The main risk is proof transfer: Schrödinger must show that its software can become a true discovery control layer, not just a useful modeling tool, while keeping internal therapeutics from becoming a capital sink. If hosted deployments hide weak demand, large customers internalize more workflow, or
SGR-3515 and partnering fail to validate the platform-derived asset engine, the stock can remain a low-multiple niche science-software name.