Futures
Access hundreds of perpetual contracts
TradFi
Gold
One platform for global traditional assets
Options
Hot
Trade European-style vanilla options
Unified Account
Maximize your capital efficiency
Demo Trading
Introduction to Futures Trading
Learn the basics of futures trading
Futures Events
Join events to earn rewards
Demo Trading
Use virtual funds to practice risk-free trading
Launch
CandyDrop
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
Launchpad
Be early to the next big token project
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
Spent over 40 yuan to generate only one PPT! Consumers are upset: they previously claimed it was 4–9 copies [Shandong Business Daily · Shan Hai News Daily 3.15 Rights Protection Express]
As AI tools go deeper into the daily work routines of “desk workers,” concerns about computing power consumption have also come to the surface. Recently, Shandong Business Daily’s 3·15 consumer rights protection column [Download Black Cat Complaint client] received a complaint from Jinan resident Ms. Liu, who reported that the TianGong AI platform under Kunlun Wanwei Technology Co., Ltd. allegedly had issues such as “misleading publicity” and “inducing consumption.” The promotion stated that 10,000 points could be used to generate 4–9 PPTs, but in actual operation, only 1 PPT was produced. After the reporter’s involvement, the platform apologized to consumers and refunded the money.
Recalling her use of TianGong AI’s PPT generation feature, Ms. Liu said she spent more than 40 yuan to buy the platform’s monthly membership and received 10,000 points. “At the time, the platform advertised that 10,000 points could generate about 4–9 PPTs. I thought the price was acceptable, so I topped up.” During use, Ms. Liu found that when she only created a 20-page PPT, the platform deducted about 11,000 points. “When I算 it, for over forty yuan, I made one PPT—too expensive.” In Ms. Liu’s view, the platform’s publicity was misleading, causing her to misjudge the real value of the AI she would be able to use with her top-up amount.
Ms. Liu added that the features of TianGong AI also did not meet her requirements. She uploaded a PPT file as a reference for the creation, but the platform’s AI neither successfully parsed that file nor gave users a conspicuous “parsing failed” notice, which resulted in the generated PPT not meeting her use needs and making it unusable.
The reporter noted that on TianGong AI’s top-up page, there indeed is a statement saying “10,000 points can generate about 4–9 PPTs,” accompanied by a notice stating “This is an estimate; the actual result applies.” The page’s top-right corner also shows the user’s available points in real time, where each task’s point consumption can be checked, but it is impossible to obtain more specific rules for point consumption and the deduction standards. When the reporter asked the platform whether the AI “can parse PPTs,” it explicitly replied, “Yes, it can.”
On issues such as publicity not matching the actual situation, details of point deductions, and the user feedback mechanism, the reporter sent an inquiry letter to Kunlun Wanwei. The company did not provide an official written response addressing the related questions raised in the reporter’s interview inquiry. A staff member said that because the company gave consumers a poor experience, it had already processed a refund and issued an apology.
“During the task, each step—such as the large model’s thinking, file parsing, searching, and more—will deduct points. The more complex the task is, the more rounds or the more time it consumes, the more points are needed. This reflects the computing resources consumed in generating the work, such as processing capability, memory, and server time.” After Ms. Liu reviewed the message notifications for TianGong AI’s web version, she found that the platform had previously provided corresponding explanations regarding this point consumption.
Tianyancha shows that Kunlun Wanwei was established in 2008, listed on the Shenzhen Stock Exchange in 2015, and is based in Beijing. Its registered capital is 1.23 billion yuan. The legal representative, chairman, and general manager is Fang Han. The company has a diversified AI business matrix comprising AI intelligent assistants, AI video, AI music and audio, AI games, AI social, and more.
According to public data, from 2022 to 2024, Kunlun Wanwei’s revenue was 4.74B yuan, 4.92B yuan, and 5.66B yuan, respectively; attributable net profits were 1.15B yuan, 1.26B yuan, and -1.6B yuan, respectively. In the first three quarters of 2025, the company’s revenue was 5.8 billion yuan, up 52% year-on-year; attributable net profit was 190 million yuan, turning from loss to profit on a quarter-on-quarter basis. It is worth noting that in recent years, Kunlun Wanwei’s overseas revenue has had a relatively high share and has increased year by year. In 2023, 2024, and the first three quarters of 2025, its overseas revenue share was approximately 86%, 91%, and 93%, respectively.
At the end of January 2026, Kunlun Wanwei released a performance forecast. For 2025, the company expects net profit to be a loss of between 1.35 billion yuan and 1.95 billion yuan. This will be the company’s second consecutive year of huge losses since it went public in 2015.
A massive flow of information and precise analysis—right on the Sina Finance app