‍Key Takeaways:

• The R&D Pipeline is Shrinking: For the first time in 30 years, the global biopharma R&D pipeline contracted in 2026, dropping 3.9% as companies prioritized capital discipline over volume.
• The Market Is Favoring Hard-To-Treat Indications: Incremental updates are being ignored by the market, while undeniable survival benefits in hard-to-treat indications (like Revolution Medicines’ daraxonrasib in pancreatic cancer) are being heavily rewarded.
• Patent Cliffs are Forcing Selectivity: With $300 billion in prescription drug revenue exposed to generic competition by 2030, Big Pharma is ruthlessly evaluating internal pipelines and external innovation.
• AI is the New Force Multiplier: Platforms like Reliant AI are automating literature review and evidence synthesis, allowing teams to meet these stricter evidence standards without sacrificing speed.

According to the European Federation of Pharmaceutical Industries and Associations (EFPIA), over the past two decades, the share of pharmaceutical R&D spend between Europe, the US, Japan, and China has decreased from 37% in 2010 to 32% in 2020, with projections indicating a further decline to 21% by 2040 if no countermeasures are taken. To this, we ask, what if the most important signal in biopharma right now is not any single breakthrough, but the tension between scientific progress and economic pressure? 

The same report also put this into context that feels less macroeconomic and tangible: of every 100 drugs that begin human testing, only around 10 reach patients. This is a productivity reality that makes today's cost pressures and shortening asset lifecycles harder to absorb, and for patients, a reality that may mean fewer life-saving drugs.

That tension is everywhere in the conversations biopharma professionals are having right now — across Reddit, at conferences, and even within the Reliant AI office. The industry is still producing meaningful innovation, but it is doing so in a far more disciplined, scrutinized, and technology-enabled environment than many teams were built for.

Clinical Wins Now Have to Carry More Weight

Not all positive trial results land the same way. Incremental progression-free survival updates in crowded categories can pass through industry conversations almost unnoticed. But when a company shows a survival benefit in a difficult indication, such as that for Revolution Medicine’s daraxonrasib, where the underlying biology has resisted decades of effort, the signal is unambiguous. The market is rewarding programs that change what is possible.

For example, in November of last year, Revolution Medicines reported that its RAS inhibitor had hit its primary survival endpoint in a phase 3 pancreatic cancer trial. Pancreatic cancer is one of the hardest indications in oncology. KRAS has been one of the hardest targets in drug development for four decades. Pancreatic cancer is one of the hardest indications in oncology, and KRAS has been one of the most notoriously “undruggable” targets in drug development for four decades. 

That is why the Phase 3 readout from Revolution Medicines did more than move a single stock. It reinforced something the industry has been quietly reorganizing around: the threshold for a meaningful clinical win has risen.

The Market Impact of Paradigm-Shifting Overall Survival (OS) Data in Oncology

Data Sources:

1. Revolution Medicines (April 2026): Daraxonrasib Demonstrates Unprecedented Overall Survival Benefit in Pivotal Phase 3 RASolute 302 Clinical Trial in Patients with Metastatic Pancreatic Cancer. 2. OncLive (May 2026): FDA Green Lights Expanded Access Protocol for Daraxonrasib in Pretreated Metastatic PDAC.
2. The ASCO Post (March 2024): Enfortumab Vedotin-ejfv Plus Pembrolizumab for Locally Advanced or Metastatic Urothelial Cancer (EV-302 Trial Data).
3. Pfizer (December 2023): Pfizer Completes Acquisition of Seagen for $43 Billion to advance Antibody-Drug Conjugate (ADC) Technology.

Two weeks later, Fierce Biotech published a separate report, which drew almost as much attention. For the first time in thirty years, the biopharma R&D pipeline had contracted.  According to Fierce Biotech, the overall number of drugs in development fell to 22,940 at the start of 2026, representing a 3.9% decline from the previous year. In contrast, from 2001 to 2025, the pipeline experienced uninterrupted year-over-year growth, nearly quadrupling from 5,995 to 23,875 drugs. This recent dip marks the first recorded decline since the mid-1990s.  

These two facts are sometimes treated as a contradiction. Many read them as proof that scientific progress and commercial reality are pulling in opposite directions. We think they tell a more useful story. The industry is not slowing down, but the bar — for what counts as an investable, defensible program — has moved. What it takes for a program to be considered investable, scalable, and strategically defensible has gone up, and the teams that recognize that shift are the ones quietly redefining how decisions get made.

