At a time when the clock is ticking for TikTok to arrive at a forced sale of its US operation to circumvent a sanction, China introduced a new export rule with an updated list of technologies prohibited from exporting in a move that would only further complicate the ongoing sales discussion between TikTok and its motley group of pursuers.
China’s ministry of commerce on Friday passed a new regulation that restricts home-grown companies to export or sell their artificial intelligence technology to a non-Chinese company. This directly affects TikTok’s sale discussions and negotiations as its AI-powered personalized recommendation tool is the heart of the company without which it loses its core value.
“TikTok is famous because of its personalized content and if that goes away it’s as good as any other app. If the algorithm is off the table then all that the buyer gets is the brand, which would make negotiations really difficult for TikTok,” Sanjeev Kumar, forecast analyst at Forrester, told KrASIA.
As it negotiates deals with companies that have made bids to buy its US business, TikTok is also said to have approached India’s largest conglomerate Reliance Industries to buy its India unit, whose valuation according to reports is pegged at USD 3 billion. Japanese conglomerate SoftBank is also orchestrating discussions with different Indian companies to help TikTok find a buyer in India. But with current changes, the deal might take longer to finalize.
Investors and analysts KrASIA spoke to said TikTok’s recommendation algorithm is its secret sauce and this is what TikTok’s prospective buyers are after.
Its recommendation engine which ensures every user gets a personalized feed is what makes the app so addictive, increasing the average time spend. According to App Annie, Indians spent 5.5 billion hours on TikTok in 2019, a six-fold jump compared to 2018.
Compared to TikTok, Indian short-video apps that mushroomed after the company was banned in June, don’t come close to TikTok’s robust recommendation engine. To create an effective AI-backed recommendation tool requires the app to have a huge user base and continuous generation of new content. The more data a company has the better predictions its AI can make.
“Indian short-video companies are very young as of now compared to TikTok’s presence in India. They will take a few years before they can match TikTok’s capabilities,” Kumar said.
However, Sumit Ghosh, founder of Chingari, an Indian short-video app, said the company has over 3.5 million daily active users which generate “pretty good data” for its AI to learn users’ interest.
“TikTok had way more data than us, but currently we also have decent amount of data to build the prediction model. Our users watch about 100-150 videos monthly on the platform which generates a lot of data. It will take another five to six months to perfect the model like the one TikTok has,” Ghosh said.
While TikTok’s sale discussions in India are still at a preliminary stage, a CNBC report said that, until last week offers from Oracle and a joint bid from Microsoft and Walmart were the top contenders but the change in regulation from China has brought in added complications to the deal. The report said that if TikTok’s core technology would not be a part of the deal, the sale value might drop significantly from its earlier range of USD 20 to 30 billion.