NVIDIA expands AI for Media suite ahead of IBC 2026, pushing video-authenticity detection to 99.3% accuracy

NVIDIA used the run-up to IBC 2026 — the broadcast industry's flagship conference, running September 11-14 in Amsterdam — to announce a major expansion of NVIDIA AI for Media, its collection of GPU-accelerated SDKs, NIM microservices, playbooks and blueprints for broadcast, sports and streaming workflows. The expansion touches video authentication, motion tracking, frame generation, upscaling, HDR conversion and localization, and arrives with a slate of media-industry partners lined up to ship it.

What's new

The centerpiece update is to the NVIDIA Synthetic Video Detector (SVD), a NIM microservice first announced at SIGGRAPH 2026 that estimates the probability that footage is authentic or AI-generated. NVIDIA says SVD's accuracy has now reached 99.3% for text-to-video content and 97.7% for image-to-video content, with the largest gains on harder image-to-video cases.

SVD is being picked up across the broadcast supply chain:

• Dalet is integrating SVD into a secure, cloud-hosted verification workflow so newsroom editorial teams can submit footage and review authenticity scores and metadata directly inside Dalet's interface.
• TwelveLabs announced general availability of Compliance by TwelveLabs, its first application on the company's video intelligence platform, which folds in SVD's frame-level authenticity signals alongside regional and custom compliance screening.
• Wowza will distribute SVD through its Video Intelligence Framework, layering it onto a Wowza Streaming Engine footprint the company says powers more than 35,000 video deployments in over 170 countries, with support for on-premises, edge, cloud, hybrid and air-gapped deployment.

Beyond authenticity detection, NVIDIA is also expanding:

• NVIDIA 3D Body Pose, which estimates 2D/3D human joint locations and angles from a single camera for sports analytics, officiating, player-safety and virtual-production use cases. Vizrt is using it in live virtual-studio environments to drive real-time 3D lighting, reflections and environmental rendering off tracked body movement.
• Video Frame Generation (VFG), which uses generative AI to interpolate new frames between existing ones, increasing frame rates 2x-4x for smoother slow-motion and high-motion replay. Ross Video is integrating VFG into its Rio Replay platform, currently supporting 6x slow-motion generation for sports replay with 8x interpolation in development.
• Video Super Resolution (VSR), which upscales video while reducing noise, blur and compression artifacts, now with new streaming-vs-quality modes and 10-bit video support, available via the NVIDIA Video Effects SDK and as a NIM microservice.
• NVIDIA TrueHDR, which converts standard-dynamic-range video to HDR in real time, plus LipSync/Active Speaker Detection for multilingual dubbing and localization, and Studio Voice for speech enhancement.
• NVIDIA Holoscan for Media, integrating a Media Exchange Layer (MXL) for connecting live media applications across distributed environments, and NVIDIA Sports Intelligence Playbooks, a framework for fine-tuning open models on proprietary sports footage that NVIDIA says improved multiple-choice accuracy from 53% to 94% in testing.

Additional integration partners named alongside the launch include AI-Media, CAMB.AI, Chyron, Panjaya and Machina Sports.

Context

SVD's debut at SIGGRAPH 2026 came as AI-generated video tools have gotten good enough that broadcasters, news organizations and streaming platforms have started treating authenticity verification as core infrastructure rather than a niche forensics tool. The IBC expansion is NVIDIA's push to embed that verification — plus a broader set of AI-accelerated production tools — directly into the software vendors media companies already run, rather than selling a standalone product.

Why it matters

The headline number is the accuracy jump on synthetic-video detection, but the more consequential move is distribution: NVIDIA isn't just improving its own model, it's getting SVD embedded into Dalet's newsroom tools, TwelveLabs' compliance product and Wowza's streaming infrastructure simultaneously. That puts AI-generated-content detection in front of editorial and compliance teams as a routine step in existing workflows, at a moment when the credibility of broadcast and news video is under increasing pressure from generative tools. The sports-side announcements — real-time 3D pose tracking, higher frame-rate interpolation for slow-motion replay — point to the same pattern: NVIDIA supplying the AI layer underneath tools broadcasters already use, rather than asking them to adopt something new.