Can Chinese AI Chip Makers Compete with Nvidia?
There is a new arms race, but we’re not talking thermonuclear war—unless we give machines the launch codes. Artificial intelligence is one of the key technologies that we cover, and it’s been a wild ride the last few years. Industries from healthcare to recruitment have embraced AI to gain efficiencies and a competitive edge. Heck, even Coca Cola is giving you Coke with AI. While the software can be sexy, it’s the hardware, or computing power, which has made many of the advancements in AI possible. One of the companies leading the charge is Nvidia (NASDAQ:NVDA), which has become the gold standard for big computing applications ranging from gaming to supercomputers to neural networks. More than a few AI chip startups have emerged in the last year or two, but the real competition will likely come from established players like AMD or Google. And then there’s China.
We recently told you about all the ways China is Kicking America’s Ass in Tech. We didn’t mention AI because China isn’t there. Yet. However, the Chinese government has plans to reach parity with the United States in AI as early as 2020. It’s putting up the money to do it, starting with a $2 billion AI business park that will be home to 400 enterprises. The country hopes to generate about $60 billion from AI technology by 2025.
Despite the official line, Chinese companies are still going with Nvidia, which signed a deal last year to provide its new Volta GPU chips into data centers run by three of China’s biggest tech companies—Alibaba, Baidu and Tencent. It’s little wonder that Nvidia’s stock gained more than 100 percent last year and is off to a scorching start in 2018. Wired reported that Nvidia is in the Chinese government’s crosshairs, urging its industry to develop a chip that is 20 times more powerful and energy efficient than Nvidia’s M40 chip used for artificial neural networks.
A Cambrian Explosion
That goal landed Beijing-based Cambricon Technologies $100 million in funding last August. Alibaba and Lenovo participated in the Series A, which was led by the Chinese government’s largest state-owned investment holding company. That investment propelled Cambricon, founded only in 2016, into the Unicorn Club of companies valued at $1 billion or more. Cambricon hopes to put its AI hardware into one billion smart devices and corner as much as 30 percent of China AI chip market in three years, China Money Market reported. The company recently appeared on CB Insights AI 100 startups list.
Last year, the company released a ton of new products, including three AI processors that can be used in all sorts of applications, from computer vision to autonomous driving to natural language generation. Cambricon also produced a couple of high performance machine learning chips for servers, one market where China lags behind despite being home to more supercomputers than the United States (many of which sport Nvidia hardware). That’s not surprising, as AI Chinese startups like Cambricon have been mainly focused on chips for mobile devices and wearables. In 2016, for example, it made $15 million in licensing fees for its Cambricon-1A chip from smartphone manufacturers and wearable device makers.
A Really, Really Smart Phone
In fact, it makes sense that much of the Chinese AI chip market has been focused on mobile applications, given the country’s emphasis on mobile technology, from social media and e-commerce on WeChat to eSports gaming. Case in point: Semiconductor manufacturer HiSilicon, owned by Chinese telecommunications giant Huawei Technologies, released a wicked fast Kirin 970 processor that features a traditional CPU, a Nvidia-like GPU and an NPU, for neural processing unit, which handles the AI workload. The NPU particularly excels at image recognition, processing a reported 2,000 images per minute. Incidentally, Huawei is the world’s No.3 smartphone manufacturer, right behind Apple and Samsung. Based on the results of that test, we’re not sure how much longer those rankings will hold.
A Marriage of Hardware and Software
DeePhi Tech is another relatively new entrant into the Chinese AI chip scene. Founded in 2016, the Beijing startup raised $40 million in a Series A last October, following an undisclosed round earlier that year. The Series A-plus round included Ant Financial (affiliated with Alibaba) and Samsung, among others. The company rose out of an academic collaboration between Tsinghua University and Stanford University. DeePhi offers what it calls its Deep Neural Network Development Kit, DNNDK, a deep learning software kit that pairs with its DPU (Deep Learning Processor Unit) hardware platform. The DPU platform is based on a field-programmable gate array (FPGA), a type of integrated circuit that the company claims can achieve an order of magnitude higher energy efficiency over GPUs on image recognition and speech detection.
Samsung is reportedly interested in the company’s AI chip technology for portable devices (based on the above test we saw, they need all the help they can get). Last year, DeePhi released a series of AI products, such as Aristotle on ZU9 FPGA Intelligent Analytic Servers, Aristotle CNN Accelerator, and AI solutions in face, speech and video analytics, according to China Daily. Targeted applications include surveillance (20 million AI-equipped security cameras in China and counting), robots and data centers.
AI on the Move
We first introduced you to another Beijing AI hardware developer called Horizon Robotics last May. Since then, the company, founded way back in 2015, has taken in $100 million in a disclosed Series A led by Intel. Its brain processing unit (BPU) is at the heart of its AI tech platform. The infusion of cash will help Horizon develop its Hugo platform for autonomous vehicles, Bloomberg reported. The company says it hopes to have its product into cars on China’s roadways by next year:
The company was expected to showcase two of its embedded AI visual chips at the Consumer Electronics Show this week in Las Vegas. The Journey 1.0 processor is designed for “intelligent” driving vehicles, while the Sunrise 1.0 processor is meant to be used in smart camera systems.
A Bit Player
A company called Bitmain out of Beijing, which specializes in the hardware used in mining bitcoins, decided it wasn’t a big leap from developing the specialized chips for trading cryptocurrencies to designing AI chips. Founded in 2013, the company took in $50 million last year from top VC firms Sequoia Capital and IDG Capital Partners. Its new AI chip, the Sophon BM1680, is specialized for both training and executing deep-learning algorithms. It’s reportedly similar to Google’s own Tensor Processing Unit AI chip, and it can be used in a variety of applications, including image and speech recognition, autonomous vehicle technology, camera surveillance, robotics and the Internet of Things. The move into developing AI hardware may be more than just opening a new revenue stream for Bitmain. Beijing has taken a dim view of cryptocurrencies and blockchain, with the government banning initial coin offerings and closing down cryptocurrency exchanges.
Many Cores is Better
Founded just last year, Shanghai-based ThinkForce Electronic Technology raised about $68 million in December, which included backing from Sequoia, among others. The company claims its team was put together with talent from chip firms likes IBM, Intel and AMD, according to China Money Network. CMN also reported that ThinkForce plans to launch an AI chip with many core processors, which are specialist multi-core processors designed for a high degree of parallel processing. The world’s fastest supercomputer, China’s Sunway TaihuLight, employs these sorts of processors. The company claims its technology is more than five times more efficient in power consumption and cost savings compared to comparable hardware from Nvidia.
Conclusion on Chinese AI Chip Makers
China is obviously serious about its AI ambitions. But 2020 seems optimistic even by its standards, considering the dearth of domestic, high-end expertise in artificial intelligence, as more than one pundit has pointed out. AI hardware also seems to be largely a nascent industry in China, and it’s hard to see any of these companies seriously contending with Nvidia anytime soon, though certainly they are poised to make serious inroads into the mobile AI market. While it’s easy to frame this competition along the lines of China versus the USA, the reality is that the global market doesn’t recognize national boundaries (except where governments put up regulatory roadblocks or financial incentives to protect technological advantages). Chinese supercomputers will likely continue to use Nvidia hardware and new smartphones will host Chinese chips. It’s the way of the world.