Plumerai: Bringing vision AI to millions of devices

Backing Plumerai’s $8.7m Series A to bring Vision AI to even the most power-constrained edge devices

We are delighted to co-lead Plumerai’s $8.7m Series A. Plumerai, a leader in ultra-efficient vision AI is redefining what’s possible for edge devices with its proprietary Tiny AI on the edge combined with Vision LLMs in the cloud - delivering advanced visual intelligence at a fraction of the cost, power, and latency of traditional solutions. We co-led the round alongside Partech, with Acclimate Ventures also joining as an investor. The company is further backed by renowned early supporters, including Tony Fadell (a creator of the iPod) and Hermann Hauser (a co-founder of Arm).

Why vision AI on edge devices needs a rethink

Despite rapid progress in AI, most vision edge devices remain far less intelligent than they could be. The challenge lies in a trade-off between on-device and cloud technologies:

• On-device AI demands expensive, power-hungry chips that are impractical for battery-powered or low-cost devices.
• Cloud AI is easier to deploy but introduces substantial operating costs, multi-second delays, lower accuracy due to video compression, and privacy concerns.

This tension has left manufacturers and users stuck with underwhelming products that fail to deliver on the promise of truly smart devices.

Plumerai’s breakthrough: combining Tiny AI with Vision LLMs

Plumerai delivers a vertically integrated AI stack that brings together the best of two worlds:

• Tiny AI on the edge – an ultra-compact (~5MB) neural model optimized to run directly on resource-constrained chips, delivering instant results with minimal power, bandwidth use, and full GDPR compliance.
• Vision LLMs in the cloud – activated only when needed and powered by pre-filtered, high-resolution inputs from Tiny AI, unlocking features such as semantic video search and intelligent action or event captions.

Rooted in the team’s early experience designing its own AI chip, Plumerai’s Tiny AI stack brings rare system-level expertise. It uniquely runs on low-cost and low-power chips priced as little as $1–$4 - dramatically lowering cost, compute, and power needs while providing unmatched accuracy and cutting latency from 7 seconds to just 0.7 seconds. Its hybrid edge–cloud architecture reduces cloud inference costs by up to 135x (based on customer benchmarks) for more advanced features like AI Video Search. And because its software is hardware-agnostic and deployable via simple over-the-air updates, manufacturers can upgrade millions of devices already in use - driving adoption at scale.

From early traction to mass-market adoption

Plumerai is already proving its value in the real world. Its AI stack powers devices from Chamberlain Group (myQ, LiftMaster), reaching over 50 million homes worldwide. Today, the company enables features such as people, vehicle, animal & package detection, familiar face & stranger identification, and multi-camera tracking, eliminating frustrating false alarms caused by shadows, lighting changes or weather conditions. And this is just the beginning. Plumerai’s roadmap includes advanced applications such as suspicious activity detection and video-based Q&A answering, opening doors to use cases in enterprise security, retail analytics, healthcare monitoring, and beyond.

The road ahead

Plumerai’s long-term vision is ambitious: to create an always-on vision AI layer for the physical world. The company’s technology has the potential to transform devices into assistants and protectors - ones that never miss a detail, adapt seamlessly to user preferences, and operate reliably across homes, offices, warehouses, and care facilities.

Behind this vision is a top-tier team led by Roeland Nusselder, combining deep technical expertise with proven commercial execution. We are excited to partner with them as they unlock the next wave of edge intelligence – with superior user experience, greater privacy, and lower cost, even on the most affordable and battery-powered devices.
 
Read more in Tech.EU

Media Contact:  

Kekst CNC 

Victoria Douglas / Mia Nolan

+44 7908 638 573 / +44 7581 063661

otbventures@kekstcnc.com