
As of November 16, 2026 there is no judgment: on April 23, 2026 the court dismissed one of seven counts without prejudice (the DMCA claim for false copyright management information) and let the three Lanham Act counts and the two California counts proceed, and the copyright count on the training copies was not challenged.
A jury trial was set for January 18, 2028, and on October 30, 2026 the court granted the parties' joint motion as modified and stayed the action until January 13, 2027 (Dkt. 81). No court has ruled on fair use in this case.
This is a separate case from the English proceedings decided on November 4, 2025, with its own court, docket and claims. DataPacts is not affiliated with any party and gives no legal advice; this page is part of the independent reference on AI training data licensing.
Dates, dockets and counts here come from the court filings named in each table, read on November 16, 2026, and how figures are sourced and dated is set out on the methodology page.
Getty v. Stability AI in the US at a Glance: Court, Docket, Parties and Status
No judgment exists in the US case: one claim was dismissed on April 23, 2026, the rest were in discovery, and on October 30, 2026 the court stayed the action until January 13, 2027 on the parties' joint motion. The table lists the case facts as of 2026-11-16.
| Field | Value | Source |
|---|---|---|
| Case | Getty Images (US), Inc. v. Stability AI, Ltd. et al. | Dkt. 1 |
| Court | U.S. District Court, N.D. Cal., San Francisco Division | Dkt. 1; Dkt. 80 caption |
| Docket | 3:25-cv-06891-TLT (discovery matters: -TLT-SK) | Dkt. 74; Dkt. 80 |
| Judge | Trina L. Thompson, J.; discovery referred to Magistrate Judge Sallie Kim | Dkt. 53; Dkt. 55; Dkt. 80 |
| Filed | 2025-08-14 | Dkt. 1 |
| Earlier action | D. Del. 1:23-cv-00135-JLH, filed 2023-02-03, voluntarily dismissed without prejudice 2025-08-14 | D. Del. Doc. 69; Dkt. 53 |
| Plaintiff | Getty Images (US), Inc. (a subsidiary of Getty Images Holdings, Inc., SEC CIK 0001898496) | Dkt. 1; SEC EDGAR |
| Defendants | Stability AI, Ltd. (Companies House 12295325, checked 2026-11-09), Stability AI, Inc., Stability AI US Services Corporation | Dkt. 1; Companies House |
| Claims | 7 counts: copyright, DMCA 1202(a), Lanham Act (3), California law (2) | Dkt. 1; Dkt. 53 |
| Rulings | 2026-04-23: Count II dismissed without prejudice; Counts III to VII survive | Dkt. 53 |
| Status | stayed until 2027-01-13 (order of 2026-10-30 granting the joint motion of 2026-10-29 as modified); joint status report due 2027-01-06; jury trial set 2028-01-18 before the stay | Dkt. 74; Dkt. 79; Dkt. 81 |
| Class action | no | Dkt. 80 |
From Delaware to San Francisco: Why Getty Refiled in 2025
Getty first sued Stability AI in the District of Delaware on February 3, 2023 (No. 1:23-cv-00135-JLH), and moved the case to California on August 14, 2025 by dismissing the Delaware action and filing a new complaint the same day.
The 2023 filing date is the one the Northern District court records in its order of April 23, 2026 (Dkt. 53), and the two actions are separate dockets in separate districts that this page keeps apart by number. The parties' statement of October 29, 2026 lists no related case (Dkt. 80).
The Delaware dismissal was a notice under Fed. R. Civ. P. 41(a)(1)(A)(i), without prejudice, and it stated that Getty would refile in the Northern District of California (D. Del. Doc. 69). Getty therefore voluntarily dismissed the action by notice; the Delaware court did not dismiss it.
The April 2026 order records that Stability AI had admitted it was subject to personal jurisdiction in the Northern District (Dkt. 53), which is the forum Getty chose for the 2025 complaint. The Delaware action is closed on that notice and carries no ruling on the merits.
