DCT
7:25-cv-00483
Arlington Tech LLC v. NVIDIA Corp
Key Events
Complaint
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: Arlington Technologies LLC (Texas)
- Defendant: NVIDIA Corporation (Delaware)
- Plaintiff’s Counsel: Nelson Bumgardner Conroy PC
- Case Identification: 7:25-cv-483, W.D. Tex., 10/22/2025
- Venue Allegations: Plaintiff alleges venue is proper in the Western District of Texas because Defendant maintains a "regular and established" place of business in the district, specifically a corporate office in Austin, and has committed the alleged acts of patent infringement in the district.
- Core Dispute: Plaintiff alleges that Defendant’s artificial intelligence platforms and software development kits for audio and video enhancement infringe five patents related to telecommunications, user interface design, and content management.
- Technical Context: The dispute centers on AI-powered software used to enhance real-time digital communications, a market segment focused on improving user experience in video conferencing, streaming, and interactive applications.
- Key Procedural History: The complaint alleges that Plaintiff sent correspondence to Defendant on October 10, 2025, notifying Defendant of the Asserted Patents, an event that may be central to allegations of pre-suit knowledge and willful infringement.
Case Timeline
| Date | Event |
|---|---|
| 2005-02-28 | U.S. Patent No. 7,702,083 Priority Date |
| 2010-04-20 | U.S. Patent No. 7,702,083 Issues |
| 2012-03-23 | U.S. Patent No. 8,862,456 Priority Date |
| 2013-10-03 | U.S. Patent No. 9,088,694 Priority Date |
| 2014-10-14 | U.S. Patent No. 8,862,456 Issues |
| 2015-07-21 | U.S. Patent No. 9,088,694 Issues |
| 2017-05-23 | U.S. Patent No. 10,574,201 Priority Date |
| 2019-05-03 | U.S. Patent No. 11,394,923 Priority Date |
| 2020-02-25 | U.S. Patent No. 10,574,201 Issues |
| 2022-07-19 | U.S. Patent No. 11,394,923 Issues |
| 2025-10-10 | Plaintiff allegedly sends notice letter to Defendant |
| 2025-10-22 | Complaint Filed |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 7,702,083 - "Method and apparatus for providing default media content to a calling party"
- Patent Identification: U.S. Patent No. 7,702,083, "Method and apparatus for providing default media content to a calling party," issued April 20, 2010.
The Invention Explained
- Problem Addressed: The patent addresses the undesirable user experience of a "dark screen" that occurs when a multimedia call (e.g., video call) is transferred to a destination that does not support the video media type (Compl. ¶20; ’083 Patent, col. 1:5-11).
- The Patented Solution: The invention proposes a system that detects when a destination endpoint is not providing a specific media type (like video) and, in response, automatically initiates the transmission of "default media" of that same type to the originating caller, thereby filling the void and preventing a negative user experience (’083 Patent, Abstract; col. 2:21-29).
- Technical Importance: This technology sought to improve the continuity and perceived quality of service in early multimedia communication systems by managing media streams during transitional periods like call transfers.
Key Claims at a Glance
- The complaint asserts infringement of at least Claim 1 (Compl. ¶36).
- Independent Claim 1 requires:
- determining that a destination endpoint is not providing a media stream of a first media type to an originating endpoint; and
- initiating transmission of default media of said first media type to said originating endpoint.
U.S. Patent No. 8,862,456 - "System and method for automatic language translation for applications"
- Patent Identification: U.S. Patent No. 8,862,456, "System and method for automatic language translation for applications," issued October 14, 2014.
The Invention Explained
- Problem Addressed: The patent identifies the inefficiency and high cost associated with traditional software internationalization, which typically requires modifying an application's source code to support different languages (’456 Patent, col. 1:12-21).
- The Patented Solution: The invention describes a middleware-based system that provides "on the fly" language translation for an application's user interface text without needing to alter the application's source code. It works by intercepting a command to display text, extracting the text for translation, querying a translation service, and then completing the display command with the translated text (’456 Patent, Abstract; col. 3:39-55).
- Technical Importance: This approach aimed to dramatically lower the barrier for localizing software applications by decoupling the translation process from the application's core code.
Key Claims at a Glance
- The complaint asserts infringement of at least Claim 1 (Compl. ¶51).
- Independent Claim 1 requires:
- intercepting, by a processor, a command to display user-interface text in a first language;
- extracting user-interface text to translate from the command;
- querying a translation mechanism using the extracted text;
- receiving translated text in a second language from the mechanism; and
- displaying the translated text.
U.S. Patent No. 9,088,694 - "Adjusting video layout"
- Patent Identification: U.S. Patent No. 9,088,694, "Adjusting video layout," issued July 21, 2015 (Compl. ¶22).
