DEEPX's new AI chip DX-M2, built on a 2nm low-power process, is expected to be released in early 2027
Automated summary · Based on collected summaries; not individually verified
Prepared by HotAI
What happened
Powered by Samsung's 2nm foundry technology, the chip delivers up to 80TOPs of NPU performance, supports LPDDR5X-9600 memory at up to 128-bit width, and can run local LLM and VLM models with over 10 billion parameters while consuming only 5W
Background
What is a vision-language model? Image questions, reading, and spatial limits
VLMs connect vision and language for tasks such as image questions and document understanding. They are a multimodal category, not a guarantee of image generation or every visual capability. Small text, counting, and spatial judgments depend on conditions. Verify sources and preserve uncertainty rather than invent details from common knowledge.
Reading notes
The event summary comes from collected excerpts. Background passages quote existing HotAI articles to explain terms and do not establish new facts about this event. Check the original sources for product results, conditions and subsequent updates.
Original source
Sources and collection timeline
| Source | Report title | Original publication time | Collected by HotAI |
|---|---|---|---|
| IT Home | DEEPX's new AI chip DX-M2, built on a 2nm low-power process, is expected to be released in early 2027 | 2026-11-11 14:20 | 2026-11-11 14:32 |