Singapore Home Team uses NVIDIA’s latest AI chips to supercharge public safety robots

Published Sep 23, 2026, 05:27 PM

Updated Sep 23, 2026, 05:27 PM

SINGAPORE - A future where firefighters tackle chemical plant blazes alongside hazard-sensing robots is one step closer to reality, as Singapore’s Home Team deploys NVIDIA’s latest artificial intelligence chips to supercharge its robotic capabilities.

The Home Team Science and Technology Agency (HTX) is the first organisation in Singapore to deploy NVIDIA’s latest GB300 chip, which was commercially available in the first half of 2026 .

GB300 is twice as fast as its predecessor GB200 in AI deep thinking. It also handles text, images and code much faster too.

The new GB300 chips provide massive computing power so HTX can train much smarter AI, specifically its NGINE AI infrastructure, to process text, video and audio at once and control robots at scale without fear of security being compromised.

Revealing this at the NVIDIA AI Day conference in Singapore on Sept 23, HTX’s chief AI officer Ang Chee Wee said: “As the first in Singapore to operationalise NVIDIA’s GB300, we are now able to train larger multimodal models and embodied AI models and run inference at scale.”

Multimodal AI models are able to process data from different mediums, such as text, images and audio, while embodied AI models are software that control robots.

He added: “Most importantly, the state-of-the-art sovereign AI infrastructure ensures that both our code and our data stay within our control. With NGINE, we can now operationalise AI at scale without compromising security.”

Ang said the new tools will allow Home Team officers to do more with rising public expectations on safety and security.

“At the same time, AI tools will also be used by bad actors, so the threats that we must deal with will get more complex,” he added .

Addressing the issue of job displacement, he said the goal is not to replace officers, but to have the AI robots take on the dull, dirty and dangerous tasks, such as dealing with hazardous materials and fighting fires.

The Home Team has already been using AI for a range of tasks , including facial recognition for immigration clearance at the checkpoints and drafting police reports with the help of the Report Lodging Co-Pilot (R-Cop) .

AI is also being used to identify and take down scam sites that impersonate the police, government agencies and banks.

So far, use of AI has saved the Home Team close to a million man-hours annually, with this number expected to accelerate in 2027, added Ang .

In late September, HTX is also expected to establish the Home Team Humanoid Robotics Centre, or H2RC, to further its work in embodied AI and humanoid robotics development. Some $100 million has been committed to the centre, which will focus on developing robots to enforce public safety.

Senior Minister of State for Digital Development and Information Tan Kiat How, who also spoke at the event, said embodied AI matters for Singapore because of the physical constraints of land, labour and energy, and its ageing workforce.

“The first wave of generative AI transformed the interface between humans and information. The next may transform the interface between intelligence and the physical world,” said Tan.

“Our opportunity is to become exceptionally good at moving technologies from possibility, to proof, to scale. We can bring frontier technology together with demanding users, researchers, engineers, regulators, capital and real-world operating environments.”

Also at the event, Singapore-based tech giant Sea announced plans to adopt NVIDIA’s Vera Rubin next-generation AI computing architecture for powering autonomous AI agents.

This is expected to pave the way to supercharge Sea’s gaming platform Garena, digital payment service Monee and e-commerce system Shopee.

This could mean reducing the time and effort by Shopee sellers to create informative product listings and improving product discovery, or improve fraud detection and credit risk assessment on Monee.

David Sun is The Straits Times’ crime correspondent. He has a background in criminology and is a licensed private investigator.

AI/artificial intelligence

发布于 2026 年 9 月 23 日下午 5:27

更新于2026年9月23日下午5:27

新加坡——随着新加坡内政团队部署英伟达最新的人工智能芯片以增强其机器人能力，消防员与危险感知机器人并肩扑灭化工厂火灾的未来距离现实又近了一步。

新加坡内政科技局 (HTX) 是新加坡第一个部署 NVIDIA 最新 GB300 芯片的机构，该芯片于 2026 年上半年上市。

GB300 在人工智能深度思考方面的​​速度是其前代产品 GB200 的两倍。它处理文本、图像和代码的速度也更快。

新的 GB300 芯片提供了强大的计算能力，因此 HTX 可以训练更智能的 AI，特别是其 NGINE AI 基础设施，以同时处理文本、视频和音频，并大规模控制机器人，而无需担心安全受到损害。

9 月 23 日，在新加坡举行的 NVIDIA AI Day 大会上，HTX 的首席人工智能官 Ang Chee Wee 透露：“作为新加坡首家将 NVIDIA GB300 投入运营的公司，我们现在能够训练更大的多模态模型和具身人工智能模型，并大规模运行推理。”

多模态人工智能模型能够处理来自不同媒介（如文本、图像和音频）的数据，而具身人工智能模型则是控制机器人的软件。

他补充道：“最重要的是，最先进的自主人工智能基础设施确保我们的代码和数据始终在我们掌控之中。借助 NGINE，我们现在可以在不损害安全性的前提下大规模部署人工智能。”

Ang表示，这些新工具将使内政团队官员能够更好地满足公众对安全保障日益增长的期望。

“与此同时，人工智能工具也会被不法分子利用，因此我们必须应对的威胁将变得更加复杂，”他补充道。

针对工作岗位流失的问题，他表示，目标不是取代警员，而是让人工智能机器人承担枯燥、肮脏和危险的任务，例如处理危险材料和灭火。

内政团队已经将人工智能用于一系列任务，包括在检查站使用面部识别进行移民许可，以及借助报告提交助手（R-Cop）起草警方报告。

人工智能也被用于识别和关闭冒充警察、政府机构和银行的诈骗网站。

Ang补充说，到目前为止，人工智能的使用每年为内政团队节省了近100万工时，预计到2027年这一数字还会加速增长。

预计HTX将于9月下旬成立内政部人形机器人中心（H2RC），以进一步推进其在具身人工智能和人形机器人开发领域的工作。该中心已获得约1亿美元的投资，将专注于开发用于维护公共安全的机器人。

新加坡数码发展及新闻部高级政务部长陈杰豪也在活动上发表了讲话，他表示，由于新加坡在土地、劳动力和能源方面的实际限制以及劳动力老龄化问题，具身人工智能对新加坡来说至关重要。

谭表示：“第一波生成式人工智能改变了人与信息之间的关系。下一波可能会改变智能与物理世界之间的关系。”

“我们的机遇在于，能够极其高效地将技术从可能性转化为验证，最终实现规模化应用。我们可以将前沿技术与要求苛刻的用户、研究人员、工程师、监管机构、资本以及真实的运行环境结合起来。”

在本次活动中，总部位于新加坡的科技巨头 Sea 还宣布计划采用 NVIDIA 的 Vera Rubin 下一代 AI 计算架构来驱动自主 AI 代理。

预计这将为Sea旗下的游戏平台Garena、数字支付服务Monee和电子商务系统Shopee的快速发展铺平道路。

这可能意味着 Shopee 卖家可以减少创建信息丰富的商品列表和改善商品发现所需的时间和精力，或者改进 Monee 上的欺诈检测和信用风险评估。

大卫·孙是《海峡时报》的犯罪记者。他拥有犯罪学背景，并且是一名持证私人侦探。

人工智能