Axelera AI

Point at a lamp. It turns on. Point at the TV. It changes channel. No remote, no voice, no cloud. Project GAIUS by @KatusDavid: multi-camera 3D triangulation, arm raycasting, gesture recognition on Metis M.2. Fully offline, fully open source. Code + docs: eu1.hubs.ly/H0vmv_j0 #EdgeAI #ComputerVision 👆🏠

We're heading to Computex 2026 in Taipei at the beginning of next month, and we've got a lot to show you. Put our booth on your itinerary now, so you won't miss out on: 🔵 Europa running a VLM to help find lost people across video streams 🔵 A robot arm running object detection LIVE on-device 🔵 Speech LLM on a Metis M.2 Max 🔵 Your first chance to see Axelera's hot new hardware range live and in action Drop us a message if you want to connect while you're there: eu1.hubs.ly/H0vgg2G0 #Computex2026 #EdgeAI #AxeleraAI #Europa #GenerativeAI #AIHardware

Wow, nice work on the Axelera community here. He actually beat our own benchmarks! 🤩 eu1.hubs.ly/H0v8WW_0

Italy has world-class engineering talent, serious research infrastructure, and a growing community of founders building for global markets. It is one of the reasons we have an engineering team in Florence and an office in Milan. The Bologna Gathering brings that ecosystem together with the investors and partners helping it scale into AI and the next generation of innovative technology. Our CEO and co-founder Fabrizio Del Maffeo is excited to join the conversation and take the stage to share his insights. If you are attending, we would love to connect. eu1.hubs.ly/H0v5ydv0 #AxeleraAI #tbg26 #ItalianTech #EuropeanTech #EdgeAI

Sweden is building one of Europe's first AI-only supercomputing systems. It will give researchers, SMEs, and public institutions cloud-style access to advanced AI workloads, supporting sovereign AI development and breakthrough applications in life sciences, materials science, and autonomous systems. Procured by the EuroHPC Joint Undertaking, we're honored to have been selected to deliver Axelera AIPUs into this landmark system. eu1.hubs.ly/H0v4ZrR0 #EdgeAI #EuroHPC #SovereignAI #HPC

A long time ago, in a cloud far, far away… TECH WARS Episode IV A NEW INFERENCE It is a period of AI war. Rebel engineers have struck a blow against the Hyperscaler Empires. Their ultimate weapon, the DATA CENTRE, an armoured GPU farm with enough power to bankrupt an entire country, is about to take over the galaxy. Pursued by the Empires' sinister latency, Princess Metis races home aboard the starship Voyager, custodian of powerful Axelera tech that can save her people and restore inference to the edge… May the 4th be with you. 🌌 #StarWarsDay #EdgeAI eu1.hubs.ly/H0t_zBb0

I sense a great disturbance in the forks... #StarWarsDay #EdgeAI eu1.hubs.ly/H0t_7h90

You don't know the power of the docs site! #StarWarsDay #EdgeAI eu1.hubs.ly/H0t_5Wh0

You've never heard of Axelera AI's Voyager? It's the SDK that made the Metis run in less than twelve parsecs. It's outrun hyperscaler GPUs. Not the local inference boxes, mind you. I'm talking about the big datacenter racks now. She's fast enough for you, old man... #StarWarsDay #EdgeAI eu1.hubs.ly/H0t_7Jj0

Our AI friends over at @Ultralytics has put up a new partner page covering YOLO on Metis, which we really want you to take a deep dive into: eu1.hubs.ly/H0tYFwB0 Essential stuff if you're deploying YOLO at the edge: • One-command export from Ultralytics straight to .axm • YOLOv8 through YOLO26 supported, no manual configuration • Four models running in parallel on independent cores per chip • No PyTorch needed at inference time Full integration docs are here: eu1.hubs.ly/H0tYHmj0 And you can join the action for free at: eu1.hubs.ly/H0tYGHp0 #EdgeAI #YOLO 🚀

