字号 ·· | 护眼
亚洲新闻台

新加坡半导体产业承诺投入逾5亿新元用于研发,全球芯片竞赛加剧Singapore semiconductor industry commits over S$500m to R&D as global chip race intensifies

点「原文对照」整页切到原文,或双击某段只看那段的原文。

新加坡:半导体业者已承诺从2026年起五年内,为新加坡的研发项目投入超过5亿新元(约3.9亿美元)。

SINGAPORE: Semiconductor industry players have committed more than S$500 million (US$390 million) to research and development projects in Singapore over five years from 2026.

根据科技研究局(A*STAR)的数据,这超过了2020年至2025年间在类似工作上投入的4亿多新元。

This exceeds the more than S$400 million spent on such efforts between 2020 and 2025, according to the Agency for Science, Technology and Research (A*STAR).

此举旨在巩固新加坡在先进芯片技术领域的优势,并助力本地企业在资金密集型行业中竞争。

It comes as the country looks to sharpen its edge in advanced chip technologies and help local firms compete in a capital-intensive industry.

A*STAR创新与企业发展副总裁杨以嘉教授表示,新加坡拥有全球“最完善、协调度最高的生态系统之一”。

Professor Yeo Yee Chia, A*STAR’s deputy chief executive for innovation and enterprise, said Singapore has “one of the most well organised and coordinated ecosystems” globally.

他补充道:“我们认为,这将成为企业选择能够拓展其研发和业务地点时的关键差异化因素。”芯片制造的下一个疆界新加坡已占全球芯片产量约十分之一,这得益于多年来在半导体制造和研发方面的投入。

“That's what we think is going to make a difference when companies select a location where they can grow their R&D and their business,” he added. THE NEXT FRONTIER IN CHIPMAKING Singapore already accounts for about one in every 10 chips made globally, following years of investment in semiconductor manufacturing and R&D.

但人工智能的崛起给芯片制造商带来了新需求,而传统让计算机运行更快的路径却变得愈发困难。

But the rise of artificial intelligence is placing new demands on chipmakers, just as the conventional route to making computers faster is becoming more difficult.

数十年来,业界通过在芯片上封装越来越多的晶体管来提升算力,大体遵循被称为摩尔定律的趋势。该趋势于1960年代首次被观察到时,晶体管尺寸以微米计。

But the rise of artificial intelligence is placing new demands on chipmakers, just as the conventional route to making computers faster is becoming more difficult. For decades, the industry improved computing performance by packing ever more transistors onto a chip, broadly following a trend known as Moore’s Law.

如今,部分芯片特征尺寸仅有几十个原子厚度,已达量子效应显著的尺度。进一步突破极限在技术上愈发艰难,成本也日益高昂。新加坡已投入十余年时间,开发可提供提升算力替代路径的技术。

When the trend was first observed in the 1960s, transistors were measured in micrometres. Today, some chip features are just dozens of atoms thick, reaching scales where quantum effects become more significant. Pushing the limits further has become increasingly technically challenging and expensive.

咨询公司Frost & Sullivan亚太区董事总经理Ravi Krishnaswamy表示:“我认为下一场竞争将聚焦于先进封装、异构集成和芯粒。”

Singapore has spent more than a decade developing technologies that could provide alternative ways to improve computing performance.

“这些是我认为将涌现更多经济增值机遇的领域。” 此类技术将使更多组件能够紧密堆叠在一起,缩短数据传输距离,并有望降低能耗。

“I think the next battle will be fought on advanced packaging, heterogeneous integration, chiplets,” said Mr Ravi Krishnaswamy, managing director for Asia-Pacific at growth consulting and market intelligence firm Frost & Sullivan.

其他研究领域包括光子学,它利用光而非电信号传输数据,从而更高效地满足数据需求旺盛的AI系统;以及专为应对数据中心巨大电力需求而设计的芯片。构建一体化生态系统叶教授表示,科技研究局(A*STAR)还看到“越来越多”在新加坡没有业务布局的海外企业来到这里,与现有合作伙伴开展合作。

“These are areas where I think more economic value-add opportunities will emerge.” Such technologies would allow more components to be packed closely together, shortening the distance data has to travel and potentially reducing energy use. Other areas of research include photonics, which uses light rather than electrical signals to move data more efficiently for data-hungry AI systems, as well as chips designed to handle the enormous power requirements of data centres BUILDING AN INTEGRATED ECOSYSTEM A*STAR is also seeing “an increasing pipeline” of overseas companies without a presence in Singapore coming here to work with existing partners, Prof Yeo said.

克里希纳斯瓦米先生表示,新加坡的优势可能在于其整合不同能力的能力。

Other areas of research include photonics, which uses light rather than electrical signals to move data more efficiently for data-hungry AI systems, as well as chips designed to handle the enormous power requirements of data centres BUILDING AN INTEGRATED ECOSYSTEM A*STAR is also seeing “an increasing pipeline” of overseas companies without a presence in Singapore coming here to work with existing partners, Prof Yeo said.

他补充道:“所以不必局限于单一细分领域。我认为,作为一个一体化的生态系统能力,我们大概能将许多这样的东西摆上台面。”基础设施是方程式的另一部分。科技研究局正寻求投资下一代工具,供企业利用以获取竞争优势。

Mr Krishnaswamy said Singapore’s advantage may lie in its ability to combine different capabilities. “So it doesn't need to be one niche. I think we can probably bring a lot of these things to the table as an integrated ecosystem capability,” he added.

