Express News | TENCENT: On March 28, approximately 0.501 billion Hong Kong dollars was spent to repurchase 0.982 million shares, with a repurchase price per share of 503.5 to 515.5 Hong Kong dollars.

BreakingsMar 28 17:30

The Hang Seng Index Company launched 2 major SZSE NEW INDEXes! A picture reveals which Technology companies made the selection.

The Hang Seng Index company announced two indices related to Technology, namely the Hang Seng Technology Top 10 Index and the Hang Seng China Technology Top 10 Select Index.

Futu NewsMar 28 15:40

Daily Bull and Bear | The Hang Seng Index night futures are high at 197 points, and there is a net Inflow of 27 million HKD into short positions; Medical stock WEIGAO GROUP soared over 15% after performance, with one warrant making a profit of 2.3 times.

As of yesterday's market close, the total market turnover was 239.641 billion HKD, with a total turnover of 21.832 billion HKD for all warrants and callable bull/bear contracts, accounting for 9.1% of the market turnover, of which good positions accounted for 5.3% and short positions accounted for 3.8%.

Futu NewsMar 28 08:45

The "seven giants" are outdated, and Wall Street's attention has shifted to the "ten heroes" of China Technology.

After DeepSeek was launched, China welcomed its own OpenAI moment. Although China's Technology stocks have already risen significantly, the road to revival may have just begun.

Barron ChineseMar 27 21:48

The top ten star Network Technology stocks in Hong Kong soared by over 14% on the first day after earnings. Which companies reported joyful results and which faced disappointment in Q4?

Among the top ten Network Technology stocks, seven companies experienced relatively small price fluctuations, all within 5%, while Alibaba and Tencent Music both soared over 14%, and Trip.com Group plummeted nearly 12%.

Futu NewsMar 27 20:09

Details of the Ant Group's domestically produced GPU training large model have been revealed! The lead researcher of the Ling model shared an article detailing the story behind it.

When using domestic acceleration cards to train large models, the Ling team faces two major challenges: the alignment of training correctness and stability. They achieved the goal of keeping the long-term training loss difference below 0.1% through basic operator alignment and mechanisms for handling spikes in loss and gradients, and validated the reliability of small model training results to predict the performance of large model training.

Hard AIMar 27 18:44