The DGX Spark 64GB launches October 23 from $4,999, with OEM-only builds, full model compatibility, and a price hike for the 128GB model.

#NVIDIA #DGXSpark #LocalAI #AIAgents #OpenSourceModels #ASUS #Qwen #EdgeAI

📢 #CoEvolution at the “Standardisation FG Landscape Maintenance on EU Funded Projects” webinar!

Our partner BEVD presented the CoEvolution approach to making #EdgeAI trustworthy, focusing on standardisation aspects.

👉 Read more: coevolution-project.eu/coevolution-...

WebApps: Integration von Webhooks, n8n und APIs

#KI #MaschinellesLernen #DeepLearning #LLM #GenerativeKI #KIModelle #KIHardware #NVIDIABlackwell #KIRegulierung #ExplainableAI #KIAssistenten #AgenticAI #MultimodaleKI #RAG #OnDeviceKI #EdgeAI

beflash.de/n8n-automati...

On-premises CourseChat, a multi-course RAG tutor for undergraduate business education, runs on twin edge AI hosts with a local vector database and LLMs served by Ollama, intended to sit behind a campus gateway…

#Semiconductors #AIInfrastructure #EdgeAI #EdTech
https://arxiv.org/abs/2610.02510

Small models, big test: PrismML’s Bonsai brings compressed vision AI to battery-limited smart glasses.

#VisionAI #EdgeAI #Wearables https://spaisee.com/article/prismml-puts-compressed-vision-ai-on-the-smart-glasses-road

NVIDIA's DGX Spark 64GB ships Oct 23 at $4,999 —
100B params on one desk.

The 128GB FE is $6,950: +$1,951 (+39%)
for double the memory, same GB10 chip.

So the price ladder isn't compute. It's memory.

Per-token billing meters a model's words, not its state. #EdgeAI

This arXiv paper proposes vFedProtoQNAS, a federated learning approach where clients train device-specific quantum neural networks and share class prototypes instead of model parameters. It targets…

#QuantumAI #FederatedLearning #EdgeAI #NeuralArchitectureSearch
https://arxiv.org/abs/2610.01718

Who checks an AI model after it leaves the factory? My lane on the Trusted AI Supply Chain panel at the EDGE AI Defense Symposium: a satellite has no recall, no re-image, no flight line, so the device must run only what is authorized, refuse tampering, and keep a record. #EdgeAI

If you think IoT maturity is about adding more devices, you are missing the real shift.

Edge AI and Swarm IoT bring intelligence closer to operations, where systems can coordinate before cloud delay weakens the response.

#EdgeAI #InternetOfThings

The first NAS run cost 22,400 GPU-days. Weight sharing cut that about 50,000x. The search strategy got the credit; cheaper evaluation did the work.

Edge AI teams: do you search against your own device's latency, or ship a backbone searched for someone else's?

#EdgeAI #Jetson

Edge AI chips must balance flexible compute, data movement, and built-in security as models evolve faster than silicon cycles. How can architects design chips that stay adaptable after tapeout?
#EdgeAI #chips #Infineon $IFNNY
chippub.substack.com/p/edge-ai-ch...

Industrial edge computing market to hit $134.75B by 2035, fueled by AI & automation

#edgeai #automation

www.soniccomponents.com/gigabyte-ai-...
GIGABYTE’s AI TOP ATOM ATAGB10-9002 packs NVIDIA Grace Blackwell computing into a compact system with 128GB unified memory, 1TB NVMe storage. #AIHardware #EdgeAI #NVIDIA #MachineLearning @nvidia.bsky.social @gigabyte.bsky.social #soniccomponents

New paper introduces Model as a Library, an offline architecture that packages community-enrolled speech models as on-device dependencies for African languages, avoiding generative hallucination by drawing vocab from…

#AI #NLP #LowResourceLanguages #EdgeAI
https://arxiv.org/abs/2609.38574

🔥 APPLY FOR **FREE HARDWARE** BUILD SOMETHING THAT MATTERS.🔥 
10 days left to submit an idea on how #EdgeAI helps communities prepare for wildfires, floods, and infrastructure failures  😍 50 semi-finalists receive an Arduino UNO™ Q #devkit
https://deveco.io/contests/qualcommforamoreresilientamerica

New federated learning approach HO-FL splits each model into a zeroth-order bottom segment and a first-order top segment, letting edge devices pick the boundary based on their memory budget while training one…

#FederatedLearning #EdgeAI #ModelTraining #AIResearch
https://arxiv.org/abs/2609.39074

Can 1-bit AI make smart glasses useful all day? PrismML’s Bonsai is stepping into the real world.

#AI #SmartGlasses #EdgeAI https://spaisee.com/article/prismml-puts-compressed-vision-ai-on-the-smart-glasses-road

AgBench offers an open benchmark for evaluating agentic AI across local, hybrid, and cloud execution on personal devices, highlighting trade-offs in latency, cost, and data exposure. Useful for developers building…

#OpenSourceAI #AIBenchmarks #AgenticAI #EdgeAI
https://arxiv.org/abs/2609.38652

A new arXiv paper, HARISSA, uses a language model's own hidden states to decide which queries deserve extra compute and which are too risky to answer, helping small local models balance efficiency and safety without sending…

#EdgeAI #OnDeviceAI #LLM #LocalInference
https://arxiv.org/abs/2609.38006