Kivu Bioscience

Here is a thing about cancer drugs that is both obvious and, in practice, extremely hard: you want the drug to kill the tumor and not kill the patient. The tumor and the patient are made of the same basic stuff, which is the entire problem. Chemotherapy solves this by poisoning everything and hoping the tumor dies first. It's a strategy with a poor bedside manner.

Antibody-drug conjugates, or ADCs, are the more elegant idea. You take an antibody - a protein that is very good at finding one specific kind of cell - and you bolt a cytotoxic payload onto it. The antibody is the guided part. The payload is the missile. In theory the whole thing floats through the bloodstream, ignores healthy tissue, docks onto the tumor, and delivers its poison at the address on the envelope. In theory.

In practice, ADCs have a long history of falling apart mid-flight. The bond between antibody and payload can be sloppy. Attach the drug at random spots and you get a batch of molecules that are all slightly different - some carrying too much payload, some too little, some destabilized in ways that make them shed their cargo into the bloodstream, which is exactly the healthy-tissue poisoning you were trying to avoid. The guided missile becomes a leaky one.

Kivu Bioscience, founded in 2024 and headquartered in San Mateo, California, has decided that this - the unglamorous chemistry of how you attach the drug - is where the game is actually won. Its tagline calls its ADCs "kinder and gentler," which is a strange phrase to apply to a class of cytotoxic cancer therapy, and also exactly the point.

The company builds on a licensed platform from Synaffix called GlycoConnect. The trick is site-specific conjugation: instead of attaching the payload wherever it happens to stick, Kivu attaches it at a single, defined location on the antibody - a spot called asparagine-297. Every molecule in the batch ends up nearly identical. Homogeneous. The drug-to-antibody ratio is fixed rather than an average of a messy distribution.

Kivu frames its approach as three pillars, and while "three pillars" is the kind of phrase that usually means a company has run its strategy through a marketing team, in this case the pillars are just the actual engineering problems: homogeneity (every molecule the same), stability (a hydrophilic linker that resists aggregation and holds the payload in place through the bloodstream), and potency (a clinically validated payload that gets into cells and resists being pumped back out).

That phrase - therapeutic window - is the one that matters. It's the gap between the dose that helps and the dose that harms. A wider window means doctors can push the drug harder without pushing the patient into toxicity. Most of Kivu's technical claims, translated out of biotech-speak, are really claims about widening that gap. Potency you can survive is worth more than potency you can't.

The people making these claims are not first-timers. Kivu is led by Mohit Trikha, a Ph.D. who has spent decades in ADC development, in the president/CEO/COO seat, backed by a leadership bench thick with M.D.s and Ph.D.s who have shipped this kind of molecule before. That matters in biotech more than almost anywhere, because the failure modes are subtle and mostly learned by having watched them happen.

And there's a build-versus-buy decision baked into the whole company that's worth admiring. Rather than invent its own conjugation chemistry from scratch - the temptation of every science startup - Kivu licensed a platform that had already been de-risked in the clinic. It's the kind of unsexy, disciplined choice that lets a 25-person team skip years of foundational work and go straight to designing drugs. Sometimes the smartest R&D decision is knowing which wheel not to reinvent.

The money agreed. In October 2024, Kivu closed a $92 million Series A led by Novo Holdings, with a syndicate spanning Gimv, Red Tree Venture Capital, HealthCap, and the existing backers BioGeneration Ventures, M Ventures, and BOM. Ninety-two million dollars is a serious Series A, the kind you raise when the preclinical data is clean and the investors can see a path to the clinic. "That data in the hands of this outstanding management team tells an incredibly compelling story," said Novo's Daniel O'Connell, which is investor-speak for: the science checks out and we trust these people to run it.