Social trading apps have fueled the current on-chain market rally, with weekly trading volume approaching $1.3 billion.

According to GMGN market data, the meme coin MOO, paired with Micron stock on Long.xyz, has exceeded $18 million in market capitalization, currently trading at $18.27 million, with a daily gain of over 330%. BlockBeats Note: Stock Meme is an emerging concept merging traditional meme coins with tokenized U.S. equities: instead of pairing with USDT or ETH, meme coins are directly paired with on-chain U.S. stock tokens (e.g., NVDA, TSLA, AAPL, etc.). This model preserves meme coins’ high volatility and community-driven speculative traits, while tapping into the popularity and narratives of real stocks. A portion of transaction fees is often redirected to the community treasury to accumulate the corresponding U.S. stock tokens, forming a dual-driven model of "sentiment-fueled speculation + real asset anchoring".

Crypto analyst Adam’s data shows that the 24-hour trading volume of decentralized exchanges (DEX) on Robinhood Chain has topped $900 million for the first time. The meme token launchpad recorded a record-high trading volume of $438 million yesterday, while real-world asset (RWA) trading volume also exceeded $200 million for the first time.

According to Bitget market data, WTI crude oil has climbed above $86 per barrel, with an intraday increase of 0.73%. Brent crude oil touched $91 per barrel, up 0.75% on the day.

Fomo co-founder @seyong announced on X that the platform has seen rapid growth recently, adding 30 new users every minute. Additional data shows that in the week ending August 24, the weekly trading volume of the social trading app neared $1.3 billion. The daily active user counts of Fomo and Pump are close to those of Polymarket, Hyperliquid, and Phantom, ranging from 60,000 to 100,000 respectively.

Beating AI News: Just after the Hy4 preview went open-source, Tencent has released an extreme quantized version of its Hunyuan model. The original model weights are nearly 1.5TB, while the new GGUF version is only around 214GB, drastically lowering the local deployment barrier for this 770B MoE model. Tencent did not uniformly quantize the entire model to 1.25-bit; instead, it applied different quantization levels based on each layer’s sensitivity to precision: non-critical layers are compressed to as low as ~1.31-bit, while sensitive layers retain 2-bit or higher precision, resulting in an average of ~2.38 bits per weight (bpw). In four benchmarks provided by Tencent, the quantized version only dropped 0.2 to 1.6 points compared to the BF16 original. After compression, Tencent also tested heterogeneous device joint inference with prima.cpp. A setup consisting of an RTX 4090 laptop and a 4-A4000 server, with only 80GB of total VRAM and 64GB of RAM, achieved an inference speed of 1.02 tokens per second—roughly 6 times faster than running the model offloaded on the laptop alone. Multiple devices with different configurations can also jointly share the model inference workload.

Japan’s 10-year government bond yield climbed to the 3% level for the first time since September 1996.