Acer has shown a compact desktop design built around Nvidia’s RTX Spark platform, a system aimed at running demanding AI and graphics workloads locally. The company presented the Acer SFF RTX Spark at its next@acer press conference in Berlin on September 2 during IFA 2026.

Acer describes the device as a small-form-factor desktop for creators, AI developers and gamers. Its local-processing focus means the work can run on the computer instead of depending on a remote cloud service for every task.

The company lists up to 6,144 Blackwell RTX GPU cores and up to a 20-core Nvidia Grace CPU for the design. It also lists up to 128GB of unified memory, which can be shared by the CPU and graphics processor rather than divided into separate pools.

Acer says the system can deliver up to 1 petaflop of AI compute. That figure describes the platform’s peak capability under the company’s stated conditions. It is not a promise that every application or consumer workload will reach that result.

TechRadar reported that Acer’s design is aimed at a niche group of users who need local AI or heavy creative performance in a compact enclosure. The publication also noted that the device’s usefulness for ordinary desktop buyers will depend on pricing, software support and actual availability.

Local processing can matter for developers who test models with private data or who need predictable access to compute when an internet connection is limited. It can also reduce the need to move large project files to a remote service, although users still need software that supports the hardware.

Acer has not announced a retail price or a shipping date for its RTX Spark design. The company says availability will be announced at a future date. The show model should therefore be treated as a confirmed product design, not as a product that can already be ordered.

The announcement gives Acer a place in Nvidia’s first RTX Spark hardware group. The next details that matter are the final chassis, cooling system, operating-system support, regional pricing and the applications that can use its local AI and graphics resources.