Adobe rebuilds Photoshop and Premiere for NVIDIA's RTX Spark
Adobe says Photoshop, Premiere and Substance 3D rearchitected for NVIDIA's RTX Spark chip run AI tasks and editing workflows two times faster, with real-time effect previews.

Updated
Why it matters
- Adobe says Photoshop, Premiere and Substance 3D are two times faster on AI tasks and editing when running on NVIDIA's RTX Spark SoC.
- RTX Spark brings local on-device AI to Windows laptops and small desktops, targeting creatives rather than DGX Spark's enterprise audience.
- Redesigned Photoshop offers real-time previews for effects like Light Leaks and the Remove tool on RTX Spark machines.
- Premiere gains speed in rendering, effects, AI-powered editing and color-grading, per Adobe VP Deepa Subramaniam.
- Substance 3D Painter accelerates texturing steps plus opening, saving and exporting large projects.
Adobe has rearchitected Photoshop, Premiere and Substance 3D to run AI tasks and editing workflows two times faster on computers powered by NVIDIA's new RTX Spark system-on-chip, according to Adobe VP Deepa Subramaniam.
RTX Spark is NVIDIA's new SoC (system-on-chip, the processor that integrates a device's core computing functions) designed to give Windows laptops and small desktops the capability to run AI tasks locally on the device, rather than in the cloud. Unlike NVIDIA's DGX Spark, which mainly targets companies and AI developers, RTX Spark aims at a broader market: creatives and novice developers. Adobe's move signals that mainstream creative software is being rebuilt around local AI hardware, not just workstations and data centers.
What did Adobe actually change?
Subramaniam said the company optimized its creative apps to "take full advantage" of RTX Spark. The scope covers three flagship products: Photoshop, the video editing software Premiere, and the 3D design platform Substance 3D. She framed the work in terms of who depends on these tools.
"Our apps like Photoshop, Premiere and Substance are trusted by the world's most talented storytellers, from photographers and designers creating iconic campaigns, to filmmakers and editors bringing award-winning film and television to life, to artists creating characters and visual effects that bring those stories to the screen," Subramaniam said.
She added: "For this audience, performance across creative workflows is critical to allow them to stay focused on their story and craft." The takeaway: Adobe is positioning hardware-software co-optimization as a productivity claim, not a novelty.
How much faster are the apps?
Subramaniam put a single headline number on the work: the optimized apps are two times faster on AI tasks and editing across workflows. That is a vendor-supplied figure without an independent benchmark, and Adobe did not specify which hardware configurations or test conditions produced it.
Per application, the claimed gains break down as follows:
- Photoshop: more responsive and faster at rendering complex workflows on RTX Spark machines; users see changes in real time, such as when applying effects like Light Leaks or removing elements with the Remove tool.
- Premiere: "real, tangible speed gains" in rendering, applying effects, AI-powered editing, and color-grading for audio and video.
- Substance 3D Painter: accelerates "the heaviest steps around texturing," plus other tasks like opening and saving large projects and exporting textures.
Note the asymmetry in the claims. Photoshop gets a concrete, verifiable behavior — real-time previews. Premiere and Substance 3D get qualitative language: "tangible," "heaviest." Adobe has not published benchmark methodology, so the two-times figure should be read as a marketing claim until third-party tests appear.
Why does local AI silicon matter for creatives?
The announcement ties two threads together: NVIDIA pushing AI-capable chips into mainstream laptops and small desktops, and Adobe rebuilding its software to use them. Local, on-device AI processing means effects, generation and editing tools can run without sending work to cloud servers — relevant for latency-sensitive tasks like real-time previews, and for professionals working with large media files.
The distinction between RTX Spark and DGX Spark also matters for the market. DGX Spark serves enterprises and AI developers; RTX Spark targets the far larger pool of everyday creatives and hobbyist developers. If the performance claims hold, laptop-class machines could handle workloads that previously required discrete high-end GPUs. If they don't, creatives on RTX Spark devices will see modest gains concentrated in specific AI-assisted features rather than across the board.
What comes next?
The immediate question is timing and verification. Adobe has announced the rearchitected apps and their claimed speedups but has not, in this announcement, detailed release dates or pricing changes tied to the optimization. Independent reviewers will need to test the two-times claim on shipping RTX Spark hardware before it can be treated as settled fact.
Source: Engadget
More from James Calloway
Show full bio
News editor covering people, launches and day-to-day practice in artificial intelligence and machine learning at AI In Context.
279 articles
Related articles
- NVIDIA's $3,700 RTX Spark Chip Takes Apple Silicon Head-On
- Microsoft prices Surface Laptop Ultra from $2,600, debuts RTX Spark workstation
- AI-Built Open Source Clones Take Aim at Adobe's Entire Creative Suite
- Microsoft's $2,599 Surface Laptop Ultra Bets on Nvidia RTX Spark
- Microsoft Gathers Nadella and NVIDIA's Jensen Huang for AI PC Reset