Friday 29 Mar 2024
By
main news image

This article first appeared in The Edge Malaysia Weekly on March 26, 2018 - April 1, 2018

WITH profit margins for personal computers razor thin and the smartphone market close to saturation, artificial intelligence (AI) has been crowned the “next big theme” for the semiconductor industry.

According to advisory firm KPMG LLP’s survey of 150 semiconductor industry leaders, AI — along with cloud computing and the internet of things (IoT) — will have a much greater impact on their revenues this year.

In fact, AI inference is expected to take off in the next two to three years and become the biggest semiconductor market.

Today, the US semiconductor industry is still dominant globally. Market observers, however, warn that its pole position might be threatened by China’s planned investment binge to build up its own chipmaking industry.

In July last year, China unveiled an ambitious development plan to become the global leader in AI by 2030. The world’s second largest economy also aims to surpass its rivals technologically by creating a domestic industry worth US$150 billion.

While it is currently unclear how much China as a whole is spending on AI, the Beijing city government has committed US$2 billion to an AI development park in the west of the capital.

So, how important is this AI race to the two superpowers? Will the US lose its dominance in the global semiconductor industry if it is defeated in the AI race? And will China become the global leader if it wins the race? More importantly, amid the ongoing and intensified US-China AI race, what should semiconductor and semiconductor-related firms do to grab a slice of the growing AI applications market?

Alan Priestley, research director at IT research and advisory firm Gartner Inc, points out that AI is still at an early stage of its development despite the intensifying race. Much of the early growth is occurring in academia and at start-ups, where initial budgets for development tools are often low.

“For technology strategic planners at many semiconductor vendors, identifying and working with the right software partner will be a key factor in winning business in the AI market,” he tells The Edge via email.

Priestley urges semiconductor players to prioritise AI-based analytics by developing a strategic plan to work only with semiconductor manufacturers that deploy it, within the next 6 to 12 months.

He recommends that they establish the adoption of AI across key areas within the chip development process. Yield improvement and predictive maintenance should be included as criteria that chip manufacturers must meet in order to win the business.

He suggests that semiconductor players identify niche AI-based analytics firms that provide customised solutions and services to the industry as partners for chip manufacturers.

Forbes contributor Paul Armstrong says semiconductor firms should start taking AI more seriously by forging partnerships, as well as exploring, testing and investing in it.

“We are years away [from] seeing the massive potential it can provide [to] a lot of industries. When quantum computing goes mainstream, it will be like continuously pouring petrol on a fire,” says Armstrong, who is also the founder of HERE/FORTH, an emerging technology advisory firm.

Globetronics Technology Bhd CEO Datuk Heng Huck Lee highlights that Malaysian semiconductor companies have broad experience and are capable of producing a wide range of sensor and wireless-related technology and products.

“I think we should be able to continue to leverage and expand into AI quite readily. All of us are globally cost-competitive enough to win many of these new opportunities,” he says.

Overall, says ViTrox Corp Bhd co-founding CEO and president Chu Jenn Weng, rising trade tensions between the US and China should benefit Southeast Asian countries. “That’s because we have been maintaining a neutral stance. We don’t choose sides. We don’t neglect any market.”

 

Who will win the race?

Interestingly, China’s intention to become the world leader in the development of AI has many in the US questioning what the Trump administration is doing to protect the country’s dominant position in the race.

The US semiconductor industry is an important contributor to the American economy. It has grown more than any other major local manufacturing industry over the last 25 years.

Executives of American technology companies, as well as AI experts in the country, have raised concerns that the Trump administration has paid little attention publicly to AI policy.

Ian Buck, vice-president of technology company Nvidia Corp’s accelerated computing business unit, has said US federal agencies are not investing enough in AI.

“I would love to see the government invest more in research funding and proactively evaluate the deployment of AI,” he said.

Globetronics’ Heng concurs. “Ian could be right. To date, we have yet to see any solid strategy from the US president on future technology development.”

In comparison, the Chinese government has a very clear focus on becoming a major player in technology, especially in semiconductor technology and global revenue capturing.

“With the strategy to invest nearly US$100 billion over a five-year period, I would not be surprised if China gains the leading position,” Heng says.

Having said that, he believes the US will still have some advantages in the race as AI also consists of the software component.

“I think the US can continue to attract the best talent in the world. With so much intellectual property controlled by the US, it is not easy for any country to take over its dominance, not even China, in the next 5 to 10 years. Beyond that, I cannot be sure,” he says.

Heng adds that as the global semiconductor industry is much more than about AI, it will be a growing pie and, therefore, whoever wins the AI race may not dominate it.

Gartner’s Priestley says, in the short term, China is facing several challenges, including a skills gap and lack of high-end computer chips, for which it is heavily dependent on foreign suppliers.

“According to a report published by the networking site LinkedIn on July 6 last year, of the AI talent pool of 1.9 million, China accounts for just over 50,000, while the US tops the ranking with its pool of 850,000,” he says.

ViTrox’s Chu is of the view that China is currently leading the AI race as it has invested more than the US.

“Whoever wins, we should benefit from AI adoption because it will complicate the assembly process of semiconductors, so much so that humans just cannot handle it and  they need to find ways to automate the process,” he says.

HERE/FORTH’s Armstrong also says “it is a great possibility” that China will become the dominant force in the global semiconductor industry if it wins the AI race.

According to a poll of 126 senior AI professionals released by professional services giant Ernst & Young Global Ltd on March 8, 40% of respondents believe the US is leading the race in AI adoption, followed by China (32%).

With data becoming more readily accessible and governments worldwide playing an active role in investing in innovation, the US and China are starting to feel the heat from other countries.

 

 

Save by subscribing to us for your print and/or digital copy.

P/S: The Edge is also available on Apple's AppStore and Androids' Google Play.

      Print
      Text Size
      Share