Atom's scientific instrument sales model
Atom Computing
This sales motion makes Atom a capital equipment company first, and a software company second. A buyer is not swiping a credit card for seats or API calls. It is choosing a 600 square foot machine, preparing a facility, running technical reviews, negotiating milestones, and then paying as hardware is built, shipped, installed, and accepted. That creates lumpy revenue, heavy customer concentration, and long periods where execution on one flagship deployment matters more than broad user adoption.
-
The clearest comparable is IQM, where 98% of 2025 revenue came from on premises system sales, cloud revenue was only €284K, and contracts bundled installation, commissioning, training, and maintenance. That is classic scientific instrument economics, few buyers, large contracts, and milestone based recognition.
-
Atom is following the same procurement pattern through QuNorth. The first named AC1000 deployment is Magne in Copenhagen, a system co designed with Microsoft and targeted for early 2027 commissioning. That is less like selling developer tools and more like placing a national lab asset that anchors a regional research ecosystem.
-
The installed base is where software like economics can start to appear. After the initial machine sale, Atom can layer in maintenance, calibration, operator training, hardware upgrades, remote diagnostics, and Microsoft linked software and error correction updates, but those are still attached to a hardware estate, not a standalone SaaS product.
The next phase is a shift from one off flagship procurements toward repeatable upgrade cycles. If Atom can turn QuNorth into a reference site and then replicate that model across sovereign quantum programs, recurring revenue will increasingly come from keeping systems current and connected, much like supercomputers and lab instruments that generate years of follow on service spend after the box lands.