Our first Demo Jam was supposed to be a low-effort thing. That is not what happened. The brief was simple: 1 to 2 minute videos of anything built with Axelera Metis and the Voyager SDK. A cool technique, a feature demo, a work-in-progress. We pitched it as low-commitment on purpose because we wanted people to actually submit. Instead, the community handed in: 🔬 A real-time closed-loop microscopy system running unattended behavioural studies on live microorganisms, in an actual research lab 🏎️ Gesture-controlled racing games with custom pipeline postprocessors 🗿📃✂️🦎🖖How to play Rock, Paper, Scissors, Lizard, Spock against an AI 🕺Real-time pose-matching party games scored on keypoint similarity 🎹Air pianos and invisible drum kits All of it running on-device. No cloud, no latency, all at the edge. The winner (and the unexpected bonus prize) is right here: eu1.hubs.ly/H0tNGY50 Thanks to everyone who turned up. Demo Jam #2 is already queued. #EdgeAI #ComputerVision #AxeleraAI

5 hours 49 minutes. That's how long one live Paramecium was tracked continuously by an unattended edge AI microscopy rig built by Winkoms Open Microscopy. A Metis M.2 running YOLOv8-seg, steering a motorised XY stage to keep a single microorganism centered in the field of view. It's being used with Dr. Humbert Salvadó's research group at Universitat de Barcelona for long-duration behavioral studies. Full write-up from FxGc on the Axelera AI community of this impressive, and highly practical, lab-based AI : eu1.hubs.ly/H0tC-3D0 #EdgeAI #ComputerVision 🔬

Real lessons from building AI inference on RISC-V. Koby Soden, Solutions Architect at Axelera AI, is speaking at RISC-V Now! on April 21st in San Jose, CA. His session, "Building an AI Inference Platform on RISC-V," covers what it takes to move from edge vision to generative AI using purpose-built RISC-V architecture and what that means for efficiency at the edge today. If you're building on RISC-V or evaluating AI inference platforms, this one is worth your time. Register: eu1.hubs.ly/H0t8vMf0 #RISCV #RISCVNow #EdgeAI

The recording of our live session with @ultralytics and Innowise is now up. See how to deploy Ultralytics YOLO models on Axelera Metis AIPUs in minutes using a single command. Read the technical blog: eu1.hubs.ly/H0tm6sp0 Watch the session: eu1.hubs.ly/H0tm3BT0 #EdgeAI #MLOps #YOLO #AxeleraAI #Ultralytics

We're looking forward to joining @ultralytics and Innowise today to share a more efficient way to deploy YOLO models. Our new integration allows for deployment on Axelera AIPUs using only a single command. We will be hosting a live demo and may have discounts available for those who join. Time: 06:00 PDT / 15:00 CEST Link: eu1.hubs.ly/H0tgFfd0 #EdgeAI #ComputerVision #YOLO #AxeleraAI #Ultralytics

MotionFlow: IP cameras become privacy-preserving activity sensors. YOLOv11-Pose on Metis classifies actions from pose geometry. Only structured state over MQTT, no video leaves the device. Running 24/7. Full source on GitHub. #EdgeAI #Privacy 🔐 eu1.hubs.ly/H0t23L70

A.B.C. (AI Book Cataloguer): point a camera at your bookshelf, detect spines with YOLOv8s on Metis M.2, and catalog your library automatically. Built by community member Denovo on @orangepixunlong. Open source. #EdgeAI #ComputerVision 📚 eu1.hubs.ly/H0t23Vb0

Offloaded CCTV object detection from a Pi 5 to a Metis Compute Board. Pi fan stopped screaming. ~30 FPS across two 1080p cameras, ~196ms latency. Zone filtering, Home Assistant webhooks, phone snapshots in 1-2 seconds. Full tutorial + code. #HomeAssistant #EdgeAI 📹 eu1.hubs.ly/H0t23Nx0 @raspberry_pi

Gesture-controlled racing, no controller needed. YOLO11n-pose-hands on Metis for 21-keypoint hand tracking. Tilt wrists to steer, fist to accelerate, open palm to brake. Works with any keyboard racing game. Built by Dilip.m. #EdgeAI #Gaming 🏎️ eu1.hubs.ly/H0t1WCm0

Our condolences to KitKat on the loss of 12 tons of inventory. It’s a tragedy that hits close to home (mostly because we really wanted one). We’ve realized that while our AI is great at security, surveillance & industrial automation, its highest calling might actually be "confectionery preservation." We’re ready to help guard the next shipment with on-site vision that doesn't blink, doesn't sleep, and most importantly, doesn't eat the evidence. #kitkatheist #StolenKitKatTracker #HaveABreak #EdgeAI