Adobe and Nvidia expanded their partnership on 15 October to optimise Photoshop, Premiere and Substance 3D for the forthcoming RTX Spark chip, with Nvidia claiming up to 2x faster AI processing, editing, colour and effects in the adapted applications. The work also covers Frame.io and Adobe Brand Concierge.
Key points
- Nvidia says RTX Spark launches on 16 October, with Adobe adapting three creative applications for the chip.
- Nvidia reports approximately 5x faster embedding generation in selected Frame.io workloads.
- Adobe plans to offer a Nemotron-based version of Brand Concierge to enterprise customers later in 2026.
- Nvidia has selected Brand Concierge for customer conversations on Nvidia.com.
Photoshop and Premiere move onto RTX Spark
RTX Spark is a system-on-chip intended to run AI tasks locally on Windows laptops and small desktops, Engadget reported. It described RTX Spark as aimed at creatives and novice developers, in contrast with Nvidia’s DGX Spark, which it said mainly targets companies and AI developers. Adobe’s changes put AI processing inside applications that professionals already use for images, video and 3D work.
Nvidia says the optimised applications use RTX Spark’s unified memory and AI computing capabilities. Its up-to-2x performance claim covers AI processing, editing, colour and effects across the applications. Deepa Subramaniam, Adobe’s vice-president of product marketing for creative professionals, said the companies are optimising Photoshop, Premiere and Substance 3D for the chip while accelerating Frame.io in the cloud.
Adobe has rebuilt Photoshop’s compositing around GPU acceleration, according to Nvidia. The company identifies removing objects with the Remove Tool and applying Gaussian Blur or Light Leaks on the canvas as tasks affected by the change. In Premiere, Nvidia says RTX Spark optimisations cover playback, rendering, effects, colour and AI-powered tools, including Enhance Speech and Object Mask.
Substance 3D Painter’s texturing work is another target. Engadget reported that the optimised application also speeds up opening and saving large projects and exporting textures. The changes span routine stages of creative production rather than a separate AI application. Nvidia’s claimed speed gain is a performance measure for those workflows, not a forecast of Adobe sales or customer adoption.
Frame.io targets video search and processing
Adobe is also applying Nvidia computing to Frame.io’s cloud media pipeline. Nvidia says the updated pipeline uses NVDEC to decode video in hardware, while TensorRT and CUDA support preprocessing and the generation of embeddings used in content discovery. This work concerns processing uploaded media and making it searchable, rather than running an editing application locally on RTX Spark.
In early results from selected Frame.io transcoding and search workloads, Nvidia reports approximately 5x faster embedding generation, a 4-5x improvement in P95 processing latency and an approximately 80% smaller search proxy. The company says moving decoding, preprocessing and embedding generation onto its GPUs lets Frame.io process files and make them searchable sooner. The performance figures apply to the selected workloads Nvidia identified.
The Frame.io work formed part of the strategic partnership Adobe and Nvidia announced in March 2026. Nvidia says that earlier arrangement also covered its open models and accelerated computing across Adobe’s creative and marketing workflows, with work extending to Substance 3D, Firefly Foundry and 3D digital twins for enterprise customers. AI Affairs previously reported that Nvidia and Microsoft opened RTX Spark laptop preorders for October delivery.
Nemotron brings Brand Concierge to Nvidia.com
Adobe and Nvidia have separately optimised Brand Concierge with Nvidia’s Nemotron open models. Nvidia says the version is intended for enterprise conversational experiences that draw on an organisation’s data and content. Unlike the RTX Spark application work, this part of the partnership concerns customer interactions delivered by a brand.
Adobe plans to make the Nemotron-based version available to enterprise customers later in 2026, Nvidia says. Nvidia has also selected Brand Concierge for personalised conversations on Nvidia.com, with responses grounded in organisational data and content.
