AutoArk Closes Two Pre‑Series A Rounds, Raising Nine‑figure RMB From Shokz, Guoruiyuan Fund, Everpine Capital and Shanghai Angel Group

Why It Matters

Eva OS could dramatically lower the cost and speed of AI‑enabled hardware development, giving startups a viable path to compete with larger model providers and accelerating the rollout of intelligent edge devices.

Key Takeaways

• AutoArk raised nine‑figure RMB (~$14 M) across two pre‑Series A rounds.
• Eva OS lets developers code hardware via natural language in ~30 minutes.
• Edge‑side model cuts voice latency to <250 ms, 70‑92% cost reduction.
• Over 2,500 enterprises use Eva OS for wearables, robots, in‑car assistants.
• Eva Pi terminal can self‑write code and update applications autonomously.

Pulse Analysis

The rise of "vibe hardware" reflects a shift from traditional, fragmented AI pipelines to unified, agent‑centric development. AutoArk’s Eva OS builds on the OpenClaw framework, embedding an AI agent directly into edge devices so that natural‑language prompts translate into fully functional code. This model sidesteps the multi‑stage engineering process that typically requires several specialists and weeks of effort, enabling a single developer—or even a non‑technical user—to launch a functional AI application in minutes. By integrating hardware awareness into the AI’s reasoning loop, Eva OS creates a feedback‑rich environment that accelerates debugging and driver optimization.

Technically, Eva OS adopts a hybrid edge‑cloud architecture that keeps latency‑sensitive tasks—speech recognition, text‑to‑speech, and visual perception—on‑device while offloading complex reasoning to the cloud. The result is sub‑250 ms voice response times and a 70‑92% reduction in compute costs compared with conventional cloud‑first solutions. Its end‑to‑end multimodal foundation model consolidates ASR, TTS, vision and language into a single sub‑gigabyte CPU‑only model, eliminating the information loss and latency penalties of modular pipelines. These efficiencies translate into tangible business benefits, such as the Qiduoduo education robot delivering a full AI experience for under $150 without subscription fees.

The broader market implications are significant. With over 2,500 enterprises already deploying Eva OS across wearables, robotics and in‑car assistants, AutoArk is positioning itself as a middle‑layer platform that could become the de‑facto OS for the next generation of intelligent terminals. Its recent funding—approximately $14 million—signals investor confidence in a solution that bridges the gap between large‑model providers and hardware manufacturers. As edge AI chips become more affordable, the ability to run sophisticated models locally will likely spur a wave of new products, giving nimble startups a competitive edge against incumbents that rely on heavyweight cloud infrastructures. The coming years will test whether Eva OS can establish a sustainable ecosystem and become the standard foundation for AI‑enabled hardware.

Deal Summary

AutoArk, a Chinese AI hardware OS startup, announced the closing of two consecutive pre‑Series A funding rounds backed by Shokz, Guoruiyuan Fund, Everpine Capital and Shanghai Angel Group. The rounds raised a cumulative nine‑figure RMB sum (undisclosed in USD) to support its Eva OS platform and upcoming hardware like Eva Pi.