🌟 Editor's Note
Welcome! This week’s roundup spotlights companies moving from pilot to production with AI: telcos opening 5G-A+AI sandboxes, manufacturers standardizing LLMs on the shop floor, housing platforms going agentic, and legaltech rolling out domain-secure copilots. Below you’ll find 12 bite-sized pieces you can paste straight into Beehiiv—each with a quick take on what’s happening, how AI is being implemented, expected impact, how it fits global trends, and sources.
🏰 Grand Korea Leisure starts an AI innovation team
What’s happening. South Korea’s casino operator Grand Korea Leisure (GKL) created a new team to push AI across its business.
How it’s implemented & impact. The team’s job is straightforward: find places where AI can make work faster or service better, then run pilots and scale what works. Examples reported by trade media include smarter customer service and using robots to move chips and cards so staff can focus on guests. That can cut queues, reduce mistakes, and improve security on the floor.
Trend fit. Hotels, casinos, and theme parks are using AI to predict demand, schedule staff, and personalize offers. This is now basic “table stakes” in top venues.
Bottom line. Expect GKL to start with quick wins (guest help, fraud checks, rostering) and only then take on bigger changes.
🏠 AMRS’ “Verom” brings AI + IoT to housing societies (India)
What’s happening. AMRS launched Verom, a housing platform that blends cameras, sensors, and simple apps to manage buildings.
How it’s implemented & impact. The idea is to connect common-area devices—doors, energy meters, CCTV—to a building app. AI can spot unusual activity, suggest ways to save power, and remind managers to fix things before they break. For residents, that means better security, fewer outages, and faster issue resolution. For building owners, it means lower running costs and higher satisfaction.
Trend fit. Around the world, “smart community” tools are moving from single apartments to whole complexes. Edge devices do quick checks locally; heavier analysis runs in the cloud.
Bottom line. Verom’s value improves as more blocks join, because shared data makes predictions better.
📶 U Mobile + Huawei open a 5G-A + AI testbed for enterprises (Malaysia)
What’s happening. U Mobile and Huawei set up an Enterprise Innovation Platform (EIP) so companies can try 5G-Advanced and AI solutions before rolling them out.
How it’s implemented & impact. Think of it as a sandbox with real network gear. Firms can test factory cameras, logistics tracking, or city sensors, and measure results on a carrier-grade setup. That shortens the time from idea to real deployment and reduces risk.
Trend fit. Telcos worldwide are moving “beyond connectivity” by bundling networks with AI apps, so customers buy outcomes (fewer defects, faster delivery), not just bandwidth.
Bottom line. Expect more paid pilots and new B2B revenue for the telco—and faster AI adoption for local manufacturers and city agencies.
🇱🇰 Sri Lanka’s first National AI Expo sets a bold growth target
What’s happening. Sri Lanka hosted its first National AI Expo and said AI could drive 10–12% of its digital-economy growth by 2030.
How it’s implemented & impact. The event gathered government, telcos, and tech firms to kickstart real projects—like AI helpdesks for public services and tools for small businesses. Making this an annual event keeps momentum and gives firms a place to meet partners and buyers.
Trend fit. Countries are using national showcases to speed up corporate adoption, set standards, and fund pilots—Sri Lanka is following that playbook.
Bottom line. Clear goals plus convening power can unlock private-sector deals; watch for procurement guides and sandboxes next.
🧩 Zindi challenges help Kenyans get hired
What’s happening. A study reported that about 1 in 5 Kenyan users on Zindi—an online AI challenge site—landed jobs or career moves.
How it’s implemented & impact. Companies post real problems (say, predicting demand or spotting fraud). Participants submit models and build a public track record. That portfolio helps hiring managers see who can “do the work,” not just talk about it. The study says strong profiles and teamwork raised the chances of getting hired.
Trend fit. Firms are increasingly using challenge platforms to find niche skills quickly—especially for data engineering and MLOps.
Bottom line. If you need AI talent, posting a challenge can be faster and cheaper than a long hiring cycle.
🧠 Innosilicon’s Fenghua No.3 GPU: a local compute push in China
What’s happening. Innosilicon announced its Fenghua No.3 GPU with a big memory budget and claims of CUDA compatibility.
How it’s implemented & impact. The card targets AI and graphics. If the specs hold up, Chinese firms get more options for training and running models without importing restricted chips. Real performance data is still limited, so early uses will likely be inference and smaller models.
Trend fit. Many countries want “sovereign compute.” Even if these chips lag the very top tier, having local supply can unlock a lot of practical work.
Bottom line. Watch for benchmarks and software support before making big bets, but expect steady improvements.
🧬 Huawei vs. Cambricon: China’s two “mini-Nvidias” race ahead
What’s happening. Coverage this week examined how Huawei and Cambricon are building home-grown AI chips and software to reduce reliance on Nvidia.
How it’s implemented & impact. New AI models like DeepSeek-V3.2-Exp are tuned to run well on China-native accelerators (Huawei Ascend, Cambricon), shrinking the software gap. This could make day-to-day AI work cheaper for Chinese firms even if raw speed trails Nvidia’s best.
Trend fit. “Good enough, available now” chips plus improving tools are often more important than chasing the absolute fastest hardware.
Bottom line. Expect faster adoption of local chips for inference, with training moving over as tools mature.
⚡ KEPCO & TNB team on AI-ready power tech (Korea–Malaysia)
What’s happening. Reports highlight cooperation between Korea’s KEPCO and Malaysia’s TNB on renewable projects and digital “smart power” tech that uses AI.
How it’s implemented & impact. Joint work includes digital power platforms that forecast demand, spot faults quickly, and improve grid stability. This can reduce outages and cut fuel costs by making plants and lines run more efficiently.
Trend fit. Utilities are adding AI to predict loads, schedule maintenance, and balance renewable energy. Results show up as fewer blackouts and lower operating costs.
Bottom line. Cross-border utility partnerships speed up adoption because teams share designs and software, not just MoUs.
🌊 Underwater data centres to cool AI cheaply (China)
What’s happening. China is moving from tests to commercial underwater data centres that use sea water to keep servers cool.
How it’s implemented & impact. The approach can cut cooling energy by around 90% and run mostly on renewables. That reduces the cost and carbon of running AI models. There are engineering challenges (saltwater corrosion, maintenance access), but new coatings and above-sea access points help.
Trend fit. As AI workloads grow, operators are experimenting with smarter cooling—underwater pods, desert locations near solar, and direct-to-chip liquid cooling.
Bottom line. If local rules allow, expect more trials where power is cheap and water conditions are stable.
🛡️ Nokia + Extreme Broadband to secure Malaysia’s AI data centres
What’s happening. Nokia and Extreme Broadband agreed to upgrade security and networks for Open DC’s data centres across Malaysia.
How it’s implemented & impact. The deal links six AI data centres with high-speed, “quantum-safe” networks and better DDoS protection. Stronger, automated networks keep AI services up, even during attacks, and make it easier to expand later.
Trend fit. As companies put AI into production, the weak point is often the network between data centres. Upgrades like this are becoming routine.
Bottom line. Expect better uptime and faster rollouts for Malaysian firms building AI services on Open DC.
🏛️ Kazakhstan creates a Ministry for AI (policy signal to companies)
What’s happening. Kazakhstan announced a government shake-up that includes a new AI Ministry—an official signal that AI projects are a priority.
How it’s implemented & impact. When countries set up dedicated AI agencies, companies usually see faster permits, clearer rules, and more pilot funding. That cuts red tape for local and foreign firms building AI products. (Specific launch details are evolving.)
Trend fit. More governments are creating AI-focused bodies to coordinate rules and support industry.
Bottom line. Expect shorter paths from proposal to pilot for domestic energy, mining, and public-service AI projects.
🧾 KPMG & industry warn: use AI in security, but keep humans in control
What’s happening. A 2025 outlook from KPMG and other studies say firms are adopting AI for cybersecurity, but they’re keeping people in the loop.
How it’s implemented & impact. Companies use AI to spot weird network behavior, filter alerts, and draft incident reports. This saves time and helps teams react faster. Surveys in Malaysia and globally show high use, but also caution—privacy, cost, and skills still matter.
Trend fit. “AI + analyst” is becoming the norm; full automation is rare.
Bottom line. Don’t just buy tools—train people and set clear rules so AI helps rather than harms.
🤝 Sarawak (Malaysia) explores AI + semiconductor ties with South Korea
What’s happening. Sarawak and South Korea’s Rebellion Inc. discussed collaboration in AI and semiconductor manufacturing—aimed at building local capability.
How it’s implemented & impact. Partnerships like this often start with talent programs, small pilot lines, and joint labs. For local companies, it can mean earlier access to new chips and design know-how; for partners, it opens a new market.
Trend fit. Regions are competing to host AI supply chains. Sharing investment and training lowers risk for both sides.
Bottom line. If deals close, expect new jobs in chip packaging, testing, and AI model deployment in Borneo.
📈 Momenta seeks funding at ~US$6B valuation (China autonomy software)
What’s happening. Autonomous-driving firm Momenta is in talks to raise funds at about a US$6B valuation.
How it’s implemented & impact. Momenta sells software that carmakers install for assisted-driving features (lane-keeping, highway pilot). Money raised would help put the software into more models and markets, which is how this sector now makes steady revenue.
Trend fit. Investors are favoring “driver-assist now, robotaxi later” companies that can ship software through automaker partners.
Bottom line. Expect more Level-2+/Level-3 features in mass-market cars in China first, then beyond.
Did You Know? In 2016, Google put DeepMind’s AI in charge of data-center cooling and cut the energy needed for cooling by up to 40%—one site even saw overall efficiency improve by about 15%.
Till next time, Use AI responsibly
AI Corporate Trends
An Elythra Studio Production