VERAXA taps AI to better pick cancer targets: the small, real step versus the big headline

By NewMediaWire · July 13, 2026.

VERAXA Biotech (Nasdaq: VRXA) has announced a collaboration with Ardigen S.A., a bioinformatics company specializing in artificial intelligence, to support the development of its BiTAC(R) platform, a family of bispecific antibodies that activate T cells against tumors (T-cell engagers), with an eye also to future antibody-drug conjugates (ADCs). In this first phase, the goal is concrete: to build AI tools capable of identifying and prioritizing combinations of two targets that act synergistically, improving accuracy against the tumor and reducing off-target toxicity, one of the major problems with this type of therapy. According to the statement, the AI will analyze large sets of preclinical and clinical data, including programs that demonstrated efficacy but failed due to toxicity, in order to learn from those failures and avoid repeating them in the design of new candidates.

This is, at bottom, one of the uses of AI in biotechnology that does have solid empirical backing, as opposed to the more diffuse promise of 'AI designs drugs from scratch.' The selection of targets and, above all, of target combinations is a combinatorial problem: with two biological objectives there are already thousands of possible pairs to choose which one to test, and a good part of clinical failures in oncology come not from a lack of efficacy but because the drug works yet damages healthy tissue. Retrospectively mining those failures with AI models so as not to stumble twice over the same stone is a bounded, verifiable application with low regulatory risk, very different from claiming that an algorithm is going to invent the cure for cancer.

That said, it is worth reading the statement for what it is: a note distributed by NewMediaWire through IBN (InvestorBrandNetwork), an investor relations service common among small-cap companies, with the 'forward-looking statements' legal disclaimer that usually accompanies this type of text. No economic terms of the agreement are detailed, nor timelines, nor concrete preclinical results arising from the collaboration; for now it is an announced strategic alliance, not a peer-validated scientific milestone. In the universe of small publicly traded biotechs, announcing a collaboration with 'AI' is also, inevitably, a message to the market as much as to science, and the two things are worth distinguishing when reading the news.

Our reading is that this type of agreement is the base material, unglamorous but cumulative, of the underlying thesis we defend about AI and health: not an abrupt turn toward the eradication of diseases, but thousands of small optimizations —better target selection, fewer drugs that fail late and expensively due to toxicity, somewhat cheaper discovery cycles— that, added up over years, can indeed bring precision oncology closer to higher clinical success rates. In the short term, however, one must be cautious about the noise: the microcap biotech industry frequently produces statements of this kind that capitalize on the 'AI' label to improve stock market perception without there yet being clinical evidence to back it. The difference between which of these players truly manages to reduce the toxicity of their programs —and which merely announces it— will be seen, as always in this sector, in the clinical trial data, not in the press release.