Keros Therapeutics is no longer the multi-program transforming-growth-factor company the market financed through late last decade. After licensing elritercept to Takeda, killing cibotercept in pulmonary arterial hypertension, and sending most of the partner cash back to stockholders, the residual equity is a cash-heavy option on one neuromuscular ligand trap. The investment debate is whether that option is being given away inside a negative enterprise value or whether the long wait to diseased-patient data justifies the discount. Cash still exceeds the entire equity capitalization, which means the market is assigning less than nothing to rinvatercept and to the leftover Takeda and Hansoh royalty claims. That is a harsh read, and it is also the entire setup.
The development that reset the company is the exclusive Takeda license that became effective in mid-January of last year. Takeda took global rights outside Greater China to elritercept and paid a $200 million upfront. The partner assumed all remaining development cost on a late-stage hematology program Keros no longer funds. That transfer is why research spending halved. The balance sheet then funded a $375 million capital return and a multi-year runway on rinvatercept alone. Shareholders exchanged a cash-burning blood program for a milestone-and-royalty claim plus a smaller, cheaper company. The subsequent repurchase of the ADAR1 and Pontifax blocks, and the tender that followed, cut the share count roughly in half and left treasury stock sitting on the balance sheet as a reminder of how much cash left the building.
The tension is that healthy-volunteer pharmacology is not a Duchenne functional endpoint. The first-in-human study showed more muscle, less fat, and higher bone density, which is exactly the ligand-trap signature, but those volunteers did not have dystrophin-deficient muscle cycling through damage and failed regeneration. The open-label patient study in Duchenne is small, and initial data does not arrive until the first half of next year. A miss there leaves a company whose partnered hematology option is real but whose internal pipeline has already lost one asset to pericardial-effusion toxicity. Competitors in Duchenne already sell exon skippers, gene therapy, and steroids; a body-composition signal that does not move function would not reopen the multiple.
The next observable that moves the residual claim is the first cut of Duchenne patient data, guided for the first half of 2027. Until then the equity lives on cash, the Takeda milestone drip, and the pledged quarter of partner proceeds received through year-end 2028. A second late-stage elritercept study, ELRiSE, dosed its first patient in July and triggered a $20 million development milestone, of which a quarter is earmarked for stockholders. That is proof the partner is still spending. It is not proof the market is wrong about rinvatercept.