Everything Is Not ‘Made in China,’ Especially Chips

Alibaba Cloud's announcement of their latest server chips has brought all eyes to China for producing microprocessors, but it seems that there is still quite a long road

Alibaba Cloud, last week, unveiled the Yitian 710 server chips built by T-Head, Alibaba Group’s chip development business. Along with it, Alibaba Cloud also announced the development of Panjiu, its proprietary servers where the new chip will be used. The company claims that the set “will enhance the cloud services by optimising computing performance and energy efficiency.” Yitian 710 is being called one of the most advanced chips, which not only is better in performance but also saves energy. This could be an important milestone in China’s chase to become self-sufficient in the semiconductor industry, but experts suggest that the country is not there yet.

Alibaba is a recent entrant in the semiconductor industry and released its first AI inference chip Hanguang 800, in 2019. Alibaba is one of the many Chinese firms that answered Beijing’s call to invest in the development of technologies and manufacturing capacities to become self-sufficient. Alibaba had announced last year that it would invest CNY 200 billion in its cloud business over three years with a focus on infrastructure development which includes servers, network, chips, and OS. But, with limited semiconductor manufacturing capabilities at the domestic level, Alibaba is likely to outsource production to either Taiwan Semiconductor Manufacturing Co. (TSMC) or South Korea’s Samsung Electronics Co., as only they have the capacity to mass-produce 5-nm chips.

The tech behind Yitian 710 and Panjiu

Yitian 710 is built upon 5nm process technology and is powered by 128 Arm cores with a 3.2GHz top clock speed. Each processor chip has 60 billion integrated transistors, is compatible with Armv9 architecture and includes 8 DDR5 channels and 96-lane PCIe 5.0, providing high memory and I/O bandwidth. The chip achieved a score of 440 in SPECint2017 (a standard benchmark to measure CPU integer processing power), surpassing that of the current state-of-the-art Arm server processor by 20% in performance and 50% in energy efficiency.
Panjiu has been developed for the next-gen of cloud-native infrastructure. The servers are optimised for both general-purpose and specialised AI computing as they have separating computing and storage. With a modular design approach for large-scale data centre deployment, these servers are expected to deliver higher economic value for a wide variety of cloud-native workloads, such as containerised applications and computed optimised workloads.

China’s Insufficiency in making Chips

Beijing has conducted drives in the past to reduce reliance on imported chips and gain semiconductor independence. The pandemic did bring about a major change, and many Chinese companies started introducing chips. Over 25,000 new semiconductor companies have been registered in China since 2020, which is triple that of what it had before. Since the beginning of 2021, big names like Huawei, Baidu, Alibaba, Oppo, Midea, Xiaomi, Gree, TCL and Haier have also upped their chip game.

Subscribe to our Newsletter

Join our editors every weekday evening as they steer you through the most significant news of the day, introduce you to fresh perspectives, and provide unexpected moments of joy
Your newsletter subscriptions are subject to AIM Privacy Policy and Terms and Conditions.

China’s dream of becoming a technology superpower is being held back due to the lack of semiconductor manufacturing. In 2020, China imported chips worth over USD 350 billion. The global hub of manufacturing can mass-produce almost anything but has failed at making microchips and especially the high-end ones. Even if big companies can make breakthroughs in the design of chips, the actual fabrication still needs to be outsourced as there are no mass-production facilities like the one by Samsung or TSMC.

Since 1999, when Nvidia invented the graphics processing units (GPUs), it has retained its lead, and currently, China has only a few early-stage startups in this segment. TSMC has been leading the foundry segment for the last 33 years, and China SMIC remains years behind in technology.

Chinese semiconductors are not very relevant in the global semiconductor market. Experts believe that Chinese players are still decades behind in manufacturing technology areas like advanced software design tools and lithography.

Not just from a nationalised perspective, but from a business perspective too, producing domestic chips makes a lot of sense as it supports a better supply chain. This came to notice after 2018 when the US prevented Huawei from purchasing American chips. For tech giants like Alibaba and Baidu, a self-designed chip can any day cater to internal needs better than an outsourced chip.

Political tensions hinder joint development

The Yitian 710 server chip by Alibaba is the closest that a Chinese company has got in terms of advancement as compared with chips of global rivals like Amazon.com Inc., Google, Intel Corp. and AMD Inc.

China’s import of semiconductors makes it at least 25% of major American semiconductor companies’ sales in the Chinese market. This mutual dependence proves to be beneficial for both countries. Tencent and Alibaba would not be the technology giants as we know them if they relied on Chinese microprocessors during their initial years. US companies, too, have benefited a lot from Chinese customers and their innovations using their chips.  

There are hundreds of semiconductors related joint collaborations that take place among companies in these countries. But, companies on both sides receive a push from the governments to reduce dependency on each other. Washington wants the American economy to be less dependent on China, and the latter clearly stated in their 14th five-year plan to achieve the goal of “technology independence.”

Though the future of the Chinese semiconductor industry is not clear yet, US companies have already started losing influence and even market share in the Chinese technology sector. The new innovations by Chinese companies are attracting local markets, and now they are way more determined to make a new semiconductor ecosystem in China.  

Wrapping up

Baidu was the first among China’s tech giants to venture into the semiconductor industry. It started chip research in 2010 and released the first product Baidu Kunlun in 2018. This August, Baidu launched Kunlun 2. In 2018, Gree Electronics announced an investment of USD $7.7 billion to make its own chips. In the same year, Alibaba acquired a domestic chipmaker which is now PingTouGe Semiconductor and has launched two AI chips since then. Tencent has also invested in Enflame, a chip startup, for four rounds, and its cloud computing company added chipmaking in 2020.

The semiconductor industry is very complex and requires years of investment without immediate returns. It took over ten years for Huawei to develop HiSilicon. Though it has made the company stand out in the global chip market, the technology is still behind the dominant players.

Meeta Ramnani
Meeta has completed PGD in Business Journalism from IIJNM, Bangalore. She comes with over six years of experience in journalism and writes about emerging enterprise technologies with a focus on digital transformation. She loves to go on bike rides and stays in touch with nature. Contact: meeta.ramnani@analyticsindiamag.com

Download our Mobile App

MachineHack | AI Hackathons, Coding & Learning

Host Hackathons & Recruit Great Data Talent!

AIMResearch Pioneering advanced AI market research

With a decade of experience under our belt, we are transforming how businesses use AI & data-driven insights to succeed.

The Gold Standard for Recognizing Excellence in Data Science and Tech Workplaces

With Best Firm Certification, you can effortlessly delve into the minds of your employees, unveil invaluable perspectives, and gain distinguished acclaim for fostering an exceptional company culture.

AIM Leaders Council

World’s Biggest Community Exclusively For Senior Executives In Data Science And Analytics.

3 Ways to Join our Community

Telegram group

Discover special offers, top stories, upcoming events, and more.

Discord Server

Stay Connected with a larger ecosystem of data science and ML Professionals

Subscribe to our Daily newsletter

Get our daily awesome stories & videos in your inbox