One assignee leads the ranked field with 49 of the 383 records in scope, well ahead of the fifth-ranked assignee at 21 and the tenth-ranked at 11. The top five assignees combined hold 37.1% of all records, but the ranking still extends to 91 companies, so the field is concentrated at the top without being a duopoly. Anyone doing competitive tracking should watch both the leader and the handful of firms just behind it, since the drop-off after fifth place is steep.

Filing peaked in 2022 at 43 records and had fallen to 25 by 2024, a 38% decline over that three-year window using the last years with complete data. That is a real cooling trend, not noise. However, publication typically lags actual filing by around 18 months, so the lower counts shown for 2025 and 2026 in most datasets are undercounts still being filled in, not confirmation that the decline is continuing at the same rate.

H01L, the general semiconductor devices classification, appears in 69.5% of the 383 records in this corpus, making it the dominant class by a wide margin. G01J (radiation and light measurement), H04N (pictorial communication) and H10F (photovoltaic and light-sensitive devices) each sit between 13% and 17% of records. Digital processing (G06F) and optics (G02B) are both under 4%, indicating that most claim activity centres on the diode structure itself rather than downstream signal handling.

The IPC composition points to G06F and G02B as the thinnest branches relative to the core H01L device claims, meaning digital readout, timestamping and microlens/fill-factor integration are comparatively under-claimed. That does not guarantee an open path — it means fewer records carry those classes, so a targeted search is still needed before drafting. Guard ring geometry for newer stacked-chip architectures is another area worth checking, since it sits at the intersection of two well-filed classes rather than being covered outright by either.

US9935231B2, assigned to AMS AG, claims a manufacturing method for building a SPAD's guard ring structure through sequential doping steps — a first well laterally surrounding the multiplication zone at a set distance, followed by a second doping process. It is a process claim tied to breakdown voltage uniformity around the avalanche region, not a claim on SPAD devices generally. Anyone designing a guard ring using a similar staged-doping sequence at a comparable lateral offset should review this claim closely, but alternative guard ring architectures that do not rely on this specific doping sequence would sit outside its scope.