Zealand Pharma deepens metabolic bet with AI-designed protein drugs
Zealand Pharma is widening its metabolic portfolio with a collaboration that could carry more than $1 billion in future value, this time moving beyond peptides and small molecules into multispecific protein therapeutics. The Danish company is paying Generate Biomedicines $30 million upfront to co-develop candidates for metabolic diseases, a move that broadens the scientific tools available to a company whose core identity has long been built around peptide engineering.
The agreement is structurally important because it signals where Zealand wants to compete next. Metabolic disease drug development is increasingly crowded, and the biology rarely bends to a single target. By pursuing multispecific proteins, Zealand is effectively betting on polypharmacology — the idea that a molecule can engage more than one biologic pathway at once and potentially produce stronger, more durable control of complex disease processes such as obesity and related metabolic disorders.
Generate will design and advance the partnered programs through lead-candidate selection, after which Zealand will take over clinical development and later-stage execution. For the initial two programs, Generate can collect the upfront payment and as much as $35 million in early preclinical milestones. The broader economics are much larger: nearly $1.2 billion in total milestones tied to development, regulatory and commercial achievements, plus tiered royalties on worldwide sales if any product reaches market.
The deal arrives at a sensitive moment for Zealand’s obesity franchise. Its pipeline still rests on two key partnered assets: a Roche-linked amylin analog and a Boehringer Ingelheim-partnered glucagon/GLP-1 receptor dual agonist. Both are in phase 3, yet neither has generated the level of enthusiasm investors were hoping for. Recent phase 3 data for the Boehringer program weighed on the company’s stock, and more recent phase 2 weight-loss and blood sugar results from the Roche asset also fell short of expectations.
One sentence captures the strategic backdrop: Zealand needed another engine.
By adding a protein modality, the company is not abandoning peptides; it is extending the same metabolic expertise into a different molecular format. Peptide drugs and protein therapeutics can overlap in therapeutic logic, but the latter can offer more elaborate target engagement and may be engineered to address multiple signaling axes at once. In obesity and cardiometabolic disease, where appetite regulation, energy balance, insulin sensitivity, and hormone signaling intersect, that broader pharmacologic reach can be attractive even when the practical path to clinic is longer.
The collaboration also reflects a geographic and scientific pairing. Zealand’s leadership described the arrangement as one that combines metabolic expertise in Copenhagen with multmodality expertise in Cambridge. In industry terms, that means Zealand is trying to pair disease know-how with platform depth, rather than building a protein-engineering engine entirely from scratch. That choice can shorten the distance between discovery and a credible development candidate, though it also makes the company dependent on Generate’s ability to translate AI-designed constructs into molecules with acceptable developability, manufacturability, and tolerability.
Generate’s platform is already attracting repeated outside validation. Amgen and Novartis have both partnered with the biotech in prior years to develop protein therapeutics across multiple disease areas, a pattern that suggests big pharma sees value in Generate’s design capabilities beyond any single indication. Internally, Generate’s pipeline is broader than metabolic disease and includes immunology, inflammation, and oncology programs. Its most advanced asset is GB-0895, an anti-TSLP monoclonal antibody in phase 3 for severe asthma.
For Zealand, the clinical implications are tied to modality diversification as much as to any single compound. Obesity drug development has taught the field that metabolic control is rarely linear. Even strong candidates can disappoint when weight-loss gains fail to translate into cleaner glycemic control, or when the cardiovascular and durability profile remains uncertain. Multispecific proteins may help by hitting complementary pathways simultaneously, but they also raise the bar for preclinical selection because complexity can introduce new safety liabilities or pharmacokinetic compromises.
The deal value also places pressure on execution. Zealand is committing only $30 million upfront, but the upside economics are substantial, and the milestone structure suggests confidence in the platform even before any human data exist. That is a familiar biotech bargain: a relatively contained near-term cost in exchange for access to a technology that could reshape the company’s longer-term pipeline if the preclinical chemistry, biology, and translational work all align.
The timing matters because Zealand’s current programs have already shown how unforgiving the obesity market can be. Strong mechanistic rationale alone does not protect a company from investor disappointment once phase 2 or phase 3 readouts arrive. The addition of multispecific protein candidates offers a second shot at relevance in metabolic disease, but the scientific risk remains substantial until Generate identifies molecules that are both potent and practical for clinical development.
What makes this partnership stand out is not only the dollar figure. It is the change in therapeutic philosophy: a company rooted in peptide biology is now adding AI-engineered proteins to its metabolic toolbox, hoping that a more flexible molecular architecture will produce better answers in one of medicine’s most competitive fields.
The immediate question is whether that strategy can produce a candidate with cleaner biology than Zealand’s current obesity assets, without repeating the same late-stage frustrations in a new molecular format.