Singapore AI News & Daily Briefing

Alibaba Cloud has opened its fifth Malaysia facility in Johor, its biggest Southeast Asian data center footprint yet. The new region will offer agentic AI services to Malaysian businesses from the second half of this year. For Singapore companies, this expansion could mean more regional cloud options and potentially more competitive pricing for AI-powered services.

Singapore's Ministry of Education is rolling out responsible AI integration across schools, focusing on tools that help teachers rather than replace them. New guidelines ensure AI supports learning without undermining critical thinking or student-teacher relationships. The approach balances embracing new technology with protecting the quality of education.

Singapore Post opened a new high-tech automated parcel sorting hub to speed up package handling across the island. The facility uses advanced automation and AI to process parcels faster and more accurately. For everyday Singaporeans, this could mean quicker deliveries for online shopping and fewer delays during peak periods like the year-end shopping rush.

Singapore is seeing more white-collar workers losing their jobs, with AI adoption, offshoring, and restructuring cited as common reasons. The National Trades Union Congress handled over 3,900 retrenchment cases recently, with more professionals now seeking union help. If you work in an office job, this is a reminder that AI-driven changes are affecting careers here at home.

A critical vulnerability called 'BadHost' was found in Starlette, a popular Python web framework downloaded 325 million times weekly. AI agents built on frameworks using Starlette could have been exposed to remote attacks. The vulnerability has been patched, but any AI agent running unpatched versions remains at risk.

Google tested its AI tutoring tool in 200+ Sierra Leone schools and found students using it gained 1.2 to 1.7 years of math progress in just 8 weeks. The AI doesn't give answers—it asks guiding questions to make kids think. Teachers loved it too, saying it helped them discover better ways to explain lessons. The catch: higher-performing students benefited most, so more work is needed to help struggling learners.

AI companies spend huge resources measuring how well their models perform on tests and benchmarks, but almost nothing measuring how those models affect the humans using them. A new essay from the Center for Humane Technology argues we're missing the most important question: what is AI doing to us? The piece raises concerns about impacts on children's brain development, mental health crises, romantic relationships, and whether AI companions are pulling lonely people away from real human connections. The author calls for long-term studies similar to how drugs are monitored after FDA approval, and argues AI companies should share user data with independent researchers.

At a robotics conference in Vienna, company AGILINK wowed crowds by having a robot twist a balloon into a dog — a task so sensitive that even slight force errors cause failure. The demo highlighted "contact intelligence," or how robots maintain stable physical interaction as objects deform and change. AGILINK also unveiled a new robotic hand with ultra-sensitive tactile sensors that can detect the weight of a sheet of paper.

Nextdoor's engineering team is shipping features in days instead of weeks using OpenAI's Codex coding tool. One engineer can now build across mobile, web, and backend—a job that used to need three separate teams. Managers who haven't coded in years are back in the codebase. The company's bottleneck has shifted from building speed to deciding what to build next.

General Motors is developing new sodium-ion batteries specifically to power AI data centers and energy storage systems. These batteries can work without active cooling, which cuts costs significantly. GM is partnering with startup Peak Energy to bring sodium-ion cells to market after 2028. The move addresses concerns that AI's massive data center buildout will drive up electricity costs for everyone.

Jun 09, 2026 6 min read Gemini 3.5 Live Translate is our latest audio model, delivering near real-time speech-to-speech translation in over 70 languages. A Anuda Weerasinghe Product Manager T Tony Lu Senior Staff Software Engineer  [Audio 3](https://storage.googleapis.com/gweb-uniblog-publish-prod/media/tts_audio_85287_umbriel_2026_06_09_15_20_53.wav) Listen to article This content is generated by Google AI. Generative AI is experimental [[duration]] minutes Twenty years ago, [translation at Google](https://blog.google/products-and-platforms/products/translate/fun-facts-google-translate-20-years/) began as one of our

Notion used Codex to build a voice input feature for web in just 3-4 hours—work that previously would've taken two engineers two weeks. One Notion engineer describes Codex as having an 'intern available 24/7' for research, bug fixes, and tweaks. Managers are coding alongside their teams again, and small teams can now tackle projects previously impossible for them.

