AI-RADAR.it
AI-RADAR.IT
News & analysis on local LLMs, stack & on-prem hardware.
Topic / Trend Rising

On-Premise and Sovereign Enterprise AI

Enterprise and privacy-focused deployments are choosing on-premise AI on local GPUs or workstations to keep sensitive data under direct control.

Detected: 2026-10-17 · Updated: 2026-10-17

Related Coverage

2026-10-15 • LocalLLaMA

Microsoft's local 137B-parameter MoE model and the on-premise coding push

Microsoft AI has developed a local version of MAI Code 1.1 Flash, a coding-optimized mixture-of-experts model with 137 billion total and 6.8 billion active parameters. The news points to a push toward self-hosted development tools on Windows, with im...

#Hardware #LLM On-Premise #DevOps
2026-10-15 • LocalLLaMA

AMD W7900 and LFM2.5: A Concrete Signal for On-Premise LLM

A 24-billion-parameter LLM tested on an AMD W7900 with a 128,000-token context and sustained speed around 95 tokens per second signals that on-premise inference is maturing beyond NVIDIA's ecosystem. AI-Radar explores trade-offs, data sovereignty, an...

2026-10-15 • LocalLLaMA

LiquidAI LFM2.5-24-A2B: Robust Performance on AMD W7900 for On-Premise LLMs

A user reported a positive experience with LiquidAI's LFM2.5-24-A2B model (24B parameters), highlighting high performance (~95 tps) on an AMD W7900 GPU. The model, requiring 30GB of VRAM for Q8 quantization with a 128k token context, proves reliable ...

#Hardware #LLM On-Premise #DevOps
2026-10-12 • LocalLLaMA

Intel's Crescent Island Becomes D-Series with Up to 480 GB LPDDR5X

Intel has turned Crescent Island into a commercial lineup: the D-Series. The D270 card combines 32 Xe3P cores, 256 XMX engines, up to 480 GB of LPDDR5X and 1.5 TB/s of bandwidth. Rated performance reaches 1.3 PetaFLOPS FP8, with support from FP4 to F...

#Hardware #LLM On-Premise #Fine-Tuning
2026-10-12 • Phoronix

From Crescent Island to D Series: Intel Targets Enterprise Inference

Intel has announced the GPU D Series, the successor to products known as Crescent Island and aimed at enterprise inference. The announcement lacks VRAM or performance details, but the renaming signals a more direct positioning in the market for on-pr...

#Hardware #LLM On-Premise #Fine-Tuning
2026-10-12 • ServeTheHome

Intel D-Series D270 Targets Inference with FP8 and 1.5 TB/s

Intel has named its Crescent Island data center inference card D-Series and revealed D270 specs: 1.3 PFLOPS dense FP8 and 1.5 TB/s memory bandwidth. A signal for on-prem LLM serving economics.

#Hardware #LLM On-Premise #Fine-Tuning
2026-10-12 • LocalLLaMA

Receipts on-device: the signal for those controlling local inference

A 0.8B model reading receipts on the phone shows on-device inference as a concrete option for personal documents. The project combines fine-tuning on donated data, JSON output and a self-hosted release valve via an OpenAI-compatible server. For those...

2026-10-12 • LocalLLaMA

On-device receipt scanning with a 0.8B vision model

A developer built a free Android app that uses a 0.8-billion-parameter vision model to read receipt photos and extract store, date, total, and line items into JSON. The model, fine-tuned on receipts, runs on the phone without sending data to third pa...

#LLM On-Premise #Fine-Tuning #DevOps
2026-10-12 • LocalLLaMA

OpenMed 3.0: On-Premise Clinical AI Securing Healthcare Data Sovereignty

OpenMed 3.0 is an Apache-2.0 open-source AI toolkit for the medical sector, designed to operate exclusively on-premise. Its core philosophy is to keep patient data on local machines, without any reliance on cloud services or external APIs, ensuring p...

#Hardware #LLM On-Premise #DevOps
2026-10-10 • LocalLLaMA

DigUp brings EmbeddingGemma 2 to the Mac for local multimodal search

DigUp is a free, open-source Mac app that runs EmbeddingGemma 2 locally to search files by content. The 865 MB Q8_0 GGUF model runs on Apple Silicon through llama.cpp and Metal; search loads only the 250 MB text encoder and responds in about a tenth ...

#LLM On-Premise #DevOps
← Back to All Topics