908 Devices is no longer a two-headed instruments company. After selling the desktop bioprocess line to Repligen, the remaining franchise is a handheld chemical lab sold to police, hazmat crews, and defense buyers who need an answer at the scene rather than a sample sent downtown. The second-quarter print is the cleanest evidence yet that the municipal channel, not a single Pentagon program, is carrying that franchise. State and local agencies accounted for more than half of revenue, VipIR crossed its hundredth lifetime shipment, and management lifted only the floor of the full-year outlook. The equity debate is whether that municipal engine, plus a just-closed Swiss screening bolt-on, can push the adjusted loss toward breakeven before earnouts and mix do the opposite.
Hardware is doing the work that recurring revenue is not. Product revenue rose to $13 million as VipIR and NIRLab units shipped. Service and contract revenue declined as mass-spec support faded with the older installed mix. Recurring revenue was $4.9 million and still near a third of sales. The dollar gain was modest and leaned on software, accessories, and FTIR service rather than a broad installed-base monetization wave. Adjusted EBITDA loss of $1.9 million was less than half the year-ago gap, as facility costs fell after the Boston consolidation and cash operating spend stayed roughly flat against faster revenue. Reported results still look weak because a noncash contingent-consideration charge from earlier deal math remains larger than the entire adjusted loss.
Cash and marketable securities ended the quarter just above $100 million with no drawn debt, even after the NIRLab cash outlay. That buffer is why the story can be about execution rather than survival. What the next several quarters have to show is whether NIRLab subscriptions and VipIR placements keep state-and-local growth intact while service revenue stops shrinking, and whether the excluded Aerosol and Vapor Chemical Agent Detector military program stays a free option or becomes a hole in the second-half setup. The open question is simple. Does municipal modernization plus a cheaper screening layer produce a self-funding handheld platform, or does the print stay a hardware spike dressed as a software story?