Researchers have developed an AI system that can automatically track how glaciers shrink by analyzing satellite images. The deep learning model, trained on radar images of glaciers, can now detect the boundaries of glaciers with accuracy down to 68 meters - comparable to human analysts. This matters because glaciers worldwide are melting faster due to climate change, and manually tracking all of them is impossible. The team used just one labeled image per glacier plus summer reference photos to adapt the model to new regions, allowing them to map monthly changes for all 145 glaciers in Norway's Svalbard archipelago over nine years.

Jun 09, 2026 4 min read Google DeepMind Accelerator selects 15 robotics companies from across Europe to join the program. Providing 3 months of intensive mentorship and technical support, enabling the integration of AI into their core products. C Carolina Parada VP Robotics, Google DeepMind ## General summary Google DeepMind is launching a three-month accelerator program to support early-stage robotics startups across Europe. These companies will receive technical mentorship, product guidance, and access to advanced artificial intelligence models to help turn their research into real-world applications. You can learn more about the selected cohort and their work on the official program page. Summaries were generated by Google AI. Generative AI is experimental. ## Bullet points * Google Dee

New software engineering graduates are facing grim headlines about unemployment rates, but experts argue the CS degree itself isn't the problem. The real issue is a broken hiring pipeline full of 'ghost jobs' that never get filled. The article breaks down actual unemployment and underemployment data showing engineers still outperform many other fields, then offers practical advice: tap your real network (referrals account for 26% of hires), consider startups for early experience, build deployed projects rather than waiting for job offers, and crucially, learn AI engineering fundamentals beyond just using tools like Copilot.

Red Hat's official NPM package channel was hijacked to push malware that spreads like a worm between machines, stealing sensitive developer credentials including GitHub secrets, npm tokens, and cloud service logins. Anyone who installed affected @redhat-cloud-services packages in the past 36 hours should assume their systems and CI/CD pipelines are compromised. Red Hat says customer systems weren't directly affected, but the breach shows how trusted software channels can be weaponized.

This article is crossposted from IEEE Spectrum ’s careers newsletter. Sign up now to get insider tips, expert advice, and practical strategies, written i n partnership with tech career development company Parsity and delivered to your inbox for free! Job Hopping as an Engineer: The Pros and Cons I’ve changed jobs more times than I ever imagined I would. In the past 12 years, I’ve worked at seven different organizations. Some of those moves were forced by layoffs. Others were deliberate bets on my own trajectory. Job hopping, done strategically, is one of the fastest ways to accelerate your compensation and reinvent your professional identity. Engineers who understand when to move and when to stay tend to out-earn and out-rank their peers who simply wait for internal recognition. Unfortunat

The White House rushed to impose export controls on AI company Anthropic after a chaotic 24-hour period. This means certain AI models and technology may no longer be sent overseas, potentially limiting global access to cutting-edge tools. For everyday users, it could affect which AI services are available internationally.

The White House and Congress are reviving a push to override state AI laws with a single federal standard. If successful, this would replace a patchwork of different state rules on AI safety, deepfakes, and kids' online protection. For users, it means AI regulations could become more uniform – but also possibly weaker than some state laws.

Google has released Gemini 3.5 Live Translate, a new AI model that translates speech in real-time across more than 70 languages while preserving the speaker's tone and pace. Unlike older translation apps that pause between exchanges, this streams continuously for smoother conversations. It's rolling out on Google Translate apps for Android and iOS, and will soon be available in Google Meet for business users.

Oracle is positioning itself as an alternative cloud provider for startups hitting the "scaling wall" — that painful phase when venture capital cloud credits dry up and infrastructure costs spiral. The company offers what it calls "converged databases" that combine multiple data formats (including AI vector data) in one system, cutting the need for complex, expensive tech stacks. The piece highlights a real example: Singapore biotech firm Mirxes used Oracle's AI infrastructure to build a 24/7 clinician assistant that answers product questions without burning out its sales team. Oracle's regional head for ASEAN is quoted pushing the message that startups need tech partners offering performance and price predictability, not just raw compute power.