- Technology Synopsis: The patent addresses the challenge of optimizing video conference layouts for displays of different sizes. The described solution involves detecting a region of interest (e.g., a participant's face) within a video feed and automatically scaling, cropping, and centering the video to ensure the region of interest remains prominent, regardless of the screen size (’694 Patent, Abstract).
- Asserted Claims: At least Claim 1 (Compl. ¶67).
- Accused Features: The complaint accuses NVIDIA's Broadcast App, which allegedly provides an "Auto Frame" feature that "dynamically track[s] your movements in real time... automatically cropping and zooming so you remain the star of the show" (Compl. ¶¶65, 68).
U.S. Patent No. 10,574,201 - "Dynamic time-weighted systems and methods for management of acoustic exposure"
- Patent Identification: U.S. Patent No. 10,574,201, "Dynamic time-weighted systems and methods for management of acoustic exposure," issued February 25, 2020 (Compl. ¶23).
- Technology Synopsis: The patent describes a system for managing a user's cumulative acoustic exposure to prevent hearing damage. The system monitors audio levels from a first source, and if those levels exceed a predefined threshold, it executes an "audio limiting action" to reduce the audio output of a second source, thereby managing the user's total sound exposure over time (’201 Patent, Abstract).
- Asserted Claims: At least Claim 16 (Compl. ¶83).
- Accused Features: The complaint accuses NVIDIA Maxine's Audio Effects SDK, which allegedly processes audio for noise removal by determining if background noise is above a threshold and, if so, executing a "background noise reduction (i.e., audio limiting action)" (Compl. ¶¶81, 85).
U.S. Patent No. 11,394,923 - "Embedding content of interest in video conferencing"
- Patent Identification: U.S. Patent No. 11,394,923, "Embedding content of interest in video conferencing," issued July 19, 2022 (Compl. ¶24).
- Technology Synopsis: The patent discloses a system that makes video conferences more interactive by monitoring the session, sensing a "context" within the content (e.g., a topic of discussion), and automatically performing a related action, such as displaying relevant content or advertising (’923 Patent, Abstract).
- Asserted Claims: At least Claim 1 (Compl. ¶99).
- Accused Features: The complaint accuses NVIDIA ACE, a platform for creating AI-powered avatars. The allegations contend that this system acquires user content (audio), senses a context (emotion), and automatically performs an action (animating an avatar in response to the emotion) (Compl. ¶¶97, 100-102).
III. The Accused Instrumentality
Product Identification
- The accused instrumentalities are a suite of NVIDIA software platforms and development kits, including NVIDIA Maxine, NVIDIA Riva, NVIDIA Broadcast App, NVIDIA Maxine's Audio Effects SDK, and NVIDIA ACE (Compl. ¶¶34, 49, 65, 81, 97).
Functionality and Market Context
- The complaint describes these products as providing real-time, AI-powered enhancements for audio and visual communications (Compl. ¶¶34, 49). Functionalities cited include generating AI avatars to use in place of live video, real-time language translation, automatic video framing, background noise removal, and animating digital characters based on emotional cues in a user's voice (Compl. ¶¶36, 51, 68, 85, 101). Figure 4 of the complaint shows a promotional image from NVIDIA demonstrating the creation of AI-powered avatars from user audio and images (Compl. p. 10). The complaint positions these products as tools for developers to integrate AI effects into video conferencing, streaming, and other applications to create more immersive and interactive experiences (Compl. p. 9, Fig. 2).
IV. Analysis of Infringement Allegations
- ’083 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| determining that a destination endpoint is not providing a media stream of a first media type to an originating endpoint | The NVIDIA Maxine products provide an option for a meeting participant to not use a live video image of themselves, which constitutes a determination that the live video media stream is not being provided. | ¶36 | col. 2:21-26 |
| initiating transmission of default media of said first media type to said originating endpoint | In place of the user's live video, the accused products initiate transmission of an AI-generated avatar of the user in a video format, which is alleged to be the "default media." | ¶¶36, 38 | col. 2:26-29 |
Identified Points of Contention:
- Scope Questions: A central question may be whether an AI-generated, dynamic avatar, created in real-time from a user's photo and audio, qualifies as "default media" within the meaning of the ’083 patent. The analysis may consider whether the patent’s use of "default" anticipated pre-stored content (e.g., a placeholder video) rather than computationally generated content.
- Technical Questions: The complaint alleges that opting not to use a camera constitutes "determining that a destination endpoint is not providing a media stream" (Compl. ¶36). A factual question may arise as to whether this user choice is technically equivalent to the automated determination process described in the patent, which is triggered by events like call transfers to incompatible systems (’083 Patent, col. 1:5-11).
