Roche faces a run of FDA decisions while older drugs fade

Roche is two businesses under one roof: a large drug company and a large diagnostics group. In the first nine months of 2026, pharma sales were CHF 36.2 billion and diagnostics CHF 10.3 billion. The investor question is less whether any single trial works and more whether a growing pipeline can replace products that are losing exclusivity.

The technology

Roche does not rest on one platform. The pipeline spans small molecules, antibodies, bispecifics, antibody-drug conjugates and, in obesity, injectable incretins. The late-stage assets are specific. Giredestrant is an oral SERD with a 30% reduction in recurrence or death in early breast cancer. Divarasib is a KRAS G12C inhibitor that Roche says beat sotorasib or adagrasib in a head-to-head phase III. Fenebrutinib is a BTK inhibitor with FDA filings accepted in both relapsing and primary progressive MS. Enicepatide is a GLP-1/GIP agonist with phase II diabetes data.

The diagnostics arm adds a second route to value. Roche's Elecsys pTau217 blood test for Alzheimer's pathology received US FDA clearance, according to the third-quarter release, and the company says it is pairing diagnostic and drug expertise in new programs.

The odds

Roche has an unusual number of filings at once. The FDA has set dates of November 30 for giredestrant in early breast cancer and December 18 for advanced disease. Gazyva in primary membranous nephropathy is due in November, and Lunsumio plus Polivy in large B-cell lymphoma by February 9, 2027.

The published base rates are favorable at this stage. In the BIO, Informa and QLS study of 2011 to 2020, 57.8% of phase III programs reached an NDA or BLA filing, and 90.6% of filings won approval. For all programs starting in phase I, only 7.9% reached approval. Roche's near-term assets are already past most of that attrition. The earlier ones, such as enicepatide in diabetes, face the lower phase II to III rate of 28.9%.

Roche has also had misses. In persevERA, giredestrant plus palbociclib missed its primary endpoint in first-line advanced disease.

The money

Roche is profitable and large. First-half sales were CHF 30.4 billion with core operating profit of CHF 11.9 billion, up 10% at constant rates. The half-year report shows free cash flow of CHF 4.2 billion and net debt of CHF 22.1 billion at June 30, up from CHF 16.2 billion at the end of 2025 after CHF 8.3 billion of dividends. The company expects to raise its dividend in francs again.

Guidance for 2026 is mid single digit sales growth and high single digit core earnings per share growth at constant rates. Roche confirmed that on October 22 after sales grew 6% in nine months.

The risks

Currency is the first. At constant rates, nine-month sales grew 6%, but in francs the number was 1%. Second is erosion of older products: Perjeta, Kadcyla, Actemra and Herceptin all fell in the nine months, according to the sales table. Third is concentration. The five growth drivers (Xolair, Hemlibra, Ocrevus, Phesgo and Vabysmo) are about 46% of pharma sales. Vabysmo grew 2% in the US against 65% in the International region.

Diagnostics faces healthcare pricing reforms in China, which held back growth in the division and cut Asia-Pacific sales by 4% at constant rates.

Our read

Roche looks like a company with near-term news flow and a steady base, not a binary bet. Two readouts could change the picture. Giredestrant in the adjuvant setting would be a new oral option in breast cancer, and divarasib's full Krascendo 1 data, which Roche has not published yet, will show whether "best-in-class" holds up. Enicepatide gives Roche a position in obesity, but it is early and the competition is well established.

The balance sheet is large enough to fund this. The main thing to watch is whether new launches grow faster than the older products decline. The next several weeks bring the FDA's first answers.