Clinical Wins Now Have to Carry More Weight

Not all positive trial results land the same way. Incremental progression-free survival updates in crowded categories can pass through industry conversations almost unnoticed. But when a company shows a survival benefit in a difficult indication, such as that for Revolution Medicine’s daraxonrasib, where the underlying biology has resisted decades of effort, the signal is unambiguous. The market is rewarding programs that change what is possible. 

That is why the Phase 3 readout from Revolution Medicines did more than move a single stock. The data were undeniable: a 13.2-month median overall survival versus 6.7 months for standard chemotherapy, representing a 60% reduction in the risk of death. We saw this same market reward play out recently when the Padcev and Keytruda combination obliterated the 30-year chemotherapy standard in advanced bladder cancer (31.5 months vs. 16.1 months median overall survival), driving a $43 billion acquisition.

These are not just clinical wins; they are paradigm shifts. They reinforce something the industry has been quietly reorganizing around: the threshold for a meaningful clinical win has risen. “Interesting” no longer earns a program a place in a portfolio. Programs are increasingly expected to demonstrate not just biological plausibility but durable differentiation, defensible economics, and a clear path to clinical adoption.

Trial design, endpoint selection, biomarker strategy, and translational evidence are no longer just internal scientific questions. They are commercial questions. The strongest programs are now the ones whose evidence packages can survive scrutiny from medical reviewers, payers, regulators, and partnership negotiators, often simultaneously.

A Smaller Pipeline Is Not the Same as a Slower Industry

The contraction of the R&D pipeline — the first such decline since the early 1990s — has drawn the most attention in recent industry coverage. The instinctive read is that the industry is shrinking, but the more accurate read is that the industry is becoming more selective. The same Fierce Biotech report notes that major players are actively trimming their pipelines. For example, Pfizer cut its R&D pipeline from 271 to 257 molecules, and Novartis trimmed 10 programs, signaling a broader industry move toward capital and pipeline discipline rather than pure volume

A smaller pipeline can reflect tougher funding conditions, higher development costs, stricter evidence standards from regulators and payers, and investor pressure on companies to stop carrying weak assets. In practice, all four are operating at once.

This is not a pessimistic view. A more disciplined pipeline can produce better science, sharper portfolio decisions, and a clearer relationship between investment and impact. But it does require teams to operate differently. Activity becomes a poor proxy for value. 

The questions that matter shift from “what are we doing?” to “what is this doing for the decision the organization needs to make?”

Capital Discipline and Lifecycle Risk Are Now Visible to Everyone

Two related industry stories help frame why this matters now.

The first is patent expiration risk, which has moved from a future-state projection to a present-day earnings reality. In its Q1 2026 earnings report, Novartis reported a 14% decline in core operating income—driven by the loss of exclusivity and steep generic erosion for blockbuster drugs like Entresto in the U.S. Novartis is not an outlier here; it is the leading edge of a much wider trend. Industry analytics firm Evaluate projects that roughly $300 billion in prescription drug revenue will be exposed to generic and biosimilar competition by 2030. That is roughly double the exposure of the last major patent cliff, and it affects almost every major player, from Merck bracing for Keytruda's eventual expiration to Bristol Myers Squibb navigating the runway for Eliquis and Opdivo.

The resulting revenue cliffs are ruthlessly reshaping how companies allocate R&D capital today. To plug these looming revenue holes, organizations are trimming low-probability internal programs and cutting headcount to free up cash, which brings us to the second story: the shape of dealmaking.

Gilead’s recent $45 million deal with Kymera to add a preclinical protein degrader to its oncology pipeline is part of a larger pattern. Large companies remain active in external innovation, but they are pursuing it earlier, more selectively, and with sharper portfolio theses than past cycles suggested. External innovation is still an essential growth lever, but it is increasingly being asked to fit a strategy, not define one.

The implication is that portfolio context now travels with every program. A clinical lead asking why their compound suddenly faces tougher prioritization questions cannot answer that by looking only at their own data. The answer often lives in patent timelines, deal flow, indication concentration, and capital allocation decisions happening elsewhere in the organization.

AI Is Becoming a Force Multiplier

Alongside the discussion of patent cliffs and pipeline pressure, a much less dramatic conversation is reshaping how biopharma teams operate day-to-day. Across functions, employees are using AI for tasks that used to be invisible: literature triage, first drafts of communications, structured extraction from PDFs, and the assembly of evidence summaries for cross-functional review.

