On September 20, 2026, PrimeBot named Ricore a compute ecosystem partner to explore edge inference for humanoid robots. The D35 is described as combining a RISC V CPU core with dataflow computing, configurable Chiplet packaging, a software toolchain and support for up to 128 GB of high speed video memory per chip.[1...
Published byEdited with GPT-6 SolImages generated with GPT Image 2
Research answer

Create a landscape editorial hero image for this Studio Global article: What did PrimeBot’s September 20, 2026 partnership announcement with Suzhou chip designer Ricore say about the Radiant D35’s role in humanoi. Article summary: PrimeBot named Suzhou-based Ricore a **compute-ecosystem partner** at its September 20 launch, framing the Radiant D35 as a potential source of on-robot inference for humanoid and embodied AI—not announcing that the chip. Topic tags: general, general web, user generated, government. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, ch
At its September 20, 2026, product launch, PrimeBot named Suzhou chip designer Ricore a compute-ecosystem partner. The plan is to explore how Ricore’s dataflow AI chips could meet the on-device inference needs of humanoid and other embodied robots—not to announce that its Radiant D35 is already installed in a PrimeBot robot.18
31
Ricore positions the D35 as an edge-inference chip for “physical AI”: running AI workloads near the robot rather than targeting cloud-based model training. The reported design combines a self-developed RISC-V CPU core with a dataflow computing architecture. Under the partnership, PrimeBot would contribute robot applications and systems-integration expertise, while Ricore would supply chip technology and the compute platform.18
26
The D35 is also described as using Chiplet packaging to configure different compute tiers. Reported software support includes a full-stack toolchain for mainstream deep-learning frameworks and large-model inference deployment. Ricore claims support for up to 128 GB of high-speed video memory per chip; the announcement does not establish that this is on-die memory or specify the memory configuration of any PrimeBot robot.18
26
31
PrimeBot announced the partnership alongside its Q1 and T1 personal robots. The collaboration offers a potential route to domestically designed compute for future embodied-robot products, but reporting describes D35 use in PrimeBot robots as a possibility, not a confirmed Q1 or T1 component choice. One report quotes a Ricore representative expressing the hope that the D35 will be used in PrimeBot robots in the future.18
29
Ricore says it was founded in 2020, holds more than 80 dataflow patents, has a team whose research received an ISCA 2026 best-paper nomination, and has obtained Intel global-procurement recognition for its RISC-V CPU-core IP. Those reported credentials do not independently validate the D35’s performance in a robot.18
Reports describe the D35 as having reached mass production, but the cited partnership coverage does not provide independently measured robot-level inference throughput, latency, power draw or energy efficiency. Nor does it identify a Q1 or T1 production configuration containing the chip. Its robot-integration status remains the key distinction between the partnership’s ambition and a demonstrated product.25
29
31
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
On September 20, 2026, PrimeBot named Ricore a compute ecosystem partner to explore edge inference for humanoid robots.
On September 20, 2026, PrimeBot named Ricore a compute ecosystem partner to explore edge inference for humanoid robots. The D35 is described as combining a RISC V CPU core with dataflow computing, configurable Chiplet packaging, a software toolchain and support for up to 128 GB of high speed video memory per chip.[18][26]