COWOS CHIPS1h ago

TSMC expands CoWoS capacity by another 40%, easing the AI packaging bottleneck

12.4K views LLAMA 5 AI MODELS3h ago

Inside Meta's 600k-GPU training cluster: a mixed H100 / B300 fleet

9.6K views INTERVIEWYouTube24:18 PEOPLE4h ago

Liang Wenfeng interview: the three calls that halved training costs again

31.5K views SKILL LAB SKILLS5h ago

We tested 5 coding skills — only two truly add 40% productivity

10.4K views HBM3E SUPPLY CHAIN6h ago

Samsung re-enters NVIDIA's approved vendor list for HBM3e

8.2K views RELAY AI MODELS7h ago

Relay vs direct API: how much latency does a model relay really add?

9.8K views RTX4090 MARKET8h ago

Used RTX 4090 prices drop 12% as consumer and data-center cards decouple

7.5K views EU-AI POLICY9h ago

EU AI Act compute disclosure kicks in: training runs above 10^26 FLOPs must register

6.9K views GAUDI 4 CHIPS11h ago

Intel's Gaudi 4 lands a major edge-inference deal, targeting 500k units

6.1K views JENSEN PEOPLE13h ago

Jensen Huang: every nation needs sovereign AI infrastructure

8.6K views DC·CN DATA CENTER14h ago

China AI data center survey: top operators at 85% utilization, the long tail below 30%

5.8K views DEEPSEEK AI MODELS16h ago

DeepSeek cuts training costs by another 40% on algorithmic efficiency

5.2K views OPEN-LLM AI MODELS18h ago

Have open models caught GPT-5o? A three-axis comparison

4.9K views SiC SUPPLY CHAIN20h ago

Wide-bandgap power shortage leaves data center PSU retrofits queued 26 weeks

4.8K views 800G CHIPS22h ago

Cisco buys AI networking silicon startup for $24B, heating up the DPU race

4.3K views SMR DATA CENTER1d ago

SMRs head for data centers: OKLO signs two cloud providers

3.9K views