(Yicai Global) March 30 -- Kuaishou Technology’s shares jumped after China’s second-biggest short video platform narrowed its net loss by almost 83 percent last year.

After popping by as much as 10 percent in the morning, Kuaishou [HKG: 1024] finished 5.6 percent up at HKD59.40 (USD7.57) in Hong Kong today

The net loss was CNY13.7 billion (USD2 billion) in the 12 months ended Dec. 31, versus CNY78.1 billion the year before, according to the Beijing-based firm’s 2022 earnings report released yesterday. Revenue jumped 16.2 percent to CNY94.2 billion, higher than the average CNY93.1 billion of analyst estimates compiled by Bloomberg.

Kuaishou had 366 million daily active users at the end of December, up 13.3 percent from a year earlier. On the firm’s earnings conference call, Chief Financial Officer Jin Bing expressed confidence that the site will exceed 400 million DAUs in the second half of this year. Monthly active users rose 10.7 percent to 640 million in 2022.

Revenue from online marketing services, Kuaishou’s biggest source of income, rose 15 percent to CNY49 billion, while that from livestreaming jumped 14 percent to CNY35.4 billion. The firm’s e-commerce gross merchandise volume surged 33 percent to CNY901.2 billion, driving the revenue from other services up 31 percent to CNY9.8 billion (USD1.4 billion).

“We are in the middle of signing full-year framework agreements with advertisers, who are cautiously optimistic about 2023,” co-founder and Chief Executive Cheng Yixiao said on the call.

“Given the pace of economic recovery and the lag in the advertising market versus the overall economy, we expect a noticeable rebound in the online marketing services market in the second half of the year,” Cheng said.

In the fourth quarter, Kuaishou narrowed its net loss 75 percent to CNY1.5 billion from a year earlier. Revenue rose 16 percent to CNY28.3 billion.

Speaking on the topic of artificial intelligence, Cheng said that by leveraging its AI team over the years, Kuaishou accumulated a deep technological foundation for generative AI, especially large language models, LLMs and multimode pre-trained large models, AITC tools, as well as the applications for these technologies.

Editor: Futura Costaglione