So, ‘Arm mini China’ is here. The Chinese joint venture is an interesting move that may provide a powerful long-term boost for the tech giant well and other big companies like it. But do smaller tech firms need to establish similar ventures to thrive in the Chinese market?
Joint venture
A JV in China has been part of Arm’s business plan since at least February 2017 and has now been confirmed.
It seems Arm has read the tea leaves and decided to cement its position in China for the long term. This — along with its innovation fund with HOPU Investments announced last year — marks Arm’s attempt to strengthen its foothold in the face of multiple threats to its considerable business here in China.
Arm’s threats in China
The pressures the company faces break down into three main areas:
1.Indigenous innovation
For the past five years, the Chinese government has aggressively promoted home-grown innovation in semiconductors as part of its ‘Made in China 2025’ programme.
The government has ploughed in huge sums directly as well as encouraging VC investment in companies such as Tsinghua Unigroup, lesser-known state-owned players such as Phytium and artificial intelligence pioneers such as Cambricon, Bitmain and Horizon Robotics.
Whilst many of these companies base their designs on Arm’s architecture, there’s a growing appetite for China to create its own competing technology. Loongson and C-Sky have been trying this for some years, with limited success. But, given the US sanctions foisted on ZTE which I wrote about here (although it now looks like the US may rescind them), the Chinese government has never been more acutely aware of its reliance on foreign semiconductor technology — nor more determined to break free.
Ironically, US actions seeking to undermine western support for the Chinese tech industry may have smoothed the path for a western giant to transfer its technology to China by making the Chinese government all the more receptive.
2.RISC-V
One of the reasons the likes of Arm and Intel have been so successful is because they’ve created mature, global ecosystems for their architectures which are extremely hard to replicate. It’s certainly proved tough for companies like Loongson — who only really have domestic government customers — to compete.
Trying to rival Arm and Intel globally is like attempting to move into the smartphone operating system market to take on Android or iOS. Even the mighty Microsoft failed at this.
But the open-source RISC-V Instruction Set Architecture (ISA), whilst still at a nascent stage, is growing extremely quickly and starting to threaten Arm.
Most Chinese semiconductor companies already have teams researching this new movement. C-Sky (recently acquired by Alibaba), for example, has realised the potential of RISC-V in enabling it to be part of a global ecosystem, joining as a platinum member early in 2017; while Taiwan’s Andes Technology has dropped its own architecture and thrown its lot in 100% with RISC-V.
I wouldn’t be surprised to see the Chinese government get behind the RISC-V project now, the same way India has. Indeed, institutes within The Chinese Academy of Sciences are already working on various RISC-V related R&D projects, with some being taped out this year.
Right now, RISC-V is mainly pitched as a replacement for Arm M0 or M3 cores in Internet of Things designs. But I don’t think it will stop there.RISC-V designs are already being considered for use in embedded PCs. There’s even one US company, Esperanto, designing an artificial intelligence chip based on thousands of RISC-V cores.
3.Japanese connection
While the 2016 acquisition of Arm by Japan’s SoftBank for over $30 billion hasn’t had a major negative impact on ARM’s Chinese sales, it’s certainly caused concern in China.
I remember meetings at the time with Chinese semiconductor players who were worried about relying so heavily on a company that was turning from British to Japanese overnight. And I was involved in at least two licensing deals in China where the Arm-SoftBank deal was a key factor in the customer switching IP.
Arm has attempted to allay Chinese firms’ fears of relying on a Japanese company. Its first tech symposia in China following the Softbank acquisition focused on this very topic and was designed to maintain Chinese customers’ trust. But these concerns have not melted away.
The Arm JV is clearly an attempt to go much further and make the company appear as Chinese as possible.
A wise move?
So, is this a wise move for Arm? Only time will tell for sure, but here are my two cents.
In the short-term, I think Arm will take a hit from this deal. Chinese sales revenues, which have increased 110-fold over the past 10 years, will roughly halve for Arm as it owns only 49% of the JV, the rest going to Chinese investors including The Bank of China and Internet firm Baidu.
It also won’t do much to boost Arm’s sales in the next couple of years. Its revenues in China have been increasing anyway, and the largest players in the market have been licensing its architecture for years. As they’ve grown, so have Arm’s sales.
But Arm is playing the long game, and here this could prove to be an extremely shrewd move.
Vitally, the JV will be seen as a Chinese company. This means it can become part of the Made In China 2025 programme.
Arm mini China is also preparing for an IPO in two to three years’ time. If this goes through smoothly — and as early as that — we can be sure it had the blessing of the Chinese government, and many of the threats I mentioned will have been addressed. This would put Arm on an exceptionally solid foundation for the long-term.
A good model to follow?
Given all this, will Chinese JVs of this kind become a trend, with other foreign semiconductor companies doing similar?
Perhaps, for some. For starters, Arm isn’t the first. Qualcomm already has JVs in China, including HXT Semi and a new partnership with Leadcore. Both are designed to help Qualcomm deal with local requirements more efficiently.
For big players like Qualcomm and Arm, it makes good sense: it enables them to cement their position in China and makes it easier for local companies to license domestically.
Smaller tech firms
But what about smaller technology companies? For these, I firmly believe there’s no necessity to follow suit — and there are significant risks that come with joint ventures.
Smaller players entering a JV would need extremely experienced legal support. In some cases, they’d need to share source code, and could face the dangers of their intellectual property leaking from the venture and of losing management staff to competing enterprises.
By contrast, using a good local distributor with the ability to deal with RMB payments, along with your own team on the ground to manage sales, is a method that works extremely well for small to medium tech firms operating in China.
Unless — as with some large semiconductor players — the Chinese government feels threatened by an acute over-reliance on your technology, going local to Arm’s extent is unlikely to be necessary or profitable.