What Getty Claims: Seven Counts Over Training Copies and Watermarked Outputs
Getty's complaint of August 14, 2025 pleads seven counts: one for copyright infringement, one for false copyright management information under the DMCA, three under the Lanham Act and two under California law. The table gives each count with its statute and its status as of 2026-11-10.
| Count | Claim | Statute | Rests on | Status 2026-11-10 | Source |
|---|---|---|---|---|---|
| I | Copyright infringement | 17 U.S.C. 101 et seq. | training copies, Exhibit A works | not challenged, pending | Dkt. 1; Dkt. 53 |
| II | False copyright management information | 17 U.S.C. 1202(a) | distorted watermarks | dismissed without prejudice 2026-04-23, not re-pleaded | Dkt. 53; Dkt. 80 |
| III | Trademark infringement | 15 U.S.C. 1114(1) | watermarks in outputs | survives | Dkt. 53 |
| IV | Unfair competition, false designation of origin | 15 U.S.C. 1125(a) | watermarks in outputs | survives | Dkt. 53 |
| V | Trademark dilution | 15 U.S.C. 1125(c) | watermarks in outputs | survives | Dkt. 53 |
| VI | Unfair competition | California law (UCL) | watermarks in outputs | survives | Dkt. 53 |
| VII | Trademark dilution | California law | watermarks in outputs | survives | Dkt. 53 |
The copyright count covers the training copies, and the other six counts turn on the watermark outputs. Every statement in this section and the three below describes what Getty alleges in its pleading, not a finding of any court.
The Training Allegations: Photographs, Captions and Metadata
Getty alleges that Stability AI copied more than 12 million photographs from its collection, with their captions and metadata, without permission or payment, to train Stable Diffusion (Dkt. 1, para. 1).
For the copyright count Getty selected 25,647 registered works, listed in Exhibit A to the complaint (Dkt. 1, para. 48). They are best called works listed in the complaint, because no court has found any of them infringed.
Getty further alleges that the original Stable Diffusion and later models were trained on content downloaded through datasets prepared by LAION, which is not a party, and it describes those datasets as more than five billion hyperlink-text pairs (Dkt. 1, paras. 62 to 63).
What the dataset holds, who runs it and its legal status are covered on the page about LAION-5B and what it contains; this page does not restate the dataset's size from any other court's judgment.
The Watermark Allegations Behind the Trademark Counts
Getty alleges that Stable Diffusion outputs carry distorted versions of the Getty Images watermark, which it says confuses users about the source of the images and implies an association with Getty (Dkt. 1, paras. 14, 84 to 87 and 98 to 102).
Those allegations support the false copyright management information count, the trademark infringement count, the false designation of origin count, the federal dilution count and the two California counts.
As the court summarized in Dkt. 53, Getty pleads use of its mark in the United States since 1995 and eight marks on the USPTO Principal Register from 2002.
What Getty Asks the Court to Order
Getty asks for findings of infringement and willfulness, a permanent injunction, destruction of every version of Stable Diffusion trained on Getty content without permission, an accounting of profits and damages (Dkt. 1, relief). The relief the pleading lists is:
- findings of infringement, including willfulness;
- a permanent injunction;
- destruction of all versions of Stable Diffusion trained on Getty content without permission;
- an accounting of profits;
- statutory damages under Counts III and IV of up to USD 2,000,000 per counterfeit mark, as pleaded on August 14, 2025 (Dkt. 1, paras. 134 and 142).
The copyright count carries its own statutory damages request in the pleading, which is not restated here. As of November 10, 2026 the USD 2,000,000 ceiling is the amount Getty asks for under Counts III and IV, not an amount any court has awarded or found due.