What is striking is how unremarkable this has become. The question has shifted from “should we use AI?” to “where is it actually useful?” That shift signals that AI in the life sciences is moving past its hype phase and into a phase where it accumulates compound returns through small, repeatable productivity gains embedded in real work.

This matters in the context of broader industry pressure. When portfolios are smaller and decisions are sharper, time savings from automating repetitive evidence work are not just productivity gains. They are decision-quality gains. They free senior analysts and strategists to spend more time on interpretation, judgment, and synthesis: the parts of the job that AI is least equipped to handle, and that the new bar most rewards.

Where Reliant Fits

This is the environment Reliant AI was built for.

The most consequential shift in biopharma right now is not any single therapeutic breakthrough or any single technology. It is the rising bar for what counts as a defensible decision. Programs are being scrutinized harder. Capital is being allocated more selectively. External innovation is being evaluated more rigorously. And the teams in the middle of all of this are being asked to move faster, with sharper evidence, in workflows largely designed for a slower era.

The timing matters. Representation learning, probabilistic reasoning, and human-in-the-loop systems are mature enough to coexist. Organizational expectations have shifted towards transparency and greater awareness of uncertainty. And the volume of biomedical knowledge has exceeded the capacity of linear search and narrative review.

But there is a larger macroeconomic forcing function at play. Reliant’s approach works not because it promises omniscience, but because it is designed for life sciences realism—both scientific and economic. As the industry enters a new era of discipline, the cost of a poor portfolio decision has never been higher. As shown above, the global R&D pipeline is contracting. Capital allocation is tightening amid looming patent cliffs. Most importantly, the threshold for clinical and commercial success has risen sharply; the market no longer rewards incremental updates, but demands unambiguous, paradigm-shifting data in notoriously difficult indications. Teams can no longer afford to advance programs based on brittle, deterministic data or isolated findings.

The original vision of the knowledge graph didn't fail; its rigid structure simply wasn't fit for purpose in a fundamentally probabilistic domain. But today, the stakes have changed. When clearing the "new bar" requires simultaneously synthesizing conflicting trial endpoints, shifting biomarker definitions, and competitive survival data, legacy infrastructure breaks down.

By keeping the best of relational mapping and leaving behind the deterministic brittleness of the past, we are no longer just building knowledge bases. Now, with the right architecture and the right humility, we are building Evidence Graphs. And they can finally do the work this industry has always needed: empowering teams to separate the signal from the noise and make high-stakes decisions with speed and conviction.

For those new to it, an Evidence Graph is more than just a knowledge graph by another name. It preserves the relational structure that made knowledge graphs valuable for entities, connections, and traversable context, but treats every connection as a claim backed by evidence, not a fact set in stone. Findings carry their provenance with them: the underlying study, the population, the endpoint definition, the date, and the strength of the signal. Conflicting evidence is represented as conflicting evidence, not flattened into a single “truth.” That is what makes the structure usable in a domain where the answer to almost any meaningful question is “it depends on the source.” We’ve written more about how this works and why it matters in our thought piece on Evidence Graphs.

That source-level transparency is not incidental. In an environment where decisions are being scrutinized more closely than ever, the ability to show where a finding came from is the difference between a recommendation that moves a program forward and one that gets sent back for verification.

The next era of biopharma will not be defined by which organizations chase the most ideas. It will be defined by which organizations can tell the signal from the noise and act on it with speed, discipline, and humility. The teams that get there first will be the ones that took the new bar seriously.

Which leaves a question worth debating: as scientific progress continues, what will matter more over the next three years: better science, better capital discipline, or better use of AI to connect the two?

If you’re exploring how to bring greater rigor and speed to your team’s evidence and decision-making workflows, our team is happy to talk through what that would look like in practice. Book a demo or reach out directly.

Frequently Asked Questions

Why did the biopharma R&D pipeline shrink in 2026?

The pipeline contracted by 3.9% at the start of 2026 due to higher development costs, stricter evidence standards, and the looming $300 billion patent cliff. Companies like Pfizer and Novartis are actively trimming weak assets to focus capital on highly differentiated, investable programs.

How is AI being used in biopharma R&D today?

AI is moving beyond hype into practical workflows. Biopharma teams use AI platforms like Reliant AI to automate literature triage, extract structured data from PDFs, build competitive landscapes, and assemble evidence summaries for cross-functional review.

What is the new bar for clinical trials in oncology?

Incremental progression-free survival (PFS) is no longer enough. The market now rewards unambiguous overall survival (OS) benefits in historically difficult indications.

‍