Exelixis EXEL buyback news

Opinion · Devil's advocate

Fact-check-o-meter

Exelixis is mostly candid, but its zanzalintinib story leans on a trial that missed half its primary goal

Exelixis's releases match its 10-Q on the numbers, and the company disclosed the FDA review extension quickly. The gaps sit in framing: the August release still pointed to a colorectal approval later this year, the missed non-liver-metastases endpoint is described as a favorable trend, and a claim about atezolizumab's contribution comes from a trial with no zanzalintinib-alone arm. The 10-Q shows a growing list of Cabometyx generic challengers, including an FDA tentative approval for Handa that the results release did not mention.

1. Press release vs filing · weight medium

   'Later this year' approval became a March 2027 PDUFA date

   The company said: "advancing preparations for the potential launch of zanzalintinib in metastatic colorectal cancer, pending approval from regulatory authorities later this year." source

   The record says: Five weeks later the 8-K said the FDA "had extended the review period" by three months. "In response to an FDA information request, Exelixis had submitted updated safety and efficacy data, which the FDA has deemed a major amendment. The updated Prescription Drug User Fee Act action date is March 3, 2027." source

   Our view: The August wording was accurate when written, and Exelixis filed the change within a day. Still, a major amendment means the FDA asked for more data, and the 8-K does not say what the request covered. Launch timing for the company's next product has moved into 2027.

2. Science · weight high

   STELLAR-303 missed its second primary endpoint

   The company said: "The results showed a non-statistically significant trend in OS favoring the combination in the NLM subgroup." source

   The record says: ClinicalTrials.gov lists two primary outcomes, overall survival in "All Randomized Participants" and in "Randomized Non-Liver Metastases (NLM) Participants". The 10-Q gives the NLM result as "stratified HR: 0.83" with "P=0.1185" and a confidence interval that runs from 0.66 to 1.05, against a stratified HR of 0.80 ("P=0.0045") in the full population. The registry lists the trial as open label (masking: none). source

   Our view: Calling a failed primary endpoint a favorable trend is common, but the interval crosses 1, so the trial did not show a survival benefit in patients without liver metastases. The approval case rests on the intent-to-treat result alone, and how the FDA weighs the NLM miss is an open question.

3. Science · weight medium

   The atezolizumab contribution claim lacks a randomized comparison

   The company said: "The STELLAR-303 data support the contribution of atezolizumab to the previously observed survival benefits of zanzalintinib in combination with atezolizumab for patients with metastatic CRC." source

   The record says: The registry shows two arms only: "XL092 + Atezolizumab" and "Regorafenib". There is no zanzalintinib-alone arm. The same release adds that the poster "demonstrated the importance of generating data around the potential impact of anti-drug antibodies". source

   Our view: Without a zanzalintinib-only arm, any estimate of what atezolizumab adds comes from post hoc or exploratory analysis, not randomization. The claim is better read as supportive analysis than as proof.

4. Legal and disclosure record · weight high

   MSN appeal won, but more Cabometyx challengers are lined up

   The company said: "the effective date of any final approval by the U.S. Food and Drug Administration of MSN's ANDA shall not be a date earlier than January 15, 2030" source

   The record says: The 10-Q lists further cases in the U.S. District Court for the District of Delaware: a consolidated MSN and Azurity trial "scheduled for November 2, 2026", a July 6, 2026 suit against Handa, and a July 6, 2026 suit against Accord, plus a Paragraph IV notice from Almatica. It adds: "On July 29, 2026, the FDA granted tentative approval to Handa's 505(b)(2) NDA." and "It is not possible at this time to determine the likelihood of an unfavorable outcome". No securities litigation or enforcement action is disclosed in the 10-Q. source

   Our view: The Federal Circuit ruling protects against MSN until 2030, subject to the November trial on the '039 patent. The Handa tentative approval, which the August 5 results release did not mention, and the Accord and Almatica filings are separate threats to Cabometyx revenue that the MSN win does not resolve.

This is opinion. We compare the company's own announcements with its filings and the scientific record, and link every source. It is not an allegation of wrongdoing and not investment or legal advice. Legal items list only matters disclosed in filings or by courts and regulators. Checked Oct 23, 2026. Companies can respond through our right of reply.