# Your TV's RS-232 port is a seriously useful automation tool - how to unlock its full potential | ZDNET X Trending * [I wore the world's first HDR10 smart glasses](https://www.zdnet.com/article/rayneo-air-4-pro-hdr10-smart-glasses-ces/) * [TCL's new E Ink tablet beats the Remarkable and Kindle](https://www.zdnet.com/article/tcl-note-a1-nxtpaper-ces/) * [Anker's new charger is one of the most unique I've ever seen](https://www.zdnet.com/article/anker-wireless-chargers-ces-2026/) * [Best laptop cooling pads](https://www.zdnet.com/home-and-office/best-laptop-cooling-pad/) * [Best flip phones](https://www.zdnet.com/article/best-flip-phone/) * [Roku TV vs Fire Stick](https://www.zdnet.com/home-and-office/home-entertainment/roku-tv-vs-fire-stick/) * [Galaxy Buds 3 Pro vs Apple AirPods Pro 3](ht

June 15 (Reuters) - iRhythm Holdings on Monday reported unauthorized activity involving data maintained on some third-party applications last ‌week, but said it has not identified any impact ‌on products, patient safety or medical device systems from the cyber attack. Here are ​some details from the company's filing: • The medtech firm said it identified unauthorized activity on June 8 and launched an investigation with external cybersecurity experts. • The company added that on June 9 it ‌received a payment demand ⁠from a "threat actor" claiming to have obtained proprietary data, patient protected health information and other personal ⁠information. • iRhythm deemed the incident material on June 10 due to the volume of potentially affected data. • Based on current investigations, the ​inci

Glance, the Singapore-based adtech subsidiary of InMobi, posted 33% revenue growth to $97.6 million for FY2025, with losses shrinking from $111 million to $106 million. The company is betting on Glance AI, a fashion recommendation tool that earns money from affiliate links and advertising. Founder Naveen Tewari expects Glance to hit profitability in June 2026.

At India's AI Impact Summit, major firms pledged billions for AI infrastructure. Reliance/Jio committed $110 billion, Adani Group $100 billion for renewable-powered data centers, and Microsoft $50 billion for Global South AI expansion. India aims to grow its data center capacity fivefold to 8 gigawatts by 2030, backed by tax holidays until 2047.

Nvidia signed deals in South Korea with memory chipmaker SK hynix, AI cloud provider Naver, and Doosan Group. The highlight is a multi-year partnership with SK hynix to develop next-generation memory for AI data centers. SK Telecom also announced plans to build a gigawatt-scale AI cloud in South Korea using Nvidia technology, with the first data center going live in 2027.

Salesforce cut fewer than 1,000 jobs across marketing, data analytics, and its Agentforce AI product team—even as the company posted record quarterly revenue of $10.26 billion. The layoffs show a pattern where AI tools drive growth for their makers while reshaping roles within the same company. CEO Marc Benioff has also said AI is replacing about 4,000 customer support roles.

Food delivery giant Swiggy partnered with Indian AI startup Sarvam to let users order food using voice commands in 11 languages including Hindi, Tamil, and Bengali. The feature works even without a smartphone through phone-based ordering. This aims to bring e-commerce to millions of Indians who struggle with English-first apps.

# Microsoft in talks to provide Anthropic Maia chips [News](https://www.techinasia.com/)Premium * [News](https://www.techinasia.com/) * [Jobs](https://www.techinasia.com/jobs/search) * [Companies](https://www.techinasia.com/companies) * [Events](https://www.techinasia.com/events/live) * [About](https://www.techinasia.com/about) * [advertise](https://www.techinasia.com/advertise?utm_source=techinasia.com&utm_medium=website&utm_campaign=intl-navbar) * [Asia Economic Summit](https://aes.techinasia.com/?utm_source=navbar-tia-homepage&utm_medium=website&utm_campaign=events-flagship-aes) * [Sign Up](https://www.techinasia.com/auth) * [Log In](https://www.techinasia.com/auth?ref=nav_login) [Subscribe](https://www.techinasia.com/subscription) [I'm fundraising](https://www.techinasia.com/companies/

_Summary:_ * Startups often hit a scaling wall when venture credits expire, finding themselves burdened by overengineered technical architectures and rising cloud costs that threaten their survival. * Oracle offers an alternative to traditional hyperscalers, providing high-performance infrastructure and “converged databases” that allow startups to simplify their tech stacks and manage multicloud environments more efficiently. * By focusing on AI readiness, Oracle helps startups in various sectors including biotech, integrating practical AI solutions that drive worker productivity and market expansion while maintaining critical security and governance. * Learn more about how Oracle can help you build, scale, and innovate with confidence through this [webcast](https://bit.ly/oracledn-webcast

Cerebras makes enormous AI chips that can run AI computations up to 15 times faster than top Nvidia GPUs, thanks to having memory built directly onto the chip rather than requiring separate DRAM. The company went public in May 2026 with a massive $5.5 billion IPO. However, getting these chips into Southeast Asia faces hurdles from export restrictions and limited cloud availability. At the SuperAI conference in Singapore, Cerebras is pushing its chips as purpose-built for AI rather than repurposed graphics hardware.

