that a patent isn’t in force when it actually is—and the AI won’t know to question it.
AI relies on structured and consistently labeled data. Mergers, divestments, name changes, and spin-offs with flawed corporate name mappings undermine the validity of AI outputs. The quest for consistency in patent data is troubled further by imprecise or varying translations of multi-lingual inputs from around the globe.
The only way to avoid this risk is using fresh, complete, curated data to feed those AI workflows. But the cost of preparing that data in-house is massive—and there are millions of record updates around the world every single week.
To ensure teams have access to fresh and complete patent data from around the world, enterprises can access purpose-built curated data suites, such as IFI CLAIMS by Digital Science.
Enterprises need to make the most of the data available
IFI CLAIMS has been addressing patent data challenges for more than 70 years. Patent data from around the world is captured, standardized, and unified into one of the industry’s most complete and dependable platforms, continuously updated and growing as patent offices deliver new information, giving teams a single reliable source for the latest global patent intelligence.
IFI’s current data coverage includes 105+ countries with bibliographic data, 62 with full text and 180 million+ patent records. Full text and attachments give AI and data science teams the detail behind the invention. Attachments include cover pages, specifications, drawings, reference images and chemistry files. Patent data is ingested, translated and parsed automatically, and checked for both technical validation before going live, and continuous editorial domain expertise oversight.
IFI supplies expiry dates, regulatory-extensions flags, terminal-disclaimers flags, standardized names and maps patents back to ultimate owners, their parent companies.
These millions of weekly updates are fed live into IFI client databases, so they always reflect the latest status and ownership updates—ensuring high quality, accurate data and drastically reducing the ambiguity which often undermines an AI workflow.
Patents form the backbone of a connected research lifecycle
Enterprises that can harness AI efficiently will pull ahead of competitors in the coming years, armed with greater competitive insight, strategic direction, and exposure mitigation.
Those enterprises have the most to gain by embracing a fully connected research lifecycle with an intelligence partner like Digital Science. This lifecycle begins with Dimensions, which links grants, publications, citations, clinical trials, patents and policy documents in one place, covering how research is funded, published and moved toward clinical use.
IFI CLAIMS supplies accurate and up-to-date patent data complete with ownership and legal status information that help define the R&D roadmap. And Altmetric captures mainstream and academic mentions of research from sources including news, blogs, policy documents, social media and patents, with 256m+ mentions of over 24 million research outputs.
Applying an AI layer across all three of these tools can accompany a target from grant to publication, through trial activity and attention, to the strategic maintenance of the patent’s position and legal status.
The common thread here is curated, structured, high quality data at every stage—from the first freedom to operate search to strategic patent extension decisions 20 years later.
With the right tools, patent data can be a strategic lever
AI is necessary for enterprises to remain competitive in today’s markets. But it’s only as good as the data under it, and for patents that means current status, correct ownership and country-by-country accuracy.
Armed with trustworthy, always-fresh patent data, enterprises can get a clear and early view of an industry’s directions, define their own priorities, and seize new opportunities as they arise.
IFI CLAIMS, as part of the Digital Science ecosystem, makes this possible.
