# Encharge : EN100

# EN100
Putting the most advanced AI where you are
## Meet EN100
The game-changing AI accelerator that delivers 200 TOPS of AI compute power in an ultra-efficient chip. Engineered to bring advanced AI capabilities to laptops, workstations, and other compute appliances, EN100 enables a new era of personalized AI at the edge.

# BENEFITS

40 TOPS/Watt

From Impossible to Indispensable
- The holy grail of AI compute: robust, scalable, and precise analog in-memory computing
- Flexible dataflow architecture with FPU, L2/L3 & LPDDR, embedded RiscV core
- NPU cores delivering up to 40 TOPS/W 8-bit compute efficiency
- Programmability and scalability from low-power to PetaOPS class AI acceleration
The EN100 AI accelerator can provide an unmatched 200+ TOPS of AI computing capacity within constraints of edge and client platforms such as laptops, significantly outperforming other NPUs, both digital and analog, across classical (CNN) and transformer (LLM) model families.

# PRODUCTS

### M.2 for Laptops
Experience unparalleled performance for personalized AI on your laptop without compromising battery life or portability.
- Peak TOPS: 200+ TOPS
- Card Power: Up to 8.25W
- CMOS Node: 16nm
- Memory: 32GB On-Board Memory (up to 68 GB/s bandwidth)
- CPU Compute: Multi-threaded RISC-V for control orchestration
- Format: M.2 2280 Card

### PCIe for Workstations
Bring data center AI performance to your desktops, workstations, and on-prem servers with GPU-level compute capacity at a fraction of the cost and power consumption. Ideal for professional AI applications in secure and sovereign multi-user environments.
- Peak TOPS: ~1 PetaOPS
- PCIe Card Power: Up to 40W
- CMOS Node: 16nm
- Memory: 128GB On-Board Memory (cumulative bandwidth of 272 GB/s)
- CPU Compute: Multi-threaded RISC-V for control orchestration
- Format: PCIe HHHL Card

# SOFTWARE

Software For Accelerating the AI of Tomorrow
EN100 allows developers to harness the transformative speed and efficiency of analog in-memory processing with seamless integration into today’s existing platforms. Developers can rapidly deploy end-to-end AI workloads on EN100 with a complete suite of software tools co-designed with EN100 to maximize efficiency, performance, utilization, and fidelity.

## ​​EnCharge Quantization Tools
Build or adapt new AI models for EnCharge silicon seamlessly, supporting industry-standard 8-bit and 4-bit quantizations within PyTorch, Keras, JAX
## EnCharge Compiler Stack
Compile AI models on EN100 NPU, creating efficient & performant end-to-end execution of models
## EnCharge Runtime Stack
Connect deployment workflows to EnCharge’s low-level runtime, device-drivers & firmware for each inference request
## EN100 Toolkit
Model-zoo support and optimized SDKs to deliver amazing AI experiences, Performance analyzer tools (debuggers, simulators), EnCharge FAE support for product integration

Mentions: AI Frameworks, Huggingface, JAX, ONNX, Keras, Pytorch, TensorFlow, Inference Frameworks, llama.cpp, Ollama.

## Unprecedented Efficiency
EN100 delivers transformative compute efficiency by leveraging analog in-memory computing—a breakthrough in AI hardware. This technology enables robust, scalable, and flexible AI compute with precision, delivering orders-of-magnitude improvements in performance and energy efficiency across a range of compute appliances compared to traditional digital NPUs.
## Maximum Programmability
Unlike specialized chips that excel only with specific models, EN100 is designed for versatility to ensure exceptional performance across a wide range of AI workloads, today and in the future. In fact, EN100 massively outperforms the next-best competitor across image recognition, vision, and LLMs/SLMs while being efficient enough to run in battery-powered platforms such as laptops.
## Optimized for On-Device Language and Reasoning Models
EN100 form factor cards are tailored for sophisticated language and reasoning models. They enable on-device deployment of an advanced class of AI previously exclusive to data centers, delivering personalized, secure responses faster than human reading while breaking through the latency of cloud-based solutions.

# A New Frontier of On-Device AI
Transformative efficiency can bring AI on-device, removing cost, privacy, and capability barriers to meaningful AI experiences:

## Creativity
Content generation and editing powered by multimodal AI that can work across text, audio, images, and video will empower creators to brainstorm, iterate, and refine their work.
## Productivity
Always-on, multimodal generative AI assistants with secure, private access to your data and preferences will help summarize video meetings and reports, create documents and spreadsheets, write code, provide real-time translations, and manage your schedule.
## Immersion
Boosted framerates, highly responsive NPCs, and procedural content generation powered by low-latency on-device AI will create a new level of gaming immersion.

# Early Access Program
## Round 1 of EN100’s Early Access Program is currently full.
Please sign up to learn more about the EN100 Round 2 Early Access Program eligibility requirements and application process when available.

© 2025 EnCharge AI. All rights reserved. Privacy Policy