AstraZeneca takes $2B Summit stake, pairs its Claudin 18.2 ADC with ivonescimab

On 29 September 2026, AstraZeneca agreed to put $2 billion into newly issued Summit Therapeutics equity, taking convertible preferred shares whose conversion implies an $18.36 common-stock price, a premium to Summit's close that day ^[1][2]. The cash comes with a clinical collaboration: AstraZeneca's Claudin 18.2 antibody-drug conjugate sonesitatug vedotin (Sone-Ve, an MMAE payload) will be tested with Summit's PD-1/VEGF bispecific ivonescimab in gastrointestinal cancers, trials starting imminently, with each company supplying its molecule, splitting costs, AstraZeneca sponsoring the studies, and each keeping rights to its own compound ^[3]. A separate non-binding memorandum contemplates pairing ivonescimab with more AstraZeneca medicines, including other ADCs, but attaches no milestone, royalty or profit-share terms and no guarantee it proceeds ^[1].

The structure is a stake rather than a buyout. Endpoints News frames it as a long-speculated deal that landed as a slice of the biotech instead of a purchase of the whole company ^[10]. What AstraZeneca is buying into is a year of overall-survival data on ivonescimab. In the updated HARMONi-2 analysis, ivonescimab monotherapy beat pembrolizumab monotherapy on OS in PD-L1-positive lung cancer, hazard ratio 0.73 (95% CI 0.57-0.95, p=0.009), median OS 30.8 versus 22.6 months, with the PD-L1-high subgroup at 0.58 ^[6]. In EGFR-mutated NSCLC after TKI failure, the global HARMONi trial's updated June 2026 cut showed an OS hazard ratio of 0.76 (95% CI 0.61-0.95, nominal p=0.0151), after the April 2025 primary analysis missed at 0.79 (p=0.057) ^[7]. In HARMONi-GI1, an ivonescimab-based regimen showed statistically significant OS superiority over a durvalumab-based regimen ^[5]. Ivonescimab was engineered by Akeso, which keeps China and other regions while Summit holds the major ex-China territories ^[3].

Ivonescimab's OS win over pembrolizumab in HARMONi-2 came against the anti-PD-1 backbone of most first-line combinations ^[6]. AstraZeneca, whose own durvalumab was the comparator that lost in HARMONi-GI1, has now tied $2 billion of its balance sheet to the molecule beating it, and pointed the combination work at gastrointestinal tumors where its Sone-Ve ADC already reads out ^[3][5]. AstraZeneca said Sone-Ve produced a statistically significant, clinically meaningful OS improvement versus investigator's choice in second-line-and-later Claudin 18.2-positive advanced gastric cancer in CLARITY-Gastric01, with data due at ESMO 2026 in a Presidential Symposium ^[3]. Pairing a checkpoint-plus-VEGF bispecific with a validated Claudin 18.2 ADC is the thesis the money is funding.

AstraZeneca is not the only large partner circling the molecule: GlaxoSmithKline's EMBOLD PanTumour-103 (NCT07602855) began pairing its own ADC risvutatug rezetecan with ivonescimab in solid tumors on 9 September 2026 ^[9].

What to watch, with dates. HARMONi-7 reads out toward 2028 ^[8].

The now-what. For anyone building a PD-1/VEGF or PD-L1/VEGF bispecific, HARMONi-2's OS hazard ratio is the number your trial design now has to answer to. For holders of Claudin 18.2 or GI-directed ADCs, ivonescimab now has a well-capitalized partner in AstraZeneca, with the GI combination trials jointly funded by both companies and sponsored by AstraZeneca, so the registration of those trials is worth tracking. For programs leaning on a pembrolizumab backbone, a PD-1/VEGF entrant with a mid-November 2026 US catalyst is now a live competitive variable.

Prognyx could not confirm from the excerpted filings what percentage of Summit or what board rights the $2 billion buys, nor the deal's closing timeline or antitrust conditions; CLARITY-Gastric01 and the planned ivonescimab-Sone-Ve GI trials had no ClinicalTrials.gov record in the set Prognyx reviewed; and no filing frames the deal as aimed at Merck beyond the head-to-head trial data.

Verdict: a well-capitalized challenge to first-line checkpoint combinations, with the nearest hard catalyst the 14 November 2026 US decision; competitors with pembrolizumab- or durvalumab-anchored programs have weeks, not quarters, to reset their read.