叶教授表示:“通过提供我们的基础设施、人才和技术开发能力,我们希望加速企业从研发到制造的进程,实现高速、大规模发展。”降低创新门槛在全球巨头投入数十亿资金建设设施和引进人才的行业中,新加坡半导体领域的中小企业(SME)需要争取一切可能的优势。

Infrastructure is another part of the equation. A*STAR is looking to invest in next-generation tools that firms can tap to gain a competitive advantage. “By providing our infrastructure, by providing our talent and technology development capabilities, we want to accelerate the companies’ capabilities to go from research to development to manufacturing at speed and at scale,” Prof Yeo said.

叶教授表示:“我们希望让这些企业拥有‘不公平优势’。通过允许它们利用我们的投资,降低其研发门槛,增强其创新和提供解决方案的能力。” 新加坡金属透镜企业MetaOptics便是一家利用该支持获取人才、知识产权和研发资源的公司。

LOWERING THE BARRIERS TO INNOVATION In an industry where global giants spend billions on facilities and talent, small- and medium-size enterprises (SMEs) in Singapore’s semiconductor sector need every edge they can get.

其首席执行官 Aloysius Chua 表示,无论总部设在何处,年轻科技公司筹集资金仍然困难。

“We want to allow such companies to have an unfair advantage,” Prof Yeo said. “By allowing them to leverage our investment, it lowers the barrier to their R&D and to their ability to innovate and provide solutions.” Singapore metalens firm MetaOptics is one company that has tapped that support for talent, intellectual property and R&D efforts.

“我认为我们早期的支持者是新加坡的一些中小企业,此外我们也获得了 A*STAR 的大力支持,”他说。

Its CEO Aloysius Chua said raising money remains difficult for young technology companies, regardless of where they are based.

MetaOptics 此后已上市并将产品推向海外,该公司开发用于生产平面光学元件的技术,可应用于从相机、传感器到高速计算连接等领域。

“I think our early supporters are some of the SMEs in Singapore, and then also we have strong support from A*STAR,” he said.

半导体设备制造商 NexGen Wafer Systems 同样受益于共享基础设施的使用。

MetaOptics, which has since gone public and taken its products overseas, develops technology for producing flat optical components that can be used in applications ranging from cameras and sensors to high-speed computing connections.

其首席执行官 Cheung Ting Kwan 表示,自行建设公司所需的所有设施将意味着投资建设自己的洁净室和其他专用设备。

Semiconductor equipment manufacturer NexGen Wafer Systems has similarly benefited from access to shared infrastructure.

“如果我们自己做,这变得相当昂贵,对于像我们这样的中小企业来说成本过高,”他说。从技术到客户 但获得基础设施和人才只能解决部分问题。

Its CEO Cheung Ting Kwan said building all the facilities the company requires itself would mean investing in its own clean room and other specialised equipment. “If we do it ourselves, this becomes quite prohibitive, expensive for a small- and medium-sized company like us,” he said.

对于新加坡半导体公司而言,另一个障碍是寻找足够的客户,将专业技术转化为可持续的业务。

FROM TECHNOLOGY TO CUSTOMERS But access to infrastructure and talent solves only part of the problem.

从事硅光子学业务的新加坡公司 Advanced Micro Foundry (AMF) 去年被跨国芯片制造商 GlobalFoundries 收购。

For Singapore semiconductor companies, another obstacle is finding enough customers to turn specialised technologies into sustainable businesses. Advanced Micro Foundry (AMF), a Singapore company working in silicon photonics, was acquired by multinational chipmaker GlobalFoundries last year.

AMF 联合创始人 Patrick Lo 表示,为小众技术寻找客户是一项挑战。

Finding customers for niche technologies was a challenge, said AMF co-founder Patrick Lo.

“我们喜欢建设晶圆厂基础设施,但希望它能得到充分利用,”现任 GlobalFoundries 技术开发与先进光子学资深研究员的罗先生说。

"We like to build the fab infrastructure, but want it to be fully utilised,” said Mr Lo, who is now a senior fellow of technology development and advanced photonics at GlobalFoundries.

“这种充分利用绝对将来自应用端。这实际上是我希望新加坡整体能拥有更多设计师涌现、更多产品开发者加以利用的方面。”

“That full utilisation will be absolutely coming from the application side. This is something I like Singapore as a whole, actually, to have more designers come up, more product developers to leverage.” It is a challenge some companies are also grappling with as they try to translate breakthroughs into commercial products.

这是一些公司在尝试将突破转化为商业产品时也在努力应对的挑战。

“That full utilisation will be absolutely coming from the application side. This is something I like Singapore as a whole, actually, to have more designers come up, more product developers to leverage.” It is a challenge some companies are also grappling with as they try to translate breakthroughs into commercial products.

“也许人们能获得大量资金,但之后呢?”蔡先生说。“如果我把技术开发到这种程度,却无法将其商业化,那么我认为作为一家公司,我们就失败了。”

“Maybe people get a lot of funding, but what's after?” Mr Chua said. “If I develop the technology to what it is, and I cannot commercialise it, then I think as a company we've failed.”