$0.30
$12.00
bear $0.30
base $1.10
bull $12.00
margin of safety $6.00
market price* $24.17
Business condition attached to the margin of safety
Positive top-line results from the Phase 3 EMPEROR study of zorevunersen on its primary endpoint (major motor seizure frequency), expected by mid-2027, with the FDA path to a rolling NDA filing confirmed.
Stoke Therapeutics is a clinical-stage biotechnology company. It has no approved product. Its lead medicine, zorevunersen, is designed to raise production of the SCN1A protein in patients with Dravet syndrome, a severe childhood epilepsy caused by a shortage of that protein. The Phase 3 study, EMPEROR, began in May 2025 (first patient dosed August 2025), and the company expects data by mid-2027 and plans a rolling FDA application in the first half of 2027. Biogen shares development and will sell the drug outside the United States, Canada and Mexico. Stoke also has a research collaboration with Acadia on SYNGAP1. What the numbers show. Owner earnings, meaning operating cash flow less working-capital swings, stock compensation and capital spending, have been negative every year we could measure: about $(2.30) per share in 2021, $(2.61) in 2022, $(2.39) in 2023, $(1.54) in 2024 and $(2.49) in 2025. Reported 2025 results look better than reality: revenue of $184.4 million and operating cash flow of $45.6 million were driven by the one-time $165.0 million upfront payment from Biogen received in February 2025, which we removed as a timing item. The 2024 to 2025 figures also mix in deferred-revenue effects that the filings did not tag separately, so those two years are a judgement rather than an exact reconciliation. Stock compensation ($32.2 million in 2025) is treated as a real cost. In the first half of 2026 operating cash flow was $(117.2) million, so the burn is rising as the Phase 3 study and launch preparation expand. The company reported $390.9 million of cash and marketable securities at December 31, 2025 and says this funds operations into 2028. By mid-2026 we estimate roughly $270 to $290 million remains, about $4 to $4.50 per share on roughly 66 million diluted shares. That is our estimate from the year-end balance and first-half burn, not a reported figure. The cash is not the value we credit, because the plan is to spend it. Our approach. We only credit what is already secured or a proven, repeating process. A drug that has not finished its pivotal trial or been approved is not in that category, so zorevunersen's eventual sales are worth $0 in the base case, even though the Phase 1/2a results and Breakthrough Therapy Designation are encouraging. Biogen's potential milestones (up to about $50 million development and $335 million commercial) depend on the same events and are also not credited. Bear case ($0.30 per share). The Phase 3 study misses or is delayed, and the company winds down or is left with little after spending its cash through the mid-2027 readout and restructuring costs. We assume only a small residual of cash per share is recoverable. Roughly 3 in 10 companies in this position end up here, but we cannot measure that from the filings; the case simply reflects that nothing in the filings guarantees value beyond the cash. Base case ($1.10 per share). Cash is spent as planned, with Biogen bearing 70 percent of the shared global development cost, through the mid-2027 readout, leaving roughly $70 to $80 million, about $1.10 per share. No product earnings are credited because none are secured. Bull case ($12 per share). EMPEROR succeeds on its primary endpoint in mid-2027, the FDA accepts a rolling application, and Biogen's $50 million of development milestones are earned. We then assume a launch in the United States, Canada and Mexico producing about $80 million a year of owner earnings by roughly 2030 and fading over the following decades, after dilution to fund the launch (about 70 million shares). The launch earnings figure is an illustrative assumption, not a filed number, because the filings do not give us a verified patient count or price; this is why the bull case is a wide estimate. What would make this wrong. Positive data would make the stock worth several times our base value; a clean failure would leave little. This is effectively a binary outcome, and the ratings reflect that the base case values the business on evidenced cash only. Because the payoff depends on one trial, the margin-of-safety price carries a very large margin of safety, and a higher margin-of-safety price applies only after the trial reads out successfully. This is our first analysis of Stoke.