Getty v. Stability AI US Docket Entries, 2023 to 2026
Ten filings carry the US case from the Delaware dismissal to the stay order. The table lists them as of 2026-11-16 with the document each rests on.
| Date | Court and entry | What it decided or reported | Source |
|---|---|---|---|
| 2025-08-14 | D. Del. Doc. 69 | notice of voluntary dismissal without prejudice, Rule 41(a)(1)(A)(i); refiling in N.D. Cal. stated | CourtListener PDF |
| 2025-08-14 | Dkt. 1 | complaint, seven counts, Exhibit A of 25,647 works | RECAP archive PDF |
| 2025-10-14 | Dkt. 24 | Stability AI's motion to dismiss Counts II to VII (argued 2026-04-07) | as recorded in Dkt. 53 |
| 2026-04-23 | Dkt. 53 | order granting in part and denying in part: Count II dismissed without prejudice, leave to amend by 2026-05-07 | RECAP archive PDF |
| 2026-05-04 | Dkt. 55 | order of reference: all discovery to a magistrate judge | RECAP archive PDF |
| 2026-05-08 | Dkt. 58 | ESI protocol and protective order: Stability AI's proposals granted (five custodians, five search terms) | RECAP archive PDF |
| 2026-05-21 | Dkt. 62 | Stability AI's answer (amended answer 2026-07-02, Dkt. 67) | RECAP archive PDF |
| 2026-08-18 | Dkt. 74 | revised case management and scheduling order: jury trial 2028-01-18, 14 days | CourtListener PDF |
| 2026-10-29 | Dkt. 80 | joint case management statement: stipulation to stay all deadlines for 60 days (joint motion, Dkt. 79); 185 photographs withdrawn 2026-10-16 | RECAP archive PDF |
| 2026-10-30 | Dkt. 81 | order granting the joint motion as modified: action stayed until 2027-01-13; joint status report due 2027-01-06; case management conference reset to 2027-01-13 | CourtListener docket |
No later entry was in the public record on 2026-11-16.
The April 23, 2026 Order on Stability AI's Motion to Dismiss
Judge Trina L. Thompson granted Stability AI's motion only on Count II, the DMCA claim for false copyright management information, and let the three Lanham Act counts and the two California counts go forward (Dkt. 53, April 23, 2026). The table sets out the outcome by count.
| Count | Moved against? | Outcome | Court's reason in one line | Source |
|---|---|---|---|---|
| I | no (not moved) | pending | not part of the motion | Dkt. 53 |
| II | yes | dismissed without prejudice | intent to induce or conceal infringement not pleaded (double scienter of 1202(a)) | Dkt. 53 |
| III | yes | denied | use in commerce and likelihood of confusion pleaded | Dkt. 53 |
| IV | yes | denied | Dastar does not bar it: Stability AI distributes its own outputs with the marks | Dkt. 53 |
| V | yes | denied | fame pleaded (four statutory factors) | Dkt. 53 |
| VI | yes | denied | unlawful, unfair and fraudulent prongs pleaded | Dkt. 53 |
| VII | yes | denied | decided with Count V | Dkt. 53 |
Stability AI moved to dismiss Counts II to VII on October 14, 2025 (Dkt. 24); the motion was fully briefed on November 4, 2025 and argued on April 7, 2026, and the order issued sixteen days later. Initial disclosures had begun on November 24, 2025 (Dkt. 53, citing Dkt. 38).
On the DMCA count, Section 1202(a) requires both knowledge and intent. The court found that Getty pleaded knowledge that the distorted watermarks were false copyright management information but not intent to induce, enable, facilitate or conceal infringement, and it read the complaint as alleging that the watermarks are byproducts of overfitting or memorization during training.
Count II was dismissed without prejudice, with leave to amend by May 7, 2026. Because no amended complaint followed, the DMCA count is out of the case as it now stands, while the same watermark allegations remain the basis of the trademark and California counts.
On the trademark counts, the court held likelihood of confusion sufficiently pleaded at this stage and held that Dastar Corp. v.
Twentieth Century Fox Film Corp., 539 U.S. 23 (2003), does not bar the false designation of origin claim, because Stability AI distributes its own outputs bearing the marks. It held fame pleaded for dilution, and the California unfair competition claim survives on the unlawful, unfair and fraudulent prongs.
Count I, the copyright claim over the training copies, was not part of the motion, so the order rules on no copyright question, and no court has ruled on fair use in this case; the general test is set out on the page on fair use and AI training in US law.
