On May 21, 2026 Flipper Devices Inc.

Flipper Devices Inc. announced the Flipper One on May 21, 2026 — a palm-sized Linux pocket computer that succeeds the Flipper Zero and repositions the company toward a developer‑oriented cyberdeck. The device combines substantially upgraded compute with an explicit engineering commitment to mainline Linux and open drivers, a change that matters for users who need a portable, auditable platform for networking, radio and on‑device AI work.

The Flipper One centers on a 2.2 GHz octa‑core RK3576 system‑on‑chip paired with a Mali‑G52 GPU and a neural processing unit rated at 6 TOPS, and it ships with 8GB of onboard RAM. Flipper Devices describes the platform as fully open: mainline kernel support is promised and the company rejects binary blobs, closed drivers, proprietary firmware and vendor‑locked board support packages in favor of an auditable software stack.

Alongside the RK3576 the product includes a secondary dual‑core RP2350 microcontroller borrowed from Raspberry Pi designs to run the user interface and power management. That microcontroller drives the display, physical buttons, touchpad, LEDs and the power subsystem, and the Flipper One can operate entirely on the RP2350 when users choose not to run Linux — a mode intended to improve idle and low‑power efficiency.

The company positions the Flipper One for builders of IP‑connected hardware, networking tools, software‑defined radio and local AI or LLM projects. Coverage of the announcement highlights practical use cases such as a portable network analyzer, an SDR testbed and offline LLM experiments that can be executed on the device’s NPU and within its 8GB RAM footprint.

For hardware and software developers the combination of an RK3576 that plays nicely with mainline Linux and the absence of vendor‑locked binaries simplifies kernel and driver work, supports custom distributions and eases integration of local inference or networking stacks. The 6 TOPS NPU and 8GB of RAM in particular enable on‑device inference for smaller language models or dedicated AI features without full dependence on cloud services. the announcement excerpt did not include shipping dates or pricing.