Galux Wins KDDF Support in Korea, Advancing AI-Designed Bispecific Cancer Immunotherapy to the Next Stage

The next wave of competition in cancer immunotherapy is increasingly not just about finding new targets, but also about who can design the molecules themselves with greater precision. South Korean AI protein design company Galux has secured support from the Korea Drug Development Fund (KDDF) to advance an AI-designed bispecific cancer immunotherapy, showing that public R&D funding is placing some of its bets on more complex antibody engineering.

According to Korea Biomedical Review, the support is focused on Galux’s AI-designed bispecific immuno-oncology treatment program. However, publicly available information remains quite limited. The report did not disclose the amount of KDDF funding, the therapeutic target, the name of the candidate molecule, or what preclinical or clinical stage the program has entered.

The core concept of a bispecific antibody is to enable the same protein molecule to recognize two targets at once. When used in cancer immunotherapy, one end may bind an antigen on a tumor cell while the other recruits T cells or other immune signals; it may also simultaneously regulate two pathways related to the tumor microenvironment. The appeal of this design is more concentrated function, but the risks are also more concentrated: affinity, selectivity, half-life, immunogenicity, and the intensity of cellular activation could each alter efficacy and toxicity if thrown out of balance.

Galux’s main technical direction is to combine artificial intelligence with physical models for de novo design of proteins and antibodies. For biomedical AI, this is more ambitious than simply screening existing molecules, because the system is not only ranking candidates, but attempting to propose new protein structures that can be produced by cells, fold stably, and bind to targets. To convince the drug development system, such platforms usually need more than computational prediction scores; they also need binding data from wet-lab experiments, functional cell assays, animal model results, and evidence of scalable production and purification.

The significance of KDDF’s support lies in pushing a platform capability closer to a drug development decision point. In recent years, South Korea has actively cultivated its domestic novel drug and biologics industries, with government-type funds often serving as early risk-sharing participants. For Galux, this kind of support may help fill resource needs for preclinical research, candidate optimization, and subsequent translation.

But this is still not a promise of efficacy. Bispecific cancer drugs have seen successful cases globally, as well as precedents of setbacks due to cytokine release, uneven target expression, tumor escape, or manufacturing complexity. Whether AI can shorten the design cycle ultimately returns to the same set of rigorous questions: whether the molecule works as expected in the human body, whether toxicity is controllable, whether manufacturing batches are stable, and whether regulators accept the logic of its design and validation.

As a result, this news is more of an R&D milestone than a clinical breakthrough. It shows that AI protein design is entering the practical field of cancer immunotherapy; what will truly be worth understanding next is not the phrase “AI-designed” itself, but whether Galux can disclose sufficiently clear targets, experimental data, and a development path, allowing this candidate therapy to move from an algorithm-generated possibility toward evidence that can be tested by medicine.