Getty filed no amended complaint (Dkt. 80), Stability AI answered on May 21, 2026 (Dkt. 62) and it amended its answer on July 2, 2026 (Dkt. 67).
Discovery, the 185 Withdrawn Photographs and the Trial Schedule
The case was in discovery under the revised scheduling order of August 18, 2026 (Dkt. 74), which sets a 14-day jury trial for January 18, 2028, until the court stayed the action on October 30, 2026 (Dkt. 81).
Discovery was referred to a magistrate judge on May 4, 2026 (Dkt. 55), and Magistrate Judge Sallie Kim is named for discovery disputes in the parties' statement (Dkt. 80). On May 8, 2026 the court adopted Stability AI's ESI proposal of five custodians and five search terms, together with its protective-order terms (Dkt. 58).
As of October 29, 2026 the parties had exchanged substantial written discovery and no depositions had been taken (Dkt. 80).
On August 18, 2026 the court also granted in part Getty's motion to modify the schedule and extended discovery by 60 days (Dkt. 73; Dkt. 68), and the parties report open discovery disputes that they hope to resolve between themselves, with remaining disputes referred to the magistrate judge (Dkt. 80).
| Step | Date | Source |
|---|---|---|
| Fact discovery cut-off | 2026-11-20 | Dkt. 74; Dkt. 80 |
| Opening expert reports | 2027-01-08 | Dkt. 74 |
| ADR completion | 2027-01-18 | Dkt. 74 |
| Rebuttal expert reports | 2027-02-26 | Dkt. 74 |
| Expert discovery cut-off | 2027-04-16 | Dkt. 74 |
| Last day to hear dispositive motions | 2027-08-31 | Dkt. 74 |
| Joint pretrial statement | 2027-10-28 | Dkt. 74 |
| Final pretrial conference | 2027-12-02 | Dkt. 74 |
| Jury trial (14 days) | 2028-01-18 | Dkt. 74 |
The table is as of 2026-08-18; the court stayed the action until 2027-01-13 on 2026-10-30 (Dkt. 81) and will decide on that date whether the stay is lifted, so these dates may be reset.
On October 16, 2026 Getty withdrew 185 photographs, covered by three registrations, from its list of Asserted Works (Dkt. 80); the filing gives no count of the works that remain, and none is computed here. Getty estimates a 14-day trial and Stability AI estimates 5 days (Dkt. 80).
Private mediation was agreed in the joint ADR report of December 19, 2025, and the ADR deadline is January 18, 2027, which is not the fact discovery cut-off, although the two dates are easy to confuse.
The Stay After Getty's Chapter 11 Filing
On October 29, 2026 the parties filed a stipulation asking the court to stay all deadlines for 60 days, so that Getty can focus on its chapter 11 case and the parties can discuss bringing this litigation to an end (Dkt. 80).
If no dismissal is filed by the end of the stay, the parties will submit a status update proposing a schedule.
On October 30, 2026 Judge Thompson granted the joint motion (Dkt. 79) as modified: the action is stayed until January 13, 2027, the parties must file a joint status report by January 6, 2027, and the court will decide on January 13, 2027 whether the stay is lifted (Dkt. 81, CourtListener docket, checked 2026-11-16).
The case is stayed, not settled: no dismissal or settlement was in the court record on November 16, 2026. The quoted phrase about bringing the litigation to an end is not read here as a settlement.
Getty Images Holdings, Inc. and certain subsidiaries filed a prepackaged chapter 11 case on October 21, 2026 in the U.S. Bankruptcy Court for the District of Delaware, and the page on Getty Images AI licensing deals and chapter 11 covers Getty's licenses and that restructuring.
What the US Case Shows Holders of Licensable Images
The US case shows that an image library's training claim can survive to discovery without any ruling on fair use, while the watermark claims carry most of the decided law so far. All that the order of April 23, 2026 decides concerns the counts about outputs that show a watermark.
