INTERNET: Baidu’s AI Obsession Spawns New Silicon Valley Center
Bottom line: Baidu’s opening of a new artificial intelligence lab in Silicon Valley is the latest move in its AI obsession, which is likely to end in failure and a quiet pullback in around two years due to mediocre execution.
I’m officially dubbing Baidu (Nasdaq: BIDU) chief executive Robin Li the “two-year attention span man”, with word that the company is setting up a new Silicon Valley office in the rush to build up its artificial intelligence (AI) capabilities. If any one of China’s “big three” Internet chiefs deserves the title of “short attention span” man, it’s most definitely Robin Li. As far as I can remember, Li has been a “flavor of the moment” guy who fixates on the latest daily hot trend, most often to jettison the idea around 2 years later when it falls out of fashion.
That’s not to say that Li is always wrong, or even that his tendency to dive into the hot trend of the moment is bad. The one notable case where his bet was on target was with the mobile Internet, where Baidu is now dominant in mobile search that is the main way people in China surf the web. But more often than not, Baidu throws tens or even hundreds of millions of dollars at the latest trends, only to abandon them a couple of years later when the company either fails to gain traction or the business emerges as money-devouring black hole.
All that said, let’s zoom in on the latest flavor of the day that Baidu is chasing, namely AI. According to the latest reports, Baidu, which previously trumpeted its formation of an AI center in Beijing, has announced plans to open a new Silicon Valley center devoted to the cause. (English article) The newest office is actually Baidu’s second in Silicon Valley, and will house about 150 people. That will roughly double its current headcount in Sunnyvale, since its original R&D facility is now home to about 200 people.
The two new sites will be led by Zhang Yaqin, a former Microsoft (Nasdaq: MSFT) executive that Baidu poached in 2014. The news comes less than a week after a top Baidu R&D executive, Andrew Ng, announced he was resigning from the company for unspecified reasons. Earlier this year, Baidu also hired Qi Lu, another former Microsoft executive with an AI background, as its new president and COO. Robin Li has also been quoted recently saying how difficult it is for Chinese companies to hire good AI talent, saying they must often pay 15 percent more than western firms would pay for comparable executives.
Anyone sensing that Robin Li is suddenly a man on a mission is correct here, in a behavioral pattern that’s quite typical for him. This particular AI initiative also dovetails a bit with one of Li’s other recent attractions for self-driving vehicles, which obviously also has a big AI component.
Lost Loves
All of this makes you wonder what happened to Li’s other loves from the last few years, namely with online-to-offline (O2O) and other web-based services. That love included quite a few initiatives, such as Baidu’s Nuomi group buying site, its Qunar online travel service, its takeout dining service and its its iQiyi online video service.
Company watchers will know it has essentially spun off Qunar to former chief rival Ctrip (Nasdaq: CTRP). It was also trying to spin off iQiyi to a separate stand-alone company, possibly for an IPO, thought that plan has stalled at the moment. Other rumors have come that it’s looking to reorganize or possibly dump Nuomi, and I wouldn’t be surprised if we start hearing similar rumors soon for the money-devouring takeout dining service.
Such is the temperament of Robin Li. A few years ago his interests were with e-commerce, but he later abandoned the company’s in-house effort, and several years dumped its joint venture with Japan’s Rakuten (Tokyo: 4755). It also took a stab at social networking, but that didn’t go very far either. As I’ve said above, the company’s one big bet that did pay off was in mobile Internet, though even there the move was mostly a no-brainer because it simply involved developing a mobile-friendly version for its core search business.
All that brings us back to the central theme of this post, namely whether the company’s latest move into AI is simply a new flavor of the day that Baidu will abandon around two years from now. The company certainly appears to be quite serious with these two AI labs in the US, as well as its driverless car initiative.
But as with many things, Baidu seems to have some issues with execution, and I do sense that it will perform equally mediocrely in both of these endeavors, ultimately ceding the ground to others who do a better job. Accordingly, I would give Baidu’s recent AI drive just a 30 percent chance of success, and argue we’re likely to see the company quietly jettison the efforts around 2 years from now when it fails to emerge as a leader in the space.
