Tencent unveiled its progress in chips for the first time this week, which immediately gave its stock price a modest boost. Silicon seems distant from the giant's main arenas of video games and social networks, so observers suggest that Tencent's move is to signal that it's aligned with China's long-run goals to self-develop semiconductors at a time its gaming unit is under a slate of regulatory assaults. Other major tech powerhouses, from Alibaba, to Baidu, to Huawei, have all answered Beijing's silicon push with their in-house chips.
On the other hand, for a company with as much data to handle as Tencent, one should wonder why it didn't get in on semiconductors earlier.
The three chips Tencent unveiled on Wednesday are all custom built: one for AI inference, one for video transcoding, and another for network interface.
There have been ample examples where internet behemoths start building their own specialized hardware to buttress their existing businesses. Back in 2018, Facebook was already hiring AI chip designers, who could potentially help tackle misinformation and process its gargantuan amount of user data.
Tencent, with over one billion monthly users on its flagship WeChat messenger, surely has a lot of digital footprints to handle.
But WeChat's steward Allen Zhang is famously reserved about exploiting personal data. Till now, WeChat's user feed is still kept chronological with native ads that sporadically pop up.
"If we analyze [users’ chat history], we can bring great advertising revenue to the company. But we don’t do that, so WeChat cares a lot about user privacy,” Zhang said at WeChat's annual conference last year. He wants WeChat to be a useful, disposable tool rather than an app that soaks up people's time with addictive, algorithmic recommendations.
But Zhang seems to have made concessions. WeChat now includes a short video section with the barebone functions of TikTok. Like TikTok, WeChat's video feature infers what users like and serves up content accordingly.
Tencent has a slew of other business lines that could benefit from greater machine learning power and are perhaps more eager to monetize: Tencent News, its news aggregator, and Tencent Video, its Netflix equivalent. Under China's strict censorship environment, content providers may also find themselves in need of more computing power to root out text, audio and videos that violate government rules.
Dowson Tong, Tencent's senior vice president, explained that the firm's AI inference chip will be used for image and video processing, natural language processing and search, among other scenarios. The chip for video transcoding, as the name implies, will ensure smoothness and low latency as Tencent handles its massive video workload. Lastly, the smart network interface card (SmartNIC) will be used for CPU server offloads, according to Tong.
Tencent won't be working on chips alone. Tong noted in his speech that the company will create an ecosystem to maintain "deep, strategic collaborations with chip domestic and overseas chip companies." For instance, Tencent has backed Shanghai-based Enflame, which specializes in AI training chips, in four funding rounds and is already applying the startup's tech to its business.