Dark | Light
[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

![AsianFinPress Avatar](https://lunarcrush.com/gi/w:24/cr:twitter::1692064986028507136.png) AsianFin 亞財社 [@AsianFinPress](/creator/twitter/AsianFinPress) on x 1659 followers
Created: 2025-07-25 22:37:01 UTC

🧠💥 Intellifusion Doubles Down on Inference AI Chips, Eyes 2026 Launch of Next-Gen Architecture



China’s Intellifusion is going all-in on AI inference, unveiling its latest X6000 Mesh accelerator (256 TOPS) and servers offering up to X PFLOPS, just ahead of the 2025 World Artificial Intelligence Conference.

🚀 CEO Chen Ning says 2025 is the tipping point:

“Inference will outpace training in growth and application.”

🔧 What’s New:

Self-developed 4th-gen NPU

Fully domestic D2D chiplet + mesh architecture

Real-time decoding of XXX video streams

Edge & data center servers (Shenmu 6203, Tianzhou 6408/680G)

📈 By the Numbers:

2024 revenue: ¥900M (↑81.3% YoY)

Q1 2025 revenue: ¥264M (↑168.2% YoY)

3-year compute deal worth ¥1.6B underway

📱 On the consumer side:

Qiancheng AI tech adopted by Huawei, OPPO, Honor

“Dr. Luka” hardware sees rapid sales

Projected 50%+ consumer growth in H1 2025

🔮 Coming 2026:
“Computing Power Building Blocks 2.0” —
featuring Nova500 NPU (FP8/FP4 native), 3D hybrid-bonded memory, NB-Mesh interconnect, and NB-Link CPU-NPU shared memory.
🗣️ “Inference is China’s shot at reshaping the Fourth Industrial Revolution,” says Chen.
“Invention means little without scale — that’s where we win.”

#AIChips #InferenceAI #ChinaTech #Intellifusion #Semiconductors #AIAcceleration #EdgeAI #HPC #AIinfrastructure

![](https://pbs.twimg.com/card_img/1948875166798544899/XUEgxZXe?format=jpg&name=800x419)

XX engagements

![Engagements Line Chart](https://lunarcrush.com/gi/w:600/p:tweet::1948875162184810964/c:line.svg)

**Related Topics**
[tops](/topic/tops)
[coins ai](/topic/coins-ai)
[inference](/topic/inference)
[$ai4](/topic/$ai4)

[Post Link](https://x.com/AsianFinPress/status/1948875162184810964)

[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

AsianFinPress Avatar AsianFin 亞財社 @AsianFinPress on x 1659 followers Created: 2025-07-25 22:37:01 UTC

🧠💥 Intellifusion Doubles Down on Inference AI Chips, Eyes 2026 Launch of Next-Gen Architecture

China’s Intellifusion is going all-in on AI inference, unveiling its latest X6000 Mesh accelerator (256 TOPS) and servers offering up to X PFLOPS, just ahead of the 2025 World Artificial Intelligence Conference.

🚀 CEO Chen Ning says 2025 is the tipping point:

“Inference will outpace training in growth and application.”

🔧 What’s New:

Self-developed 4th-gen NPU

Fully domestic D2D chiplet + mesh architecture

Real-time decoding of XXX video streams

Edge & data center servers (Shenmu 6203, Tianzhou 6408/680G)

📈 By the Numbers:

2024 revenue: ¥900M (↑81.3% YoY)

Q1 2025 revenue: ¥264M (↑168.2% YoY)

3-year compute deal worth ¥1.6B underway

📱 On the consumer side:

Qiancheng AI tech adopted by Huawei, OPPO, Honor

“Dr. Luka” hardware sees rapid sales

Projected 50%+ consumer growth in H1 2025

🔮 Coming 2026: “Computing Power Building Blocks 2.0” — featuring Nova500 NPU (FP8/FP4 native), 3D hybrid-bonded memory, NB-Mesh interconnect, and NB-Link CPU-NPU shared memory. 🗣️ “Inference is China’s shot at reshaping the Fourth Industrial Revolution,” says Chen. “Invention means little without scale — that’s where we win.”

#AIChips #InferenceAI #ChinaTech #Intellifusion #Semiconductors #AIAcceleration #EdgeAI #HPC #AIinfrastructure

XX engagements

Engagements Line Chart

Related Topics tops coins ai inference $ai4

Post Link

post/tweet::1948875162184810964
/post/tweet::1948875162184810964