No single assignee dominates this dataset; the field is fragmented across device makers, a university patent foundation and several newer entrants. University of Virginia Patent Foundation shows up repeatedly through co-assignee pairs tied to named inventors, reflecting foundational closed-loop research. Established device makers such as Abbott Diabetes Care Inc. and Medtronic MiniMed, Inc. hold historical portfolios but show no new filings in the latest tracked year, so today's ranking reflects accumulated filings rather than current filing speed.

It fell sharply from a 2017 peak of 16 filings down to just 1 in 2022, then began climbing again, reaching 4 in the most recent partial year. Because patent publication typically lags actual filing by around 18 months, that 2026 figure understates true filing activity for the period. The overall shape looks like a first innovation wave being absorbed, followed by the early stage of a second one, rather than a market in decline.

The core classifications in this landscape are A61M5/172 for insulin infusion pump hardware, H04W4/80 for device-to-device wireless communication such as Bluetooth pairing, and G16H40 for healthcare informatics and device management software. In practice, most records classify primarily under A61M (111 of 129) and secondarily under G16H (88), meaning connectivity is usually claimed as a feature attached to the physical pump rather than as a freestanding software invention. G06F and G05B, the general digital-processing and control classifications, appear far less often.

US20140276554A1, assigned to Animas Corporation, claims a method for switching between a base insulin delivery profile and a temporary delivery profile based on continuous glucose monitor readings, including both an exact-amount ('hard') and a nominal ('soft') temporary profile mode. Anyone building a closed-loop or hybrid closed-loop system that dynamically overrides a baseline dosing schedule during a defined time window should check their dosing-logic claims against this family specifically. It is one of the more heavily cited records in this corpus, meaning examiners frequently cite it against adjacent filings, so a workaround typically needs to avoid the base/temporary profile switching structure itself, not just the connectivity layer around it.

The clearest gaps sit away from the crowded A61M device core: smartphone-mediated dose-adjustment interfaces, multi-sensor redundancy across Bluetooth mesh networks, edge-processed glucose trend alerting, and cloud-side version control for insulin delivery profiles all show thinner claim density relative to the core pump-plus-connectivity combination. These branches sit in classifications like G06F and G05B that appear far less often in this dataset than A61M and G16H. That does not guarantee they are unclaimed elsewhere, but within this corpus they carry less filing pressure than the core device claims.