The Healthcare AI Institute / Transparency Registry
Data report · updated weekly · as of 2026-09-27

The State of FDA AI Clearances

Every count on this page is computed live from the Transparency Registry, which mirrors openFDA public records for the FDA's 9 AI-specific 510(k) product codes. The numbers refresh with the weekly sync. Facts only: this report counts public records and never evaluates any product.

568clearances since 2018
266distinct applicants
102cleared in 2026 so far
43busiest quarter (2026 Q2)

Clearances by year

7
2018
13
2019
39
2020
43
2021
62
2022
77
2023
98
2024
127
2025
102
2026

Striped bar = year in progress. Decision dates from public 510(k) records.

The curve is still steep.

2026 has already recorded 102 clearances against 127 in all of 2025. Annual clearances under these codes have risen every full year on record.

Clearances by quarter

15
2024 Q1
21
2024 Q2
29
2024 Q3
33
2024 Q4
33
2025 Q1
34
2025 Q2
33
2025 Q3
27
2025 Q4
28
2026 Q1
43
2026 Q2
31
2026 Q3
New applicants keep arriving.

40 organisations received their first clearance under these codes in 2026 so far, out of 266 distinct applicants since 2018. The field is widening, not consolidating into a fixed roster.

By product code

CodeFDA device nameClearances
QIHAutomated Radiological Image Processing Software329
QASRadiological Computer-Assisted Triage And Notification Software87
QKBRadiological Image Processing Software For Radiation Therapy49
QFMRadiological Computer-Assisted Prioritization Software For Lesions41
QDQRadiological Computer Assisted Detection/Diagnosis Software For Lesions Suspicious For Cancer39
QBSRadiological Computer Assisted Detection/Diagnosis Software For Fracture11
QJUImage Acquisition And/Or Optimization Guided By Artificial Intelligence9
QMSAdjunctive Open Loop Fluid Therapy Recommender2
QVDRadiological Machine Learning Based Quantitative Imaging Software With Change Control Plan1

Clearances held, by applicant

A count of public records, most-cleared first. Holding more clearances is a fact about volume, not a statement about any product.

ApplicantFirst clearanceClearances
Aidoc Medical , Ltd.2018-08-0134
Siemens Medical Solutions USA, Inc.2020-07-3017
Ischemaview, Inc.2020-07-0913
Ever Fortune.Ai, Co., Ltd.2022-04-0412
Dia Imaging Analysis, Ltd.2020-06-2310
DeepHealth, Inc.2021-04-169
Viz. Ai, Inc.2020-03-189
Screenpoint Medical B.V.2018-11-218
Siemens Healthcare GmbH2022-04-158
Mim Software, Inc.2020-07-027

Method

The registry mirrors openFDA 510(k) records for the AI-specific product codes the FDA has created since 2018, refreshed weekly by an automated sync. Decision dates, applicants, and product codes are reproduced as published; nothing is edited. A year or quarter in progress is marked as partial. The full row-level dataset is downloadable: CSV · JSON. Counts are CC BY 4.0: reuse them with attribution.

How to cite

The Healthcare AI Institute, The State of FDA AI Clearances, healthcareaiops.com/registry/report/ (data as of 2026-09-27).

A dated snapshot of the row-level dataset is archived with a DOI: 10.5281/zenodo.22651260 — cite that version for reproducibility.

Evaluating a purchase in one of these categories? Start with the free vendor due-diligence checklist, verify the clearance in the registry, and keep the governance record in HAI-OS.
The diligence kit — free, no email needed.

The vendor due-diligence checklist (interactive, stores nothing) · the contract rider (paste-ready, CC BY) · the live clearance report (weekly, citable) · the record verifier.

One email each Monday: what changed in FDA AI clearances, in two minutes. Unsubscribe any time.

Going deeper? The Compliance & Regulatory AI Toolkit ($97) is the complete evaluation workbook, and HAI-OS keeps the governance record once a tool is deployed.