Adobe and NVIDIA have expanded their partnership so that several of Adobe’s flagship creative and marketing tools run faster on NVIDIA hardware and open models, the companies said in a joint release on 15 October 2026.
The announcement covers three tracks: desktop creative apps optimised for NVIDIA’s RTX Spark superchip; cloud acceleration for Frame.io; and Brand Concierge tuned with NVIDIA Nemotron open models. The performance numbers in the release are the vendors’ own claims — “up to” figures and “early” results from selected workloads — not independent benchmarks.
What is being optimised
According to the companies, Adobe is bringing RTX Spark — which NVIDIA launched earlier this month — to its agentic creative applications Photoshop, Premiere and Substance 3D. Optimised to use RTX Spark’s unified memory and AI compute, those apps are said to deliver up to 2x faster AI processing, editing, colour and effects.
Jason Paul, vice president of local AI at NVIDIA, said: “Powered by RTX Spark, Adobe’s optimized tools deliver up to 2x faster performance to accelerate these new workflows.” Deepa Subramaniam, Adobe’s vice president of product marketing for creative professionals, said the companies are “delivering the performance creatives need to bring their ideas to life” by optimising those apps for RTX Spark and accelerating Frame.io in the cloud.
Photoshop, Premiere and Substance 3D — as Adobe describes them
Adobe says it rearchitected Photoshop around GPU-accelerated compositing so tasks such as the Remove Tool and on-canvas effects like Gaussian Blur and Light Leaks feel faster and more responsive.
In Premiere, the companies say some of the most demanding post-production paths — playback, rendering, effects, colour and AI tools — are RTX-accelerated and tuned for RTX Spark, with tangible gains claimed for features including Enhance Speech and Object Mask.
For Substance 3D Painter, RTX Spark is said to speed the heaviest texturing steps. None of those claims includes a public third-party test suite or a ship date for the optimised builds in the release.
Frame.io cloud numbers (selected workloads)
Adobe is also using NVIDIA accelerated computing for Frame.io cloud media processing and content discovery. The updated pipeline, the release says, uses NVDEC for hardware-accelerated decoding and TensorRT with CUDA for preprocessing and embedding generation.
Early performance results from selected Frame.io transcode and search workloads, as stated by Adobe and NVIDIA:
- Approximately 5x faster embedding generation
- 4–5x improvement in P95 processing latency
- An approximately 80% smaller search proxy
Those are vendor figures on chosen workloads. TSN has not reproduced them. The companies say the practical effect for teams would be that uploads become searchable sooner and large libraries refresh more efficiently.
Brand Concierge on Nemotron
Adobe and NVIDIA also worked to optimise Adobe Brand Concierge with NVIDIA Nemotron open models. That version “will be available to enterprise customers later this year,” the release says. NVIDIA has selected Brand Concierge to power personalised conversational experiences on NVIDIA.com, grounded in organisational data and content, while using Nemotron for responses at scale.
Context: March partnership and Adobe MAX
The work builds on a strategic partnership the companies announced in March, covering creative and marketing workflows, open models, libraries and accelerated computing across video, images, 3D and other media — including earlier GPU acceleration for Frame.io and work around Substance 3D, Firefly Foundry and 3D digital twin workflows.
Adobe and NVIDIA say they will showcase the latest creator advancements with hardware partners at Adobe MAX, 10–12 November.
Why it matters
Creators already juggle image, video, audio and 3D in one project; agents and generative tools add another layer of compute demand. Pairing Adobe’s mainstream apps with a local high-memory chip (RTX Spark) and cloud GPU paths is meant to keep interactive editing and review from becoming the bottleneck. Separately, putting Brand Concierge on open Nemotron models is Adobe’s bid to give enterprises more model choice for conversational brand experiences.
Readers should keep the “up to 2x” and Frame.io multipliers in the same bucket as other vendor performance slides: useful as directional claims, not as guaranteed speed-ups on every machine or every project.
What’s still unknown
- When the RTX Spark-optimised Photoshop, Premiere and Substance 3D builds ship, and on which SKUs.
- Pricing changes, if any, tied to the optimisations.
- How the selected Frame.io workloads compare with a typical customer library.
- Which Nemotron model sizes power Brand Concierge at general availability.
What to watch
Watch Adobe MAX in November for demos and any firmer availability dates, and watch NVIDIA.com for the live Brand Concierge experience the companies describe.
What this means
- Partnership expanded across apps and cloud. Photoshop, Premiere, Substance 3D, Frame.io and Brand Concierge are in scope.
- Speed claims are the vendors’. “Up to 2x” and Frame.io multipliers come from Adobe/NVIDIA, including early selected-workload results.
- Brand Concierge on Nemotron is later this year. Enterprise availability is forward-looking in the release.
Sources
Related stories
- NVIDIA Readies Its RTX Spark Laptops for Local AI Agents With a New Open Model, Nemotron 3.5 Bolt
- NVIDIA Releases Nemotron 3.5 Super VL, an Open 120B Multimodal MoE for Agents
- Meta Releases the Weights for Muse Spark Open, and Apple Shows It Running Across Two to Four Mac Studios
Follow @tsnmediahub on X for breaking AI news as it happens.
Follow on X
