Why this autumn matters
Cancer drug development is not one story but several running at once, and this quarter they come to a head. The first catalyst has already arrived. On October 8, topline Phase 3 results from the EXCALIBER-RRMM trial showed that the ZDd regimen delivered a median progression-free survival of 42 months, against 20 months for the DVd comparison arm. The hazard ratio was 0.49 (p<0.000001), which the results describe as a 51% reduction in the risk of progression or death. Median follow-up was 23 months, and safety was reported as consistent with prior findings.1 In plain terms, patients on one regimen went about twice as long without their disease worsening. That is a topline result, though, and the full data are still to come.
Two further dated events sit ahead of us. Moderna and Merck plan to present detailed results for their personalized mRNA cancer vaccine, intismeran, at the ESMO cancer congress on October 24, and to engage with regulators about potential filings.2 Three weeks later, on November 14, 2026, the US Food and Drug Administration has a decision date for Summit Therapeutics' application for ivonescimab. That application rests on results from the HARMONi trial.3 Neither outcome is known. Both are scheduled, and a scheduled date is not a result.
Combinations are becoming the plan
On October 2, Summit announced a clinical trial collaboration with Daiichi Sankyo and AstraZeneca. It will test ivonescimab alongside Datroway across several solid tumor settings, intending to include breast and lung cancer.4 The significance is the pairing. Instead of asking whether one new drug beats the current standard, the industry is increasingly testing two different kinds of therapy together.
Europe offers a related signal. On July 24, the European Medicines Agency's CHMP committee adopted a positive opinion recommending Gilead's Trodelvy plus Keytruda as a first-line treatment for PD-L1-positive metastatic triple-negative breast cancer.5 That is a recommendation, not a final authorization. Evandro de Azambuja, speaking in the announcement, said: "People with metastatic triple-negative breast cancer need effective treatment options as early as possible in the course of their disease, particularly when their tumors express PD-L1."5 Telix, meanwhile, dosed the first patient in LUTEON, which it describes as the first Phase 3 study of a radiopharmaceutical therapy for relapsed or recurrent clear cell renal cell carcinoma.6
What the SEC filings show about one big player
Merck is the one company in this story whose finances we can state as checked fact. Its figures come from SEC filings and were verified against that source. Merck is co-developing intismeran with Moderna. We should be plain about the limits: these filings say nothing about intismeran's cost or prospects. They only show how a very large company is funding its overall business.
Merck's capital spending, the money it puts into plants, labs and equipment, was $3.863 billion in 2023 and $3.372 billion in 2024. It then rose to $4.112 billion in 2025, roughly 22% higher than the year before.7 For scale, that 2025 figure is not far short of the $4.395 billion Merck reported as its cost of revenue in the second quarter of 2026 alone.8 The first-quarter pattern is less tidy: $861 million in 2024, $1.328 billion in 2025 and $991 million in 2026.7 A single quarter can mislead, and the full-year trend is the steadier guide.
The cash picture is more volatile. Merck held $6.841 billion at the end of 2023, $13.242 billion at the end of 2024 and $14.565 billion at the end of 2025.9 The quarterly readings in between moved a great deal, peaking at $18.169 billion in the third quarter of 2025. Cash then fell to $5.327 billion in the first quarter of 2026 and stood at $6.849 billion in the second quarter.9 Our dataset does not say why cash fell by about $9 billion between the end of 2025 and March 2026, and we will not guess. Readers should treat it as a question for Merck's own filings. It would be wrong to read it as a verdict on the company.
Merck's cost of revenue, what it spends to produce what it sells, was $16.126 billion in 2023, $15.193 billion in 2024 and $16.382 billion in 2025.8 Quarterly it rose from $3.419 billion in Q1 2025 to $4.195 billion in Q1 2026, and from $3.557 billion in Q2 2025 to $4.395 billion in Q2 2026.8 That is a rise of about 23% in both quarters. Without Merck's revenue in our dossier, we cannot say whether sales rose by as much, so this is a cost trend, not a margin trend.
Uneven money across the sector
Commercial companies and early-stage ones are in very different positions. Novocure reported second-quarter 2026 net revenues of $184 million, up 16% year on year, with 18% growth in global active patients. It also received a CE Mark for Optune Pax in locally advanced pancreatic cancer, with Germany the first EU market to launch and more than 280 active patients on the device as of June 30.10 Frank Leonard, speaking for the company, said: "This was our strongest quarter to date, with record net revenues and active patients on therapy."10
At the other end, ADC Therapeutics announced job cuts that it projects will deliver "$10 million in annualised cost savings". It says it "maintains a cash runway into 2028, supporting ongoing regulatory and clinical plans for ZYNLONTA combinations."11 Cutting staff to stretch cash is a rational choice for a small developer, and it also shows how little room such firms have if a trial disappoints.
The AI-discovery pitch, and how far to trust it
AstraZeneca is positioning artificial intelligence as a way to shorten discovery timelines. Puja Sapra, in an MIT Technology Review piece, describes a "lab of the future" in which "a self-driving car uses sensors and models to navigate its environment, this system uses AI to make predictions, robotic systems to execute experiments, and instruments to generate data."12 She names proprietary data as the edge: "Data is our differentiator."12 She also concedes the hard part: "One of the hardest problems in de novo design is predicting whether a computationally generated molecule will be safe in the human body."12 And she stresses that "human oversight remains at the heart of this approach."12
Two cautions apply. First, this is a company executive describing a vision, with no trial results or timelines attached. Second, our own source check is unflattering: of 11 claims we checked from that MIT Technology Review document, none held up. We do not take that to mean Sapra's words are misquoted, since the quotes above are verbatim. It does mean the surrounding article's factual claims should not be treated as established, and nothing in it shows that AI-designed drugs have yet reached patients.
How much to trust each source
We measure how often claims drawn from a source have held up. The NewsEOD press-release feed carrying the Novocure and Gilead items held up in 56% of 4,956 checked claims. The feed carrying the ADC Therapeutics item held up in 45% of 520, and the Telix item's source in 31% of 2,928. These are company announcements, so the quotes are reliably the companies' own words, but the figures inside them deserve independent confirmation. Merck's SEC-sourced numbers are the firmest ground in this piece.
What to watch
- October 24: whether the detailed intismeran data at ESMO match the headline expectations, and what Moderna and Merck say about regulatory filings.
- November 14: the FDA's decision date for ivonescimab. A decision on that date is not guaranteed by the date itself.
- EXCALIBER-RRMM: the full data behind the 42-month versus 20-month topline, including safety.
- Merck's next filing: whether cash rebuilds from $6.849 billion and whether capital spending continues above 2024's level.
- Smaller developers: whether cash-stretching cuts like ADC Therapeutics' hold the 2028 runway.
