Oncology R&D in 2026 feels like a turning point, the rules of the game are quietly being rewritten around integration of science, data, and real‑world care. The organizations that thrive from here will be those that can connect these dots with discipline, not just ambition.
If the last decade in oncology was about discovering new toys, this one is about learning to use them wisely. Pipelines are now packed with antibody–drug conjugates (ADCs), bispecifics, cell and gene therapies, radioligand therapies, and increasingly sophisticated small molecules. For leadership teams, the core question has shifted from “Do we have an ADC?” to “Do we understand where an ADC genuinely changes the therapeutic equation for a specific biology and patient population?”
ADCs, for example have become one of the most dynamic segments of the oncology market, with hundreds of trials and billions in global spend, particularly in solid tumors. But the strategic value of an ADC does not come from the modality label; it comes from getting three things right: target selection, payload strategy, and combination design. The same logic applies to bispecifics and radioligands.
The more mature oncology organizations are now doing three things differently:
This is a subtle but important mindset shift. Instead of chasing every emerging modality, leading teams are curating a deliberate toolkit aligned to the clinical questions that matter in their chosen tumors.
We talk a lot about precision medicine, but in oncology R&D is quietly becoming the operational baseline. The most forward‑leaning sponsors now assume that biomarker strategy, diagnostics, and real‑world data will shape every phase of development, from FIH to post‑marketing.
In early development, I see the real differentiator as how seriously teams take translational coherence. It is not enough to “include” ctDNA, minimal residual disease, or complex signatures in the protocol. The critical questions are: Are these markers driving inclusion and escalation decisions? Are they wired into early stop/go criteria? Do they genuinely help the team say “no” sooner when the biology doesn’t hold?
Regulators are pushing in the same direction. Initiatives focused on dose optimization and better characterization of exposure–response are forcing oncology away from a “maximum tolerated dose at all costs” mindset. This is uncomfortable for programs that have relied on aggressive dosing to eke out marginal signals, but it is absolutely the right direction for patients and for long‑term asset value.
Where I think the field is still under‑innovating is in how we connect precision development with real‑world practice. Too often, we design exquisitely biomarker‑driven Phase II/III trials and then release those therapies into clinical environments where diagnostic access, turnaround times, and data capture lag far behind. The next wave of leadership in oncology R&D will come from teams that design trials with implementation in mind: understanding where the test will actually be run, how quickly, for whom, and at what cost.
AI inevitably enters every oncology conversation now, but the signal‑to‑noise ratio is still painfully mixed. The interesting shift in 2026 is that the conversation has become more pragmatic: less about building “AI strategies” and more about embedding AI into specific, high‑value decisions.
In my experience, the meaningful use cases cluster around four areas:
My litmus test for AI in oncology R&D is simple: if you removed the AI tomorrow, would the decision process meaningfully degrade? If the honest answer is “not really,” then you are probably still in experimentation mode, not impact mode.
Oncology has historically enjoyed a kind of “special status” in portfolios. That landscape is maturing. Capital is more selective, regulators are more demanding, and trial complexity is higher. The result is a quiet but important pivot from volume to quality.
We are already seeing:
I view this as a necessary reset. Oncology R&D cannot sustainably be a game of running as many shots on goal as possible and hoping a few cross the line. The bar is moving toward a clear mechanistic hypothesis and a line of sight to real‑world adoption.
It is no longer enough to champion bold science; you also have to be comfortable saying “stop” when the evidence doesn’t justify the next increment of spend, and to do so early, while you still have room to redeploy capital and talent.
If I look across the different strands, modality evolution, precision, AI, capital discipline, a picture emerges of what “good” looks like for oncology R&D in 2026. It is simply about a consistent, intellectually honest way of making decisions.
Three traits stand out in organizations that are setting the pace:
Oncology R&D has always been a space where ambition runs high. In 2026, the differentiator is more about who is the most disciplined in translating that ambition into decisions that stand up to the lived reality of patients and health systems.