Getty's own licensing record is a separate matter: its public licenses with Perplexity (October 31, 2025) and OpenAI (June 21, 2026) are display licenses in which training use is not stated, and nothing in this case sets a price for any license.
How image holders license for training is set out on the page on selling photos and images for AI training.
The case also shows how the claims divide by evidence. The training claim rests on the allegation that a collection was copied and on a list of registered works, while the watermark claims rest on what the outputs display and on the mark's use, so the two halves of the case would be tested on different proof.
The complaint lists registered works, and the court tested the trademark counts against registrations and use of the mark since 1995. Registrations and records of use are therefore what this record relied on, which is a fact about the record and not advice.
Where the US Getty Case Sits Among Other Image-Training Cases
The US Getty case is one of three image-training cases against Stability AI or its dataset supplier in the tracker, next to the English proceedings and the German dispute over LAION.
Every tracked case with its docket and status is on the case tracker of AI copyright lawsuits over training data. The table compares the three as of 2026-11-10.
| Case | Court and docket | Status | Page |
|---|---|---|---|
| Getty Images v. Stability AI (UK) | High Court, Claim No. IL-2023-000007; [2025] EWHC 2863 (Ch) | judgment 2025-11-04: secondary infringement dismissed, trademark infringement found in part; permission to appeal granted to Getty on 2025-12-16; Court of Appeal (CA-2026-000007) part heard 2026-11-11, judgment reserved 2026-11-12 (checked 2026-11-16) | separate UK page, not yet published |
| Andersen v. Stability AI | N.D. Cal. 3:23-cv-00201-WHO | discovery as of 2026-11-09 | Andersen page, not yet published |
| Kneschke v. LAION | BGH I ZR 281/25 | pronouncement set for 2026-12-17 | Kneschke page |
The three cases differ in forum and in the question put to the court: the US case tests watermark claims first and has not reached the copyright count, while the German case turns on the text and data mining exceptions of German copyright law.
Stability AI's licensing deals and practices belong on its buyer profile, Stability AI's licensing and acquisition, which is not yet published.
More About Getty Images v. Stability AI in the US
This section lists what is decided and what is open, and answers the questions the case prompts. It is information, not legal advice.
Decided Versus Still Open in the US Case
The dismissal of the DMCA count and the survival of the other counts are decided; the copyright claim, any fair use defense, whether the stay is lifted and the trial are open.
| Decided | Still open |
|---|---|
| Delaware action ended by notice, without prejudice | Count I (copyright) on the merits |
| Count II dismissed without prejudice and not re-pleaded | any fair use defense |
| Counts III to VII survive the pleadings | whether the stay ordered 2026-10-30 is lifted on 2027-01-13 |
| trial date set (before the stay) | jury trial on 2028-01-18 |
Frequently Asked Questions About Getty Images v. Stability AI
Are the US and UK Getty cases the same lawsuit?
No, they are two lawsuits in two courts with two dockets. The UK High Court decided its case on November 4, 2025 after Getty abandoned the UK training claim, while the training claim is live in the US.
Did Getty win against Stability AI?
Not in the US, where there is no judgment. On April 23, 2026 the court dismissed one of the six challenged counts and let five proceed, with the copyright count untouched; in the UK, trademark infringement was found in part and secondary infringement was dismissed.
What is the citation for Getty Images v. Stability AI?
The US case is Getty Images (US), Inc. v. Stability AI, Ltd. et al., N.D. Cal. No. 3:25-cv-06891-TLT (earlier D. Del. No. 1:23-cv-00135-JLH). The UK judgment is a different case, cited as [2025] EWHC 2863 (Ch).
Has anyone appealed in the US case?
No, there is no final judgment to appeal as of November 16, 2026. The appeal in the news is the UK one.
Is the Getty Images lawsuit a class action?
No, Getty sues alone. The joint statement says "This is not a class action" (Dkt. 80).
Why did Getty drop the Delaware case?
Getty dismissed it without prejudice to refile in the Northern District of California, where Stability AI admitted personal jurisdiction. The notice of August 14, 2025 stated the refiling, and the new complaint was filed the same day.