Schrödinger is forcing a hosted, ratable software model onto a physics-based molecular-discovery franchise that still draws a large share of reported profit from collaborator exits rather than from license economics. Second-quarter annual contract value rose twenty-seven percent, and management held the full-year ACV band that still implies low-double-digit growth. Trailing four-quarter ACV reached $208 million, which is the cleaner read on demand than a software line depressed by recognition timing. The print therefore asks whether contract growth is finally decoupling from GAAP software sales that fell as hosted mix jumped. That gap between bookings quality and reported revenue is the investment debate.
The hosted shift is the mechanism, not a footnote. Hosted licenses already represented 47 percent of software revenue in the quarter, versus less than a third a year earlier, and each extra point of hosted mix temporarily knocks a few million off recognized software sales. Software gross margin compressed into the low seventies as cloud cost replaced high-margin on-premise recognition. Net income flipped positive only because other income included a large fair-value gain on Eli Lilly's purchase of Ajax Therapeutics, a company Schrödinger co-founded. Strip that mark and the quarter is still an operating loss. The income statement is celebrating an exit; the operating statement is still funding a platform transition.
Cash including marketable securities ended the quarter above $400 million after a mid-year cash receipt tied to the Ajax close. The next test is whether third-quarter ACV lands inside the guided band and whether Bunsen, the new agentic co-scientist, starts showing up as throughput at Bristol Myers Squibb rather than as a product announcement. Can Schrödinger convert physics accuracy and an agentic interface into durable hosted ACV without needing another collaborator sale to dress the income statement?