’456 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| intercepting, by a processor... a command to display user-interface text on the screen in a first language... | NVIDIA Riva, which is described as a collection of GPU-accelerated microservices, intercepts a command to display user-interface text in a first language. | ¶52 | col. 2:15-18 |
| extracting user-interface text to translate from the command | The accused products extract the user-interface text from the command in order to translate it. | ¶53 | col. 2:18-19 |
| querying a translation mechanism by use of the extracted user-interface text | NVIDIA Riva queries a translation mechanism, such as its NeMo Machine Translation models, using the extracted text. | ¶53 | col. 2:19-20 |
| receiving translated user-interface text in a second language from the translation mechanism | The accused products receive the translated text from the translation mechanism. | ¶54 | col. 2:20-22 |
| displaying the translated user-interface text in the second language | NVIDIA Riva displays the translated text. The complaint provides Figure 9, a screenshot of a Riva demonstration translating "Riva translation uses Megatron model..." from English to Hindi. | ¶54 | col. 2:22-23 |
- Identified Points of Contention:
- Scope Questions: The claim requires "intercepting" a command. The dispute may focus on whether the accused NVIDIA Riva product performs a true, low-level interception of a standard display command without the application's explicit cooperation, or if it functions as a software library that an application developer must consciously call to perform translation.
- Technical Questions: The complaint alleges Riva "intercepts, by a graphics processing unit (GPU)" a command (Compl. ¶52). A key evidentiary question will be whether the complaint provides sufficient technical detail to demonstrate that this specific architectural implementation meets the claim limitation, particularly as the patent describes the invention in the context of a Java Virtual Machine (JVM) layer (’456 Patent, col. 3:45-50).
V. Key Claim Terms for Construction
- The Term: "default media" (’083 Patent, Claim 1)
- Context and Importance: The definition of this term is critical because the infringement allegation hinges on whether an AI-generated avatar can be considered "default media." Practitioners may focus on this term because it directly links the patent's original technological context (preventing a "dark screen" during a call transfer) to the modern, accused functionality (AI avatars).
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The patent does not narrowly define the content of the "default media," referring to it as a "default video stream" that can be provided until the call is connected to a destination that delivers its own video content, suggesting its purpose is to fill a void (’083 Patent, col. 2:40-47).
- Evidence for a Narrower Interpretation: The specification provides examples of default media as coming from a "media library" or an "interactive media server," which could suggest pre-existing or pre-recorded content rather than content generated dynamically in real-time based on user input (’083 Patent, Fig. 1).
- The Term: "intercepting" (’456 Patent, Claim 1)
- Context and Importance: This term is central to the non-obviousness of the patented solution and the infringement case. The core of the invention is enabling translation without modifying the application's source code. Whether NVIDIA Riva "intercepts" commands or is simply called by an application via an API will likely be a dispositive issue.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The patent describes the invention in the context of a "middleware layer" that "intercepts and translates textual messages" (’456 Patent, col. 2:17-18). This could be argued to cover any software architecture that sits between an application and the operating system to modify display text.
- Evidence for a Narrower Interpretation: The background explicitly contrasts the invention with methods that require developers to "alter the computer source code of an application" (’456 Patent, col. 2:53-56). The specification's solution involves modifying a Java Virtual Machine layer to intercept calls, a technique designed to be transparent to the application itself (col. 5:1-5). Evidence that using the accused Riva product requires application code changes could support a narrower construction that excludes such implementations.
VI. Other Allegations
- Indirect Infringement: For each asserted patent, the complaint alleges that Defendant induces infringement by, among other things, providing instructions, user manuals, advertisements, and software development kits that "facilitate, direct or encourage the use of infringing functionality" by end users and developers (Compl. ¶¶14, 41, 57, 73, 89, 105).
- Willful Infringement: The complaint alleges willful infringement based on Defendant's alleged knowledge of the patents, stemming at least from the receipt of Plaintiff's correspondence on October 10, 2025, and its subsequent and continued infringing conduct (Compl. ¶¶28, 43, 59, 75, 91, 107).
VII. Analyst’s Conclusion: Key Questions for the Case
- A core issue will be one of definitional scope: Can the term "default media," as used in a 2005-priority patent to describe a placeholder for a missing video stream, be construed to cover a dynamically generated, AI-powered avatar as alleged to be used in NVIDIA's Maxine product?
- A second primary issue will be one of architectural function: Does the accused NVIDIA Riva software perform a true, low-level "interception" of display commands transparently to an application, as required by the '456 patent, or does it operate as a high-level API that requires explicit, source-code-level integration by developers?
- A broader evidentiary question will be one of technological translation: Across all five asserted patents, which were filed over a 14-year span, the case will likely examine whether the specific AI/ML-based mechanisms in NVIDIA's modern software platforms perform the same function in substantially the same way as the methods described in the patents, which are often framed in the context of earlier technologies like IVR systems and Java Virtual Machines.