Credit：Visual China

“Right now, we actually have one of the largest inventories of AI chips in China among all the players. And one of the key things that we have done was actually we were the first to put in order for H800 and that allow us to have a pretty good inventory of H 800 chips. So we have enough chips to continue our development of Hunyuan for at least a couple more generations”, Tencent President Martin Lau told analysts at an earnings call following the third quarter report. Lau said the new U.S. export control “does not really affect the development Hunyuan and our AI capability in the near future”.

However, Lau admitted the export control did have the potential impact on Tencent’s cloud computing business. “Going forward, we feel that the chip ban does actually affect our ability to resell (use of) these AI chips through our cloud services, so that's one area that may be impacted, and going forward, we will have to figure out ways to make our the usage of our AI chips more efficient,” Lau said. Tencent will try to see whether it can offload a lot of the inference capability to lower-performance AI chip, so that it can retain the majority of its high-performance AI chips for training, and will try to find the domestic source for these training chips, according to Lau.

The United States introduced on October 18 a rule about additional export controls, with a aim to limit China’s access to advanced semiconductors that could fuel breakthroughs in AI and sophisticated computers. The rule was supposed to come into effect following a 30-day public comment period. But AI chip giant Nvidia disclosed on October 25 that the U.S. government informed the licensing requirements, which is applicable to products having a “total processing performance” of 4800 or more and designed or marketed for datacenters, was effective immediately. The Wall Street Journal reported late October that Nvidia could be forced to cancel next year orders from Chinese buyers, suggesting the company may face a hit of at least $5 billion as orders from major Chinese companies next year exceeded $5 billion.

Hunyuan is Tencent’s proprietary large language model (LLM) that debuted and opened to enterprises in China for tests and building apps on Tencent Cloud on September 7. The foundation model empowers enterprises to build powerful tools, in addition to training their own unique large models derived from Tencent’s Model-as-a-Service (MaaS) offering, creating creates a virtuous cycle in which enterprises refine their large models with Hunyuan to create uniquely intelligent services across their operations. Hunyuan has also been connected to 50 of Tencent’s own products, including Tencent Cloud, Tencent Marketing Solutions, Tencent Games, Tencent fintech services, Tencent Meeting, Tencent Docs, Weixin Search, QQ Browser, and other core offerings.

At the financial report, Tencent said it is upgrading Hunyuan and deploying it for in-house products, as well as making it available to enterprise customers. “We are increasing investment in our AI models, providing new features to our products and enhancing our targeting capabilities for both content and advertising,” said Tencent Chairman and CEO Pony Ma, or Ma Huateng.

Advertising is Tencent’s division with the fastest growth in the third quarter. Revenues from online advertising increased 20% year-over year to a record high RMB25.7 billion, propelled by robust advertising demand for Video Accounts, mobile ad network, and Weixin Search. Martin Lau said Tencent is continuously increasing the size of Hunyuan and preparing for the next generation of the model, which is set to be a mixture of exports architecture. He added that the Shenzhen-based company expects Hunyuan can provide a lot of tools for business customers and further improve advertising business efficiency, really merging the advertising stage and the selling stage.  

763人已贊賞 >

敬原創(chuàng)，有鈦度，得贊賞

贊賞支持

0

0

0

掃描下載App