China is aggressively integrating AI into its military capabilities, viewing AI supremacy as central to global power status. The country is investing heavily in autonomous weapons, AI-enhanced surveillance, and decision-support systems. This military AI race has global security implications as nations compete for technological advantage.

Alibaba released new AI models specifically designed to power robots, marking a shift in the AI industry away from chatbots toward physical AI agents that can interact with the real world. This reflects a broader trend of AI companies developing specialized models for robotics applications rather than general-purpose conversational AI.

Qualcomm CEO Cristiano Amon declared 2026 'the year of AI agents' at MWC Barcelona. He believes AI agents will shift the digital world from smartphone-centric to agent-centric models. Qualcomm also launched the Snapdragon Wear Elite chip for AI-powered wearables, with Samsung, Google, and Motorola already signed on to build devices. Lenovo revealed its own AI 'super agent' called Qira. The idea is that wearables like smartwatches and smart glasses could feed agents constant real-time data about your surroundings.

Singapore's IMDA and Microsoft have signed an MOU to work together on AI safety and security. They will conduct joint research on agentic AI, develop evaluation tools and benchmarks, share best practices on AI governance, and explore how governments can responsibly access frontier AI models for safety testing. The partnership aims to build a trusted AI ecosystem through collaboration between government and industry.

China-ASEAN Investment Cooperation Fund II (CAF II), a $1 billion fund managed by ESR, has made a strategic investment in Racks Central, a Singapore-headquartered data center platform with growing regional presence. This signals continued investor confidence in Southeast Asia's data center infrastructure despite rising energy and land costs.

Enterprise identity management firm SailPoint is acquiring Tel Aviv-based Entro Security for an undisclosed amount. The deal aims to tackle the surge in machine identities—bot accounts, API keys, and service credentials—created as companies deploy AI agents across their operations. These non-human identities are a growing security vulnerability that traditional tools don't fully address.

Singapore fintech Aspire reveals that AI platform competition in Asia is heating up, with Anthropic's Claude rapidly gaining ground on OpenAI's ChatGPT. The data shows an increasing number of startups across Asia are now using multiple AI platforms simultaneously rather than picking just one, as AI becomes a daily tool for operations.

Singapore has launched a new $10 million lab called the Materials Data Foundry, a collaboration between NUS and the University of Toronto, to use AI to rapidly discover recipes for making next-generation semiconductor chips and clean hydrogen. Part of Singapore's $120 million national AI-for-Science programme, the lab can run thousands of experiments in days instead of months, potentially accelerating breakthroughs in critical technologies that power everything from phones to renewable energy.

[ ](https://profiles.yahoo.com/brands/reuters/) Reuters Tue, 16 June 2026 at 6:58 am UTC 1 min read FILE PHOTO: AI (Artificial Intelligence) letters and robot hand miniature in this illustration taken, June 23, 2023. REUTERS/Dado Ruvic/Illustration/File Photo/File Photo PARIS, June 16 (Reuters) - French Prime Minister Sebastien Lecornu said on ‌Tuesday the gover

An octopus does not need its central brain to control every grab, twist, or squeeze. Now, robots may not need to either. Engineers have published research highlighting a robot arm modeled after an octopus's, which can register a touch and begin holding an object without waiting for instructions from a central computer, according to [Earth.com](https://www.earth.com/news/octopus-inspired-robot-arm-can-sense-and-grip-on-its-own/). ### What happened? Researchers in Italy have developed a soft underwater robotic arm that mimics the limb of an octopus, leading to some fascinating findings. An octopus was chosen for inspiration since its tentacles don't always require instructions from its central brain. The project was led by Barbara Mazzolai at the Genoa-based Italian Institute of Technology,

Respond.io, a Malaysian startup using AI agents to handle customer service messages at scale, has raised $62.5 million and is eyeing acquisitions in North America and Europe. The company charges businesses per conversation rather than per user seat, making it potentially cheaper for high-volume support operations.

Nvidia CEO Jensen Huang, who grew up in Singapore, says AI is advancing so fast that society needs entirely new 'social norms' to handle it. He's calling for fresh rules around how AI is used in daily life—from deepfakes and job displacement to privacy and truth online. This matters because Huang leads the world's top AI chip company, so his views shape how the industry thinks about its responsibilities. Everyday impact: if new norms are created, they could affect how your job, privacy, and online information change.

G7 leaders are reportedly discussing creating a 'trusted partners' framework that would give certain countries preferential access to cutting-edge US AI models. The talks aim to balance AI innovation with security concerns, potentially creating tiers of access based on geopolitical alliances and safety standards. This could reshape how AI technology is shared across nations.

Intel has begun production of its new 18A-P chip process, a key milestone for the company as it works to regain ground against TSMC and Samsung in advanced chip manufacturing. The 18A process uses new transistor architecture and is designed to meet surging demand from AI companies needing powerful, efficient chips.

Microsoft's discussions with Oracle about sharing cloud infrastructure have reportedly broken down, according to Business Insider. The talks had explored potential collaboration on data center and cloud resources, likely driven by growing AI demand straining both companies' infrastructure capacities.

OpenAI burned through $39 billion in 2025, up from $5 billion the year before, and the company is now weighing dramatic price hikes for ChatGPT users. Instead of flat monthly subscriptions, CEO Sam Altman has suggested a future where users pay per token like a utility meter. Some observers have compared this to a drug dealer model: give it away cheap to hook users, then raise prices once they're dependent. A $200 monthly Pro plan can cost OpenAI up to $14,000 if used heavily. If pricing shifts to consumption-based billing, everyday users could see their AI bills fluctuate unpredictably.

RoboStrategy CEO Andrew Kang says advances in AI are making autonomous machines smarter, and that robotics could be the market's next big investment theme. He points to falling costs and growing capabilities as key drivers.

A French startup is taking a different approach to AI robots by designing machines that don't look like humans. While many companies race to build humanoid robots, this startup believes specialized, non-humanoid designs can better handle specific tasks. The idea is that a robot designed to clean floors doesn't need to walk on two legs — it just needs to clean well. This signals a growing debate in the robotics industry about what form AI machines should take.

Many businesses are scrambling to build out their AI infrastructure — the foundational systems needed to run AI applications at scale. The article covers why companies are finding this challenging and what they're doing to catch up.

A marketing executive shares how his fear of AI diminished the more he actually used it. Rather than replacing his job, he found AI handled repetitive tasks like drafting copy and generating ideas, freeing him to focus on strategy and creativity. His takeaway: anxiety about AI often fades once you get hands-on experience with it.

Lorong AI is Singapore's government-backed hub for AI practitioners to meet, collaborate, and share ideas. Launched by MDDI in January 2025, it now has over 260 paying members and has recorded 4,000 event attendees in 2025 alone. The hub recently expanded from a 70-person WeWork space to a 140-seat venue at one-north's Vidacity. Over 14 collaborations have already sprouted from connections made there, including a research paper on detecting hostile AI agents. The initiative aims to build the human networks Singapore needs to compete in the global AI race, complementing the upcoming Kampong AI development.

Malaysians see strong productivity gains from artificial intelligence (AI) as many as eight in ten Malaysians report that AI tools have helped them save time and improve efficiency, a poll from Ipsos showed Tuesday. This is reinforcing expectations that AI will continue to reshape how they work in the years ahead, according to the report. These tangible benefits help explain the growing acceptance and adoption of AI across the country. While Malaysians generally believe AI can improve their own jobs, only around half are confident that AI will improve the broader job market and economy. This gap indicates the view on AI is anchored in daily productivity, while its wider economic and societal impact remains less fully grasped. Meanwhile, more than 60 percent of Malaysians are comfortable wi

The artificial intelligence (AI) platform race is becoming increasingly competitive across Asia's startup ecosystem, as new data from Singapore-based Aspire shows Anthropic's Claude rapidly gaining ground on OpenAI's ChatGPT, while a growing number of startups are adopting multiple AI platforms simultaneously as AI becomes embedded into day-to-day operations.

New research from NTT DATA shows the global insurance industry is at a turning point as risks outpace its capacity to absorb them, with cyber protection emerging as the single largest source of uninsured risk worldwide. AI is both creating new cyber risks and potentially offering solutions for insurers to better assess and price risk.