Singtel, Microsoft and NVIDIA Dial Up AI Over 5G | NVIDIA Weblog






For telcos around the globe, one of many greatest challenges to upgrading networks has all the time been the query, “If you happen to construct it, will they arrive?”

Asia’s main telco, Singtel, believes the important thing to serving to prospects innovate with AI throughout industries — for every thing from visitors and video analytics to conversational AI avatars powered by giant language fashions (LLMs) — is to supply multi-access edge compute companies on its high-speed, ultra-low-latency 5G community.

Multi-access edge computing, or MEC, strikes the computing of visitors and companies from a centralized cloud to the sting of the community, the place it’s nearer to the client. Doing so reduces community latency and lowers prices by means of sharing of community sources.

Singtel is collaborating with Microsoft and NVIDIA to mix AI and 5G, so enterprises can increase their innovation and productiveness. Utilizing NVIDIA’s full-stack accelerated computing platform optimized for Microsoft Azure Public MEC, the telco is creating options that allow prospects to leverage AI video analytics for a number of use circumstances and to deploy 5G conversational avatars powered by LLMs.

From Sea to Shore

Singtel has been rolling out enterprise 5G and MEC throughout ports, airports, manufacturing services and different places. Along with operating low-latency purposes on the edge utilizing Singtel’s 5G community, the answer has the potential to remodel operations in sectors similar to public security, city planning, healthcare, banking, civil service, transportation and logistics. It additionally presents excessive safety for public sector prospects and higher efficiency for finish customers, enabling new clever edge eventualities.

Prospects can use these capabilities by means of Microsoft Azure, solely paying for the quantity of compute and storage they use for the length by which they use it. This replicates the cloud consumption mannequin on the community edge and lets customers save on extra operational overhead.

Edge Applied sciences

Singtel is working with video analytics software-makers collaborating in NVIDIA Inception, a free program that provides startups go-to-market help, experience and expertise. These ISVs will be capable to use the NVIDIA Jetson Orin module for edge AI and robotics along side Microsoft MEC to determine visitors flows at airports and different high-population areas, retail video analytics and different use circumstances.

Singtel and NVIDIA are additionally showcasing their expertise and options, together with a real-time LLM-powered avatar developed by system integrator Quantiphi and based mostly on NVIDIA Omniverse digital twin expertise, at a Could 11 launch occasion in Singapore. The avatar, constructed with NVIDIA Riva speech AI and the NeMo Megatron transformer mannequin, allows individuals to work together in pure language on any subject of curiosity. Companies can deploy these avatars anyplace over 5G.

Utilizing Singtel’s high-speed, low-latency 5G — mixed with NVIDIA AI accelerated infrastructure and capabilities — enterprises can discover use circumstances on every thing from pc imaginative and prescient and combined actuality to autonomous guided autos.

Singtel plans to broaden these new capabilities past Singapore to different international locations and affiliated telcos, as properly. This collaboration will assist redefine what’s attainable by means of the highly effective mixture of compute and next-generation networks, unlocking new operational efficiencies, income streams and buyer experiences.

Share this


Ashes 2023: Steve Smith takes a hilarious jibe at English media over haircut controversy; shares an image with Alex Carey

From Mitchell Starc’s catch to Jonny Bairstow’s controversial dismissal, the Ashes 2023 has been within the headlines for thus many causes....

FAA Publishes Air Taxi Implementation Plan

picture: FAA The FAA has issued an implementation plan for integrating air